Wikipedia talk:Did you know/Statistics/Archive 1

Worst of DYK

We can also learn what not to do from really bad hooks. The hooks that trigger a "Who cares?" response. I don't think we want a "Worst of DYK" page, but if people want to offer examples of particularly ineffective hooks that illustrate what not to do, they could be included here.Cbl62 (talk) 15:33, 20 November 2008 (UTC)[reply]

Best?

Is this really the right name for this page? They may have had the most page views, but were they the best-written, best-sourced, most interesting etc? I doubt that they were all. "Best" is always someone's opinion, so can never be determined in the form of a list. – How do you turn this on (talk) 23:33, 20 November 2008 (UTC)[reply]

Point taken, but while best-written and most interesting are subjective, page views are objective. Since the goal of DYK is to attract readers and editors to newly created/expanded pages, these are the hooks that have performed best in accomplishing that objective. Cbl62 (talk) 09:05, 13 December 2008 (UTC)[reply]

something with traffic statistics link

I tried to check hits of Ganesha in 07-2008,[1] it shows 0 hits from 13-31, which is impossible for an article which gets atleast 1k hits everyday. --Redtigerxyz ^Talk 10:14, 13 December 2008 (UTC)[reply]

Henrik's page view tool was not operating during the last half of July 2008, and so there are no stats available for that time. Cbl62 (talk) 16:31, 13 December 2008 (UTC)[reply]

Split page?

This page is currently 156kb long and growing by the month. I presume something is being done about it and I imagine that when 2009 rolls round all the 2008 months would go into a 2008 archive but I just thought I'd point it out in case it hasn't been thought of. Also how about making each month a separate page which can then be placed on this as a sequence in a template-like fashion? That ought to shorten it down nicely and remove all the excess weight from the main page! ;)

I'd do it myself if no one objected, or at least help out. --➨♀♂Candlewicke _S^T_# :) 03:11, 28 December 2008 (UTC)[reply]

I agree. Once we hit Jan. 2009, we can put the months from 2008 into an archive. Cbl62 (talk) 08:57, 28 December 2008 (UTC)[reply]

Self-serving busy work?

Except for the odd DYK from May of this year, all the rest come from August 2008 to December 2008. Looks like a lot of busy work (for those editors gathering more trinkets, like for RFA or something). Do you really want to encourage editors to spend time going through their DYKs over the last many years to get listed on this page? Really? Does this sort of thing contribute to the worth of Wikipedia? Also, please indicate where consensus for the inclusion rules were developed and any statistical research supporting the reasoning behind the rules. —Mattisse (Talk) 18:39, 29 December 2008 (UTC)[reply]

Actually, on second thought, this looks like an easy way of racking up edits without having to do any thinking. And with zero possibilities for conflicts! —Mattisse (Talk) 18:56, 29 December 2008 (UTC)[reply]

The point is?

I perused down the many different section of this page, many of the same sections repeating the same article/picture listed in other sections. Just for fun I picked an article, Needlegun_scaler which was a DYK April 10, 2008. True, it had over 16,000 views the day it was on the main page, but ZERO views for the remainder of the month. —Mattisse (Talk) 18:52, 29 December 2008 (UTC)[reply]

The point, as I see it, is that the information on page views gives DYK participants useful information on the types of hooks that attract the most interest on the main page. Not every DYK hook can or should be one that generates >5,000 views (DYK is not just for sensational facts and we need to maintain subject matter balance), but information on page views provides info on user interests and good clues for writing effective hooks and selecting effective images. --Orlady (talk) 19:25, 29 December 2008 (UTC)[reply]

I share Mattisse's concerns. A DYK with the following would get many, many page views, but does that help anyone? (Apologies for the language in advance, but it's to demonstrate the issue):

Did you know that former US President Bill Clinton once fucked a White House intern with a cigar before ejaculating over her dress?

A list of "what works" is not really worthwhile. A page explaining why such hooks work would be, in my own view GTD 20:28, 29 December 2008 (UTC)[reply]

An essay on "what works" would be a nice feature, but I don't know how one would begin to write such an essay without first compiling information about effective hooks. Furthermore, even with information about page views it is not entirely clear why some hooks were particularly successful. What could have led someone to predict that Ivan Castro (soldier) would get more page views than Todd Palin, or that Akutan Zero and Millard House would be exceptionally successful as DYK entries? --Orlady (talk) 21:04, 29 December 2008 (UTC)[reply]

Where does the "useful" information on the "value" of hooks come from? Raw view numbers seem fairly meaningless as a measure of hook value. Do you control for the "first in queue" effect? Do you control for nationality or fame of subject? For example, anything with president-elect Obama in it would get a fair number of hits, regardless of the hook. Whereas, the same exact hook with the president-elect of Niger would get much less. I have looked at hit rates of FAs and found them quite low compared to bad articles on popular subjects. —Mattisse (Talk) 21:23, 29 December 2008 (UTC)[reply]

And some would get, for example, Slashdotted no matter how good or bad the hook was. There are too many variables to derive any useful information GTD 21:38, 29 December 2008 (UTC)[reply]

Further, these "statistics" do not take into account Wikipedia's enormous increase in readership, nor the "partnerships" Wikipedia is forming with other organizations to increase wikipedia traffic. Thus, comparisons over time of "hits" per page are enormously misleading. I am tempted to nominate this collection of misinformation for deletion. —Mattisse (Talk) 22:30, 29 December 2008 (UTC)[reply]

As far as I'm concerned, a DYK hook is effective if it induces just one knowledgable and interested person to visit an article and improve it -- or link to it from another article. Notwithstanding the numerous factors that affect page views (including hook position, number of hours the hook was on the main page, time of day the hook appeared, day of the week the hook appeared, etc., etc.), I think it is interesting to see which hooks attracted the most attention, and it does help me in writing/editing more effective hooks. If you don't like the page, please just ignore it. --Orlady (talk) 01:25, 30 December 2008 (UTC)[reply]

Just wandered by and wanted to counter some of these old grumbles by saying - what a great page, entertainment has its place alongside education, and well done to all for maintaining it. Ben Mac Dui 19:59, 10 May 2011 (UTC)[reply]

Thanks for the compliment. It's a team effort, and a bit of fun. —Bruce1ee^talk 05:14, 11 May 2011 (UTC)[reply]

Observation

The current all-time top twenty hit hooks are all produced by different nominators. So, either nobody found the magic hit-making formula yet, or a formula only works once... Oceanh (talk) 02:02, 30 December 2008 (UTC).[reply]

As has been pointed out above, there are so many variables impacting page hits that have nothing to do with the hook, that it is very unlikely anything about a particular hook is going to have a noticeable impact, with the exception of certain "words" or a famous name like Britney Spears. Most of the time hooks are of little interest and don't make much sense anyway.

That is why this whole project, Wikipedia:DYKBEST is a fools errand and a way of diddling time away that could be used in building an encyclopedia. But editors love toys, trinkets, and awards, even if meaningless. —Mattisse (Talk) 02:12, 30 December 2008 (UTC)[reply]

I agree that there are all sorts of factors that can impact page views, including time of day that it is on the Main Page, whether it has the #1 slot with photo, and weekend vs. weekday. But what I have found in keeping these statistics is that a really interesting and well-written hook can get 5,000 page views even if it's the last hook in the list and featured in the middle of the night or on a weekend. And a really boring, poorly-written hook won't garner 500 page views even if it's featured in prime time. Too many hooks that make DYK and the main page are completely boring, and by showing the types of hooks that work, we can improve the quality of the content on the main page, and also achieve our goal of attracting editors/readers to new content. Cbl62 (talk) 02:28, 30 December 2008 (UTC)[reply]

Could you provide some examples of subjects most people would not know about or be interested in that a well-written hook brought in a large number of page views? I check page traffic frequently and find page traffic depends far more on the subject than anything else. Has the number of DYK's changed? The list on the main page looks longer and more boring lately. I just counted up one person's DYKs for December and found he has 58 so far this month, with as many as 14 on a single day. And the articles were boring, My checking them may count as a page view, but it doesn't count as a page read. I also noticed that very little "improvement" of the articles took place after their DYK, except by bots. —Mattisse (Talk) 02:41, 30 December 2008 (UTC)[reply]

Leading hook

Who determines which hook gets the favored spot of lead hook? —Mattisse (Talk) 02:44, 30 December 2008 (UTC)[reply]

Let us look at this from a statistical point of view. If we assume that roughly one fourth of the nominations have a supporting picture, the chance is about 50% for a nomination with picture to be lead hook. Hooks with "good" pictures have higher odds, as the selecting editor presumably picks pictures appealing to her. Oceanh (talk) 03:10, 30 December 2008 (UTC).[reply]

Article traffic statistics down

Just FYI, Henrik's article traffic tool seems to not have been collecting data since 18 March (so that's 4 day so far: 18, 19, 20, and today). rʨanaɢ ^talk/_contribs 18:18, 21 March 2009 (UTC)[reply]

nvm, seems to be back now, and even the old data is up. rʨanaɢ ^talk/_contribs 09:55, 24 March 2009 (UTC)[reply]

DYK stats for non-DYK article?

It looks like Biosphere 2 got a lot of hits in June 2009 when the Ed Bass article was featured in DYK. Is it fair to feature the hook on the DYKstats page, when the article that got the hits wasn't the new article that the hook was supposed to showcase? --Orlady (talk) 15:12, 17 July 2009 (UTC)[reply]

No, I am going to remove it. --BorgQueen (talk) 15:13, 17 July 2009 (UTC)[reply]

Glad to know that we agree. :-) --Orlady (talk) 16:22, 17 July 2009 (UTC)[reply]

Now wait just a minute here :-) The DYK certainly did result in the hits. Why wouldn't it qualify? Not, perhaps, the intended effect, but still, a very noticeable jump. Frank | talk 16:49, 17 July 2009 (UTC)[reply]

Irrelevant. The purpose of this list is to showcase DYK articles with large number of hits, not some random article the hook happened to link to. Besides, do you expect us to check the hits of every single article a hook links to? --BorgQueen (talk) 17:02, 17 July 2009 (UTC)[reply]

Certainly not, and I don't think anyone expects it's an all-inclusive list anyway. I don't feel strongly about this; I'm not fussing so much as wistfully looking on... I was just pleased with the result. I had actually created Ed Bass over a year earlier, and then expanded it with newfound research superpowers, so the article itself is (dare I say it) somewhat dear to me... Frank | talk 17:10, 17 July 2009 (UTC)[reply]

Although it's inevitable that nontarget articles will get attention from DYK hooks, When I see a nontarget article getting a whole lot more hits than the DYK article, it tells me we need to work harder to write hooks that draw the reader's attention to the DYK article. (I had a DYK a while back for biography articles about two of the discoverers of the element promethium, and it was understandable but a bit disappointing to see that the article about promethium got the lion's share of the attention.) --Orlady (talk) 17:23, 17 July 2009 (UTC)[reply]

Sigh. It's a real shame, too, because Bass turns out to be quite an interesting guy. Frank | talk 17:25, 17 July 2009 (UTC)[reply]

(ec) First, getting listed here does not mean that the article was well-written, and may not even mean that the hook was well-written, i.e. you shouldn't take this list too seriously. Second, getting emotionally attached to an article you wrote is, in a long-term perspective, not a good idea. It often leads to claims or ideas of article ownership, and I've seen some excellent editors leave Wikipedia when their beloved creations get ruined or messed up for one reason or another. Being proud of your contribution to this project is one thing, falling in love with articles you wrote is quite another. The former motivates you, the latter blinds you. --BorgQueen (talk) 17:29, 17 July 2009 (UTC)[reply]

Oh, I know, I know...that's why I put the "(dare I say it)" bit in parens there. Besides, I have no emotions. Why do you think I was interested in writing an article about a guy who wants to replicate Mars on Earth? ;-) Frank | talk 17:33, 17 July 2009 (UTC)[reply]

On 31 July my Dnestr radar got 2303, but the link to K Project got 8207! It's still getting over 1000 a day, clearly as a a result of being a link on DYK earlier. Previous baseline was less than 100. Secretlondon (talk) 11:12, 9 August 2012 (UTC)[reply]

That's interesting!! I can't believe that DYK alone was responsible for the uptick in viewership, though. There must be some sort of external attention directed to The K Project, possibly as a result of DYK or possibly just a coincidence independent of DYK. If you look at stats history for most successful DYK hooks, you will see that attention falls off immediately when the DYK hook leaves the main page. --Orlady (talk) 13:54, 9 August 2012 (UTC)[reply]

Page view statistics

Where did you get the page view statistics for 200908 because Henrik's counter only shows 200907 on my pages. Mario1987 20:17, 4 August 2009 (UTC)[reply]

The way I get it is by just changing the url at the top from 200907 to 200908. It's not on the drop-down menu, but it works. Little Mountain 5 18:32, 6 August 2009 (UTC)[reply]

Thanks. Mario1987 20:52, 6 August 2009 (UTC)[reply]

Amphibious helicopter gets 11k or 29k?

I just counted the main link target and collateral link page views related to the DYK entry Amphibious helicopter and got 11,000 for the main one plus 18,000 for the various other links. Can I say that it got 29k or should I just list it as 11k? Binksternet (talk) 23:48, 29 September 2009 (UTC)[reply]

I believe it should only get 11k page views. The other linked articles in the hook may have attracted plenty of hits but the DYK article is what is being highlighted here, and it only got 11k. See also the DYK stats for non-DYK article? discussion above. --Bruce1ee^talk 07:19, 30 September 2009 (UTC)[reply]

Data loss in late September?

The page view stats tool seems to have little or no data for several dates in the last week of September 2009. Any information or theories on whether the logs exist -- or whether the data are totally lost? --Orlady (talk) 18:01, 11 October 2009 (UTC)[reply]

The raw data still exists here, 24 hourly files for each day. If you want to manually extract stats for an article, download the files (this can be quite time consuming), uncompress them, then using a text editor, find the article in each file. There are two numbers after each article name, the first is the number of page views for that article in that hour (I don't know what the second number is). What I don't understand is, if the raw data still exists, why stats.grok.se can't be rerun for those missing days. --Bruce1ee^talk 05:19, 12 October 2009 (UTC)[reply]

odd patterns

Just wondering...I don't want to argue for my hook being listed here, but I wonder if anyone familiar with DYK stats has an idea about the odd pattern shown in [2] for April 2008 and first day of May, after it was DYK on April 23, 2008. Carter_Hall_(Millwood,_Virginia). It's like it had a 2nd life. 2.5k hits first day, drop to nothing but then climb to total 8k over 8 days. Not like any others i've seen. --doncram

Another one I don't understand at all is Ponce City Hall, which shows zero(!) hits on day before, day of, and day after DYK posting on 10/15/2009, but some hits before and since. And it had an included picture. doncram (talk) 02:47, 22 October 2009 (UTC)[reply]

stats.grok not working?

It appears that stats.grok.se isn't working at present, with no updates in a few days. And http://dammit.lt/wikistats/archive/2009/10/ only has dumps up to 11 October. Is there another service which is available, as I just wanted to check on page views for a recent DYK of mine. Cheers, --Russavia ^{I'm chanting as we speak} 11:36, 23 October 2009 (UTC)[reply]

The files for the last few days are here. They haven't been archived yet. --Bruce1ee^talk 11:47, 23 October 2009 (UTC)[reply]

A user familiar with how it works has stated that on certain days the program for the tool had problems with the complilation of the raw data and therefore did not update. Unfortunately on those days the information will most doubtfully stay blank. It is unfortunate with all the users that enjoy this handy tool. Many users have tried to reach Henrik, but am unsure of his responce. ♠ B.s.n. ♥R.N. 12:04, 23 October 2009 (UTC)[reply]

Bruce Lee, Is there any chance you might be able (or willing) to create a newer tool :) ♠ B.s.n. ♥R.N. 12:09, 23 October 2009 (UTC)[reply]

I'd love to, but I'm not a programmer and I wouldn't know where to start. On a couple of occasions I have downloaded files and extracted stats manually, but it's a very time consuming process. --Bruce1ee^talk 13:04, 23 October 2009 (UTC)[reply]

Tell me about it, I am compiling stats for my Air Botswana expansion at moment, and after compiling only 8 hours work of stats so far I get 8591 views, which is obviously well over the 1900 that stats.grok.se tells me for 20 October. Unfortunately, I had to download 22 hours of worthless stats in order just to get this far :D --Russavia ^{I'm chanting as we speak} 13:22, 23 October 2009 (UTC)[reply]

Not sure...

... if I'm missing something, but I think another one of mine should qualify [3] (Czestochowa massacre) for November, with 8.2k. I don't know 100% sure how this works so I might be wrong.radek (talk) 02:42, 7 December 2009 (UTC)[reply]

It does qualify, and I've added it. BTW it's 7.5k, not 8.2k (it's the number of views on the day it was on the Main Page). --Bruce1ee^talk 07:04, 7 December 2009 (UTC)[reply]

just a question.

I was looking at the page views for Trail of the Whispering Giants that I got to DYK, and the day it was on there, it got 7.7k hits, then the next day it got 9.8k. Is this because it was technically the next day in parts of the world when it was up, or am I missing some reason for it to have gotten so many hits the day after it appeared? I'm just a little lost.--Found5dollar (talk) 18:59, 15 December 2009 (UTC)[reply]

As DYK is currently running three queues a day for eight hours each, 16 of the hooks will be shown for brief periods on two days. This appears to be to case with your hook. Must say 17000+ views is quite impressive! Calmer Waters 19:49, 15 December 2009 (UTC)[reply]

Thanks for clearing this up! I'm happy so many people got exposed to these incredible sculptures.--Found5dollar (talk) 01:31, 16 December 2009 (UTC)[reply]

? Lead DYK

How can you tell if an article was the lead DYK? Midford Castle was a DYK on the front page on 18 January 2008 & got approximately 12225 hits, the section on this page is for non lead articles & I can't remember that far back whether it was the lead, but I do know it included the symbol (♣) which might have made it more interesting.— Rod ^talk 17:12, 24 December 2009 (UTC)[reply]

Did you check the archive? --BorgQueen (talk) 17:14, 24 December 2009 (UTC)[reply]

I looked at Wikipedia:Recent additions 200 but can't tell from that.— Rod ^talk 17:28, 24 December 2009 (UTC)[reply]

According to the archive page, it was not the lead, since the hook didn't have "(pictured)". --BorgQueen (talk) 17:31, 24 December 2009 (UTC)[reply]

We didn't have a picture for the article then. Shall I add it to the list?— Rod ^talk 17:36, 24 December 2009 (UTC)[reply]

Sure, go ahead. --BorgQueen (talk) 17:40, 24 December 2009 (UTC)[reply]

zero views?

hey, i have a question. i got Winnipesaukee Playhouse into DYK but when i looked at it's trafic statistics, it says it had zero views on the 8th, the day it was on the front page... anyone know why this is?--Found5dollar (talk) 15:13, 10 February 2010 (UTC)[reply]

It appears to be a problem with the source files. Five hours are missing on February 8 (see here). A similar problem occurred on January 23 & 24 (see Henrik's talk page). --Bruce1ee^talk 15:40, 10 February 2010 (UTC)[reply]

Ah so it is just a problem that nothign was counted during that time period because of a glitch. I was just worried that less people looked at it when it was on then main page than when it wasn't.--Found5dollar (talk) 21:15, 10 February 2010 (UTC)[reply]

Distortion in DYKstats

Something that's been bugging me for a while about the leader table - it doesn't take account of the length of time the hook was on the main page. For example, the leading hook, Ivan Castro (soldier), was up for 12 hours - twice that of the normal 6-hour time period.

I would like to suggest then, that the hook view numbers be adjusted according to the length of time they appeared on the mainpage, and that the leader board be sorted accordingly. So, for example, the Ivan Castro page views would be multiplied by 6/12, the standard time period divided by the actual time period. That way, we would get a much more accurate list of the hooks which were actually the most popular. Gatoclass (talk) 01:29, 29 March 2010 (UTC)[reply]

Full support, but we need to do that for all hooks in the list. Suggest for now adding hours for those we know; e.g., "Dec. 8, 2008; 12 hrs"; having the data it is easy to (re)process it. Materialscientist (talk) 01:40, 29 March 2010 (UTC)[reply]

With regard to this point, what determines whether a hook is up for 12 hrs or just 6 hrs?radek (talk) 07:06, 29 March 2010 (UTC)[reply]

Well, I guess we could just make a list here, couldn't we? Might as well get started -

Ivan Castro - 12 Gatoclass (talk) 02:11, 29 March 2010 (UTC)[reply]
Leonard Siffleet - 6½ Gatoclass (talk) 07:03, 29 March 2010 (UTC)[reply]

Hmm, just realized another problem - some articles get lots of views even without being on the main page. The Todd Palin page, for example, was getting around 30k hits already at the time it was featured. We may need to adjust for that too. Gatoclass (talk) 02:15, 29 March 2010 (UTC)[reply]

On the other hand, the Palin hook pattern is very odd - it seems he got a lot of hits for a week or two around the time the hook was featured, and then the number of hits dropped dramatically. Not sure what to make of that. Gatoclass (talk) 02:21, 29 March 2010 (UTC)[reply]

Such viewing pattern - high number of views around the featuring day - is not unusual for DYK stats and there is no easy solution. In science, we would do background subtraction. The simplest form is: I = I_n – (I_n+1 + I_n-1)/2 where I_n is number of views on day n. This is problematic here because views for some hooks are divided between two report days. Thus maybe I = I_n – I_n-1. Materialscientist (talk) 04:00, 29 March 2010 (UTC)[reply]

I don't think that is quite the same situation. If you take a look at the Todd Palin page, he was getting a lot of hits for about a week, and then suddenly they dropped away again by an order of magnitude. Maybe he'd been in the news for some reason or another. Whatever the reason, I'm not sure how to handle that particular one. But hopefully, that kind of pattern does not occur too regularly. Gatoclass (talk) 06:56, 29 March 2010 (UTC)[reply]

The activity dropped significantly more than a day after DYK which is long enough; if use "background correction" then the count for this hook should be 29,900. Materialscientist (talk) 07:07, 29 March 2010 (UTC)[reply]

There are many factors that impact the number of views that a hook gets. The issue is not limited to length of time on the main page, but also includes "background" interest in the hook (as MatSci pointed out), weekend vs. weekday, daytime vs. overnight time slot, lead slot vs. buried in the middle, days when data is missing, etc. With the data we have available, I don't think we can accurately adjust the hit counts to account for all these variables. As for number of hours on the main page, we have no data to show how many hits a 12-hour hook received in the first six vs. last six hours. If the first six hours were daytime and the next six hours were overnight, it's likely that far more than 50% of the hits came during the daytime hours. Further, I suspect that a hook that's up for 12 hours of equally desirable viewing time still achieves diminishing hits in the later hours that it's on the main page. If we simply cut the hits in half, that will not be an accurate reflection of the hooks achieved in the first six hours. That said, I share your concern about statistical distortion. That's why, for example, we established a second "all time best" list for non-lead hits. I've sometimes thought it would be interesting to generate a list of the hooks that have gotten the most hooks on a weekend time slot or an overnight time slot.

We could try similar statistical analyses to balance out hit level based on when a hook was featured, but I think that would be (a) very time consuming, (b) imprecise, and (c) undesirable.

With respect to the "background noise" issue (as in Todd Palin or the recent hook about the Green Lantern movie), BorgQueen and I had developed a practice in the early days of excluding hooks from DYKSTATS unless there was a marked increase (of at least 5,000 hits) above the background level while the hook was on DYK. I believe the Todd Palin hook was the subject of discussion at the time as to whether or not it should be included - the hit pattern was complicated by the ups and downs of Palin news at the 2008 Republican National Convention that was going on at the same time. We should take care not to give DYKSTATS credit to a hook that has simply ridden a wave of publicity unrelated to being featured in DYK.

I believe that the data doesn't allow us to accurately reduce the number of hits, and that the best way to deal with the skewing is to provide more information. I recently started tracking some of the data in my sandbox here. In this vein, we could add an additional column(s) to the DYKSTATS leaders to reflect key points impacting on hits. Or we could create a separate chart identifying issues that potentially impact the hits for the leaders. I think the latter is probably preferable to avoid unduly cluttering the main DYKSTATS leader board. But if someone wants to experiment in user space, that might provide some specific data for consideration. Cbl62 (talk) 18:22, 29 March 2010 (UTC)[reply]

I'm not at all persuaded by the argument that we shouldn't adjust the numbers by the number of hours on the page because of this or that other variable. Many of the other variables you mention can't be measured reliably if at all, whereas the number of hours a hook spends on the front page is obviously a major distortion and one that can be corrected readily. Yes, sometimes hooks do get featured at more or less popular times but that's just the luck of the draw. When a hook spends twice the time on the front page as another however, it's obviously going to get far more hits, and something needs to be done to correct that. Gatoclass (talk) 07:52, 31 March 2010 (UTC)[reply]

I agree that the other variables I mentioned can't be measured reliably. But the same holds true with hours on the main page. Simply cutting the hits by 50% for a 12-hour hook is not reliable. Hit levels are likely to be highest in the first 6 hours a hook is on the Main Page, and there are many other factors impacting whether the hit levels will be higher in hours 1, 2, 3, etc. We have no way of determining how many hits a hook received in the first 6 hours. Cbl62 (talk) 06:18, 20 April 2010 (UTC)[reply]

Green Lantern (film)

The Green Lantern (film) article was getting heavy hits for several day before and after being featured on DYK. Over the past 15 days, it's been averaging around 3,500 hooks a day. So the "boost" while it was featured on DYK was only around 4,000 extra hits per day. Under DYKSTATS Rule #3, I think we have to subtract the baseline of 3,500 hooks a day to measure the DYK "boost" and not include the background. The purpose of DYKSTATS is to measure the boost in views resulting from being on DYK, and not simply to reward articles about topics that are timely and getting lots of hits. For this reason, I'm inclined to remove it from the all-time leader board. Anyone disagree? Cbl62 (talk) 05:34, 31 March 2010 (UTC)[reply]

Fine with me. There should be several cases like that in monthly stats, but I won't bother with correcting those. Materialscientist (talk) 05:55, 31 March 2010 (UTC)[reply]

Basically, most movie and "recent celebrity" stats are blown up. Materialscientist (talk) 05:57, 31 March 2010 (UTC)[reply]

I agree with you that there's probably no need to fix them all in the monthly stats, but for the all-time list, we should not include it unless the "boost" is greater than 11,000 hits. Cbl62 (talk) 06:19, 31 March 2010 (UTC)[reply]

If you check out the background hits on Fishsticks (South Park) and The Ring (South Park), they should probably also be dropped from the all-time list. In the case of The Ring, the article received an average of more than 12,000 hits per day during the 10-day period before and after DYK, and so the "boost" on the DYK day was only 7,000 above the average. In the case of Fishsticks, the ten-day average is about 19,000, and the "boost" on the DYK day was about 4,000 above average. Under Rule #3, neither of these should be on the all-time list. Cbl62 (talk) 06:30, 31 March 2010 (UTC)[reply]

That would seem to make sense. Gatoclass (talk) 09:48, 31 March 2010 (UTC)[reply]

Picoaza

Interestingly, not only did the uber-obscure Picoazá article get more than 5,000 page views from the DYK entry on April 1st - but the work done to expand the article in order to turn it into a DYK entry directly caused two other articles to be created to fill in red-links from Picoaza!

Manteño civilization (for which archeological evidence was discovered at Picoaza).
Jacinto Jijón y Caamaño (who was the archeologist who found that evidence).

This is a shining example of the value of promoting minor articles on the front page in this way.

Not bad for something that started as nothing more than a way to get a silly line from write-in candidate onto the April fools day front page!

Many thanks DYK team! SteveBaker (talk) 03:35, 3 April 2010 (UTC)[reply]

April 1 2010

provided 13 hooks which formally qualify for the "Hall of fame" (20k+/11k+ views). How should we add them there? The pictured 20k+ section is not a problem, but the 11k+ section does not distinguish April 1 or not, and I hesitate mixing April 1 with serious hooks there. Materialscientist (talk) 04:45, 5 April 2010 (UTC)[reply]

I don't see why there should not be dates for the 11k section as well, so I would just go ahead and add them together with the date. Gatoclass (talk) 04:50, 5 April 2010 (UTC)[reply]

Brad Stevens

Should Brad Stevens qualify for the HOF? I believe the hook straddled April 5-6 in wikitime, but I'm sure the hits are almost entirely due to the national championship game being played rather than the hook being awesome. Thoughts?

For reference, its "my" article, and currently is listed (not by me) and appears to be using the April 4 number. Unless I am misunderstanding something that isn't correct. --ThaddeusB (talk) 04:15, 8 April 2010 (UTC)[reply]

Don't know about this one - it was specifically rushed to the queues so that the DYK launch would be right before the game - can't tell how to separate these 2 factors. Materialscientist (talk) 04:12, 8 April 2010 (UTC)[reply]

If you look at the stats for the five-five day period surrounding the Final Four, the article averaged about 30,000 hits per day. Not surprisingly, the biggest hit days were the days when Stevens' team played games against the Spartans and Blue Devils. Using a five-day average of 30,000 hits, Stevens' article only received 11,000 hits above the five-day average on the DYK day. That was also the day of the Butler-MSU game. Accordingly, I don't think it belongs in the HOF. Cbl62 (talk) 15:02, 14 April 2010 (UTC)[reply]

Proposed revision to HOF

The time is coming when we may need to re-think the thresshold for the HOF lists. Originally, we chose the 11,000 and 20,000 thresholds based on a review of data indicating that these were pretty extraordinary achievements. We've now got almost 75 non-lead hooks that qualify and almost 50 lead hooks that qualify. At some point, the lists will be too long. Maybe already on the non-lead hooks? Should we modify the threshold for the HOF lists? Maybe limit the list to the Top 50 hit recipients? Top 75? What's a reasonable upper limit for such a list? If we limit the lists in this manner, I would strongly favor being more rigorous in excluding hooks for items that were "in the news" and thus receiving high numbers of hits independent of DYK. Thoughts? Cbl62 (talk) 15:15, 14 April 2010 (UTC)[reply]

Multis

I notice that Cbl has removed a number of multi-hooks from the lists. I support the inclusion of multis in the list, I think it penalizes users who write multis to exclude them. Multis are written for one reason only, which is to save space for the benefit of the project, but the downside is less interesting hooks and less attention for the individual articles. So users are already penalized by writing multis. There is also on average far more work in putting together a multi than in submitting a single article, and that effort is fully deserving of recognition. I therefore believe that multis should be included in the "view leaders" list where appropriate. Gatoclass (talk) 21:49, 27 April 2010 (UTC)[reply]

Fully agree. Otherwise we would stimulate authors to cheat, breaking up multis into singles even when they fit nicely into one hook. Materialscientist (talk) 22:18, 27 April 2010 (UTC)[reply]

Since the inception, we have not added the counts of multis. In the past month, some people have been adding them. I removed them because that's how it's been done from inception. If there is consensus to change the rules (is there?), I don't feel strongly about it. But this is definitely a change in how things have been done. Cbl62 (talk) 05:08, 29 April 2010 (UTC)[reply]

I understand that, but would you please consider joining the "consensus" (I guess three of us are an active part of the DYK stats page)? Multies are common, and hook reviewers and selectors tend to push for combining them. If this reflects negatively on the total views then we're artificially creating an unhealthy situation when the nominators "benefit" from avoiding submissions which can be combined. On the other hand, allowing them add the views does not harm anything. Materialscientist (talk) 05:24, 29 April 2010 (UTC)[reply]

It certainly makes sense to encourage multis, and we have done that already with a special chart on the "Hall of Fame" subpage. So long as the composite count is limited to the featured new/expanded articles (and doesn't include non-new/expanded articles that also happen to be linked in the hook), I'm OK with the proposal. Cbl62 (talk) 05:48, 29 April 2010 (UTC)[reply]

I'm sure everyone can agree that non-DYK links should not be counted, so this issue appears to be settled. Gatoclass (talk) 06:20, 29 April 2010 (UTC)[reply]

Not everyone has agreed in the past that non-DYK links should not be counted. See "DYK stats for non-DYK article" above. I just want to be sure that by opening up DYKSTATS to composite counts, someone doesn't misconstrue that as a reversal of the prior resolution on the non-DYK links. Cbl62 (talk) 08:18, 29 April 2010 (UTC)[reply]

Fishsticks (South Park) and The Ring (South Park)

Recently, I've learned most of my previous South Park hooks were removed from the "Non-lead hooks over 11,000 views" list because the articles were already very popular on non-DYK days, so the popularity on the DYK hook day was disproportionately high. I see the merit in this logic, although I think a blanket removal of all hooks from popular articles is not at all fair, if there is evidence of a viewership spike on the day of the hook because of DYK. However, it was recently explained to me that if the hook date viewership exceeds the average viewership of the average views (half the sum of the date before and after the hook date), then it warrants inclusion. By that logic, I readded Fishsticks (South Park) and The Ring (South Park) because in both cases, those criteria were satisfied. (I did not readd others like Eat, Pray, Queef, Pinewood Derby (South Park), The Coon because they don't fit the criteria.) However, Fishsticks (South Park) was recently removed yet again because it was claimed the spike didn't exceed 11,000 of the past five days prior to the hook. This is a different criteria than was explained to me, and frankly, I don't think it's fair, because it includes dates that were disproportionately higher because of the original airdate and repeat dates. If you considered only the five days after the hook date instead of before, it would pass no problem. And if you compare it to the average viewership days in the last half of the month, it also more than exceeds 11,000 the average. Finally, the big test: if you add up every day of the month besides the hook date (including the two monstrous dates, which are 48,300 and 38,000) and take the average viewership, they come out to be 9,945. The hook is 23,500, which exceeds that average by more than 11,000. By any fair measure, Fishsticks (South Park) deserves to be in this list, and I think my argument is sound (and admittedly long-winded). I am planning to restore it, but wanted to put my logic here first. I welcome any replies... — Hun ter Ka hn 04:33, 29 April 2010 (UTC)[reply]

Re-added with adjusted views. After adjustment, both "parks" marginally exceed 11k, and I understand Clb62's removals, but we've got to keep some consistent count. Subtracting a half sum of the views on the days before and after DYK is the only proposal we had so far. Materialscientist (talk) 04:45, 29 April 2010 (UTC)[reply]

Adjusting the count to correct for the background noise is a reasonable way to resolve the issue. Mat Sci came up with the "background subtraction" idea, and it's a good one. But one day before and one day after may not be the right formula. Oftentimes, a hook gets heavier than normal views on the day after the DYK feature. If we only go +1/-1, we may penalize a hook that continues to draw attention the day after it was featured. I had been looking to +5/-5 days as a way to remove the background noise. Maybe five days is too much, but I think 1 day is too little. I suggest either +2/-2 or +3/-3 as a better screening formula. Cbl62 (talk) 05:27, 29 April 2010 (UTC)[reply]

"One day" silently includes that the views should drop to background level (we often have DYK related views spread over two days). Sure, it is a crude procedure, but we need to keep consistency. A few days is clearly too long; two days is Ok, but hardly better than one. Materialscientist (talk) 05:41, 29 April 2010 (UTC)[reply]

As a former history major who avoided "statistics" class like the plague while in college, I will defer to your judgment as the science guy. If you think +1/-1 is a better system, it's ok with me. I just want to avoid penalizing hooks that continue to draw attention in the day or 2 after being featured on DYK. I do agree consistency is desirable. There may be special cases where a different approach is needed, but if special cases arise, we can discuss here on the talk page. Cbl62 (talk) 05:54, 29 April 2010 (UTC)[reply]

Per the discussion above, I went ahead and used the +1/-1 "background subtraction" to adjust the hook count for Todd Palin and Brad Stevens. If anyone thinks these adjustments are unwarranted or inaccurate, let me know. Cbl62 (talk) 06:04, 29 April 2010 (UTC)[reply]

Good hooks

The 08:00, 11 May 2010 update had a particulary fine selection of hooks, it seems. Fully 50% of the hooks qualified for DYKSTATS - Crispin Black, Ricardo Blas, Jr., HMS Tynedale (L96), and Chronic Tacos - and, in addition to that, two "non-DYK" articles linked - RFA Sir Galahad (1966) (9,200 hits!) and Jimmy Hoffa - would have qualified had they qualified.[/Tautology]. Good work folks! - The Bushranger _{Return fire}^{Flank speed} 02:33, 14 May 2010 (UTC)[reply]

Wait, I'm confused...

... Would an article be added to the particular month if the page views are above 5,000 for simply just the day which it featured, or for the month containing the day in which it is featured? --^{SteelersFanUK06} HereWeGo2010! 01:14, 28 June 2010 (UTC)[reply]

Not sure I understand the question. The article can be added to the list if it received >5000 views on the day it was featured on the main page (with a note that those views should originate from the main-page featuring, and not just from the article popularity, that can be seen from the stats of the other days). Materialscientist (talk) 01:22, 28 June 2010 (UTC)[reply]

Oh okay I understand. Basically I thought because the page views tool displayed an entire month that the page was meant to be listed if they were >5,000 in the entire month, not just the single day. Thanks for clearing that up. --^{SteelersFanUK06} HereWeGo2010! 01:26, 28 June 2010 (UTC)[reply]

Larissa Riquelme

I added Larissa Riquelme (an article I nominated for Did You Know) to the statistics list. Admittedly, it had received over 5,000 views on the two most recent days before its DYK appearance for which statistics are available (four of the intervening days are missing). On the other hand, the 13,600 views the article received on its DYK appearance day were 7,000 more than it has received on any other day so far. I hope this works for everyone, and if not, I'm sure someone will say so. --Metropolitan90 (talk) 16:44, 14 July 2010 (UTC)[reply]

Corrected the stats, as agreed at "Distortion in DYKstats" section above. Materialscientist (talk) 22:30, 14 July 2010 (UTC)[reply]

Listing of Paul (octopus) and Mani the Parakeet

Hi there, I draw your attention to the listing of the DYK hook on Paul (octopus) and Mani the Parakeet under the "All-time DYK page view leaders" section. Rule 3 on the listing of DYK hooks states that "The purpose is to measure the boost in views from the DYK hook. Accordingly, an article that is consistently drawing large numbers of views and does not have at least a pronounced boost from its inclusion on DYK is not eligible for "STATS."" If you apply Rule 3 on the listing of that hook, it seems the hook might be non-eligible for listing as it is likely the articles involved were already drawing in large numbers of views when it appeared on the Main Page. If there is consensus on its ineligibility, it would have to be removed. What do you think? AngChenrui (talk) 05:26, 17 July 2010 (UTC)[reply]

It is the same story as 2010 Moscow Victory Day Parade below it - those DYK hooks appeared on the main page on the day(s) when the articles would be in focus of attention anyway. However, I do not see any reasonable way to estimate how many views originated from DYK activity. Materialscientist (talk) 05:38, 17 July 2010 (UTC)[reply]

We could do the +1/-1 background subtraction that has been done before. The problem is there is no data for the -1 day. If we look just at the +1 day, the "boost" on the DYK is not extraordinary from what was seen on the +1 day. Some sort of correction appears to be in order. Cbl62 (talk) 05:45, 17 July 2010 (UTC)[reply]

When you also account for the hits for the rename of Paul (octopus), the hits on the +1 day are actually higher then the hits on the DYK day. See http://stats.grok.se/en/201007/Paul%20(octopus). There were about 173,000 hits on the +1 day (July 12) for Paul, as compared to 152,000 hits ont he actual DYK day (July 11). Cbl62 (talk) 05:48, 17 July 2010 (UTC)[reply]

(ec) The actual number of views is above 320k because there were two articles + redirect (one article got moved), thus I did some very rude background subtraction and rounded to ~300k - as I mentioned, I see no way to have an accurate estimate in this case (yes, stats were pretty irregular and often distorted these days). Materialscientist (talk) 05:51, 17 July 2010 (UTC)[reply]

MatSci. I think your numbers are way off. You are counting hits for both July 11 and 12, but the hook was on DYK only on July 11. We don't credit the hook with hits from the day after the DYK. On the +1 day (July 12), the total hits for Paul (octopus) and Paul the Octopus are approximately ~~172,000~~180,000. On the DYK day, Paul got 152,000 hits. So, Paul got more hits on the +1 day than the DYK day. As for the parakeet, the hook got WAY less views (18.4 on DYK day and 9.5 on the +1 day). Two key pieces of evidence suggest the DYK boost was minimal (and nowhere near 300,000 hits). Those two key facts are (1) the Paul article got more hits on the day after the DYK (suggesting that the media attention rather then DYK is the driving force in the hits), and (2) the parakeet article did not attract big numbers even on the DYK day (again suggesting that the big hits for the octopus were not driven by DYK). Cbl62 (talk) 06:05, 17 July 2010 (UTC)[reply]

I know that a stable fraction of hooks get their views split up between two days. Do we know how to discard this effect? Materialscientist (talk) 06:18, 17 July 2010 (UTC)[reply]

If a hook is on DYK during any part of 2 days, we count both days. If a hook is on DYK for only a single day, we only count the one day. That is how we've done it from the beginning. In this case, Paul and Mani were on DYK exclusively on July 11. Cbl62 (talk) 06:23, 17 July 2010 (UTC)[reply]

<indent, ec>No hook has crossed the UTC day break in many months, yet the stats get split (I know it from some of my obscure noms). More convincing however is that other hooks of the very same DYK set got all their views on 11 July (example). Thus I would suggest ~160k (~150k + ~10k from Mani_the_Parakeet, dicarding the redirect, as its stats are quite odd in time). Thoughts? Materialscientist (talk) 06:29, 17 July 2010 (UTC)[reply]

Your example proves my point. When an article is not an "in the news" item, the hooks fall off dramatically the day after the DYK. But when an article is "in the news", it continues to get big hits after the DYK. That's because it's "in the news", not because it's on DYK. In this case, the stats seem to indicate that the huge hit levels received by Paul were caused by the "in the news" phenomenon and not a "DYK boost" of anywhere near 160,000. Even if we give the hooks the maximum benefit of the doubt, here’s what my calculations show based on the agreed +1/-1 background subtraction:
(1) Paul and Mani were featured on DYK exclusively on July 11. The hook did not spill over into July 12. The combined articles received 170,600 hits (152,200 + 18,400) on July 11.
(2) We then look at the hits received on the day before and after. Unfortunately, no data exists for July 10. Giving maximum benefit of the doubt to the hook, let’s use “0” as the hook count for July 10. We then look at the hook count for July 12, which is 182,800 (9,500 + 27,500 + 145,800).
(3) We establish a baseline by averaging the two days, resulting in a baseline of 91,400 (182,800 divided by 2).
(4) We subtract the baseline from the hits received on the DYK day. That results in a corrected hit total of 79,200 (170,600 – 91,400). Based on the foregoing, I cannot see how we would justify a DYK count above 79,200. Even that is inflated greatly because of the lack of data for July 10. But at least it is based on the methodology that has been agreed. Cbl62 (talk) 06:44, 17 July 2010 (UTC)[reply]

Background subtraction only works if it is relatively stable in time. Both the 10 May Parade and the Octopi were on the main page ITN. Imagine, for the sake of exercise, that the 10 May Parade was at DYK section on 9 May and ITN on 10 May. If so, it is not correct to subtract the data of these days. It goes back to separation of DYK/ITN views, which I don't know how to do. Materialscientist (talk) 06:54, 17 July 2010 (UTC)[reply]

I thought we weren't supposed to be promoting DYKs that are also up on ITN. If that happened, I agree that it could result in greater distortion. What day was Paul on ITN. If it was on ITN on July 12, you may have a point. It may have received an ITN boost on July 12 that is distorting the background subtraction. But if the ITN was on July 11 or 13, it's not so problematic. I don't know where the ITN archives are to verify when it was up on ITN. Do you? Cbl62 (talk) 07:03, 17 July 2010 (UTC)[reply]

For the 10 May Parade, things are clear - the same article slipped to ITN after it was featured on DYK. Octopi case is complex: they were featured at DYK on 11 July. On 12 July, ITN posted the world cup hook. It did not contain either of the octopi in the hook, but Paul was linked from within the hook's articles (in "see also", marginally, I agree). Counts are odd in that the unbolded ITN articles had many more views than bolded, i.e. go figure how the viewers got to the Paul's page .. Materialscientist (talk) 07:08, 17 July 2010 (UTC)[reply]

Do you remember Gato's suggestion - adding a tag on how long the article spent on the main page? I think we should just tag these two as being featured on DYK and ITN for two days - any views correction would be highly speculative here. Materialscientist (talk) 07:15, 17 July 2010 (UTC)[reply]

If Paul wasn't featured on ITN on July 12, I don't see it as a distortion problem. The reasons Paul got huge hits on July 12 are myriad but I presume the bulk is because tens of thousands of people around the world heard about Paul on the news and did a google search (or wiki search or other search) that brought them to the Wikipedia article. I have to sleep now, so think it over. But I think a correction to 79,200 is more than fair to the Paul hook. Even that figure IMO greatly distorts the DYK boost, but it has the benefit of being arrived at by a consistent methodology. Cbl62 (talk) 07:19, 17 July 2010 (UTC)[reply]

Although I don't follow Wikipedia:In the news (ITN) that closely, I doubt that Paul the octopus was actually featured there. However, he was mentioned in "the news", that is, newspapers, television, web sites, etc. throughout the world. I agree that we have to subtract the background hits since that is what had to be done with the DYK hook I nominated for Larissa Riquelme around the same time. --Metropolitan90 (talk) 19:59, 18 July 2010 (UTC)[reply]

Not to drag this back up, but shouldn't the traffic count (and the title, and nominations, and...) include the ones for the parakeet, too, since from the bolding it appears to have been a double-hook? - The Bushranger _{Return fire}^{Flank speed} 00:57, 16 November 2010 (UTC)[reply]

2010 Moscow Victory Day Parade

If we apply the agreed +1/-1 background subtraction method, 2010 Moscow Victory Day Parade does not belong on the all-time list. Here are my calculations:
(1) The article was featured on DYK on May 9 only. It's listing did not spill into May 10. It received 53,000 hits on the DYK day (May 9).
(2) We establish a baseline average from the day before and the day after DYK. The article received 84,500 hits on the day after the DYK (May 10) and 1,100 hits on the day before the DYK (May 8).
(3) Averaging the +1 and the -1 day, you get a baseline of 42,800 hits (84,500 + 1,100 = 85,600 divided by 2).
(4) When you subtract the baseline hits from the DYK day hits, the presumed "boost" using background reduction is 11,200 (53,000 – 42,800).
(5) Conclusion: The hook does not belong on the all-time list because the corrected "boost" of 11,200 hits is insufficient. Cbl62 (talk) 06:29, 17 July 2010 (UTC)[reply]

Here again, the other hooks of the same DYK sets were exclusively viewed on 9 May (example), but. I have no idea how to subtract background, because it comes from ITN, where that article was featured around 10 May - ITN lasts longer than DYK. My personal opinion is most those views came from the main page, but how to split DYK and ITN? Materialscientist (talk) 06:35, 17 July 2010 (UTC)[reply]

I apologize for the multi-track discussion I've now started. But it's an important point. The Black-throated Gray Warbler example again proves my point. A "normal" DYK get big hits on the day it's featured and then falls off dramatically. That's true of the Black-throated Gray Warbler. The fact that an "in the news" item gets hits the day before and after it's featured on DYK is due to the fact that it's "in the news." The only way to eliminate the "in the news" distortion is to strictly apply the +1/-1 background substraction formula that you suggested. Otherwise, the all-time list is going to be dominated by hooks that get huge hits because they are "in the news" and not because of a true "DYK boost." My view is that we have to pick a formula (which I thought we had done) and stick to it. Cbl62 (talk) 06:53, 17 July 2010 (UTC)[reply]

Well, as I recall, this hook got ITN well after being featured at DYK. Then the DYK/ITN views should be independent, but I don't know the time stamps (crucial here is how the WP stats server gets its views, they don't seem to be UTC daytime bound). Materialscientist (talk) 06:57, 17 July 2010 (UTC)[reply]

In this discussion, I was using the phrase "in the news" in the non-wikipedia sense. I hadn't realized that the parade was a Wikipedia ITN item. If the ITN came after May 10, it should not distort the +1/-1 background subtraction. And as for Henrik's counter, it's my understanding that it is strictly bound by UTC time. And if he's missing data for part of a day, then he chucks the whole day and doesn't report the results. See this diff from his Talk page. Cbl62 (talk) 07:12, 17 July 2010 (UTC)[reply]

Well, DYK sets do not cross the UTC day break, but I do get some tail views for obscure articles (example). Materialscientist (talk) 07:18, 17 July 2010 (UTC)[reply]

I've seen that kind of tail, too, but only on hook groups that expire at exactly 00:00. I've not seen that on a hook group expiring at 6:00, 12:00 or 18:00. My belief has been that this is because the bot changes the cue at exactly 00:00 but it may be a few minutes later before it gets reset and actually loaded to the main page. I may be way off base, but that was my understanding. BTW Paul and Moscow do not fall into that, as they were mid-day shifts not ending at 00:00. Cbl62 (talk) 07:27, 17 July 2010 (UTC)[reply]

The fractions are always too large, i.e. they correspond to about 1 hour or longer, not minutes. Maybe there is a time delay in the flow of raw data to Henrik :) Materialscientist (talk) 07:30, 17 July 2010 (UTC)[reply]

Archiving June 2010 toplist

Hey, is there any reason why the June list of top articles was not archived yet? - ArCgon (talk) 16:05, 25 July 2010 (UTC)[reply]

Some editors still add into the last month. There were major interruptions in the WP stats in June-July. Materialscientist (talk) 22:42, 25 July 2010 (UTC)[reply]

An option for solving DYK stats

Hello, I had posted something similar earlier. What I am proposing is something along the lines of the following:

As a page in the DYK section goes onto the main page, a bot creates a temporary page based off of the article, either something with a prefix or suffix like /Paul the octopus_dyk
this page is simply a redirect to the page /Paul the octopus
1. If it is a possibility, remove the small "redirected from...." text in this instance.
Henrik's traffic stat page will be set to monitor the temporary page /Paul the octopus_dyk and not the page /Paul the octopus
1. This will allow for measurement of page hits only off of the DYK space, as any hits from either searches, or potential ITN items will be linked to the regular page /Paul the octopus and not the DYK one.
After the article is removed from the main page DYK, the stats can be archived and the redirect removed permanently (again, likely by some bot).

Regarding the other gigantic variables, like if you find out your DYK is going up in Chicago at 1AM... or its on a beautiful Friday night... or during the Superbowl or any other reason which would affect traffic to the site, I believe there is nothing you can really do about that and it's not worth complaining about. This is also the same for articles that appear during a DYK when they are relevant to viewers current interest, like Paul's during the World Cup or whatever, that's the name of the game! Anyway, I hope this is helpful, let me know what you think! - Theornamentalist (talk) 02:24, 26 July 2010 (UTC)[reply]

Technically speaking, it is a good solution. Pros are accurate stats. Cons are (i) double redirects from the main page, (ii) confusing article names for admins directly fixing main page errors. (iii) creating/deleting 32+ useless pages/day (some noms are multiple, i.e. some 12,000 pages/year), i.e. spamming WP. Pros and cons are lightweight, thus IMO there should be a strong motion for it to make it real (i.e. ask Shubinator to tweak the DYK bot). One technical comment is the redirect should be full protected, against vandalism, I hope the bot can do that. Materialscientist (talk) 03:26, 26 July 2010 (UTC)[reply]

I'm not sure if a bot can do that; I've left a message for Shubinator. Either way, I agree with a lot of the stuff in the relevant items mentioned in prior discussions, the entire top DYK is in need of a great overhaul; I think that even making the format simply smaller, like smaller pics, more condensed boxes and stuff like that. Also, I think that we could potentially create separate top scores for ITN DYK's as opposed to obscure ones that have attained many hits by great hooks or something. I know that my proposal would create a lot of spam, but I think it would be worth it (as trivial as top DYK kind of is). I mean, this system exists for a reason, and we should try to do it right, whichever way that may be - Theornamentalist (talk) 20:46, 26 July 2010 (UTC)[reply]

The core of what you're proposing is technically possible. I'm more concerned about the edge cases and general impact. For example, it's easy enough to figure out the DYK article and replace the link when it's formatted like [[Article name]] or [[Article name|Displayed name]]. However, a not-insignificant fraction of DYK articles are linked through a template or other means (military ship DYKs are a prime example). So if we still went ahead with the proposal, most DYK articles would have a dedicated redirect but some wouldn't. This creates more problems with maintenance in the long run; it would make it tougher for an admin to step in if something happened to the bot or an article needed to be swapped out. Also, while redirects are cheap, they're not free, and I'm hesitant to intentionally add redirects to the Main Page. Shubinator (talk) 05:58, 27 July 2010 (UTC)[reply]

I didn't know about template links; I understand your concern for avoiding any potentially complex maintenance issues. Well, figured the idea was worth a shot, hopefully we can figure out something one of these days. - Theornamentalist (talk) 17:40, 27 July 2010 (UTC)[reply]

Henrik's traffic counter ain't workin'

for 28 Jul and later. Kayau Voting IS evil 02:26, 1 August 2010 (UTC)[reply]

It was quite unstable this summer. Experience tells that some time during next week Henrik might recover the missing stats. Materialscientist (talk) 02:48, 1 August 2010 (UTC)[reply]

I think it is still not working correctly. I see pages which used to get hundreds of hits a day, suddenly with none or a handful. - Tim1965 (talk) 13:14, 23 August 2010 (UTC)[reply]

When I am in doubt, I look at the current and past stats of some article like diamond (stable FA with very few edits but many views). Materialscientist (talk) 23:39, 23 August 2010 (UTC)[reply]

I can't help but notice that the article diamond went from 4,600 page views on 2010-08-21 to zero over the past six days. My sense is that Henrik's counter is picking up some articles, but not all. And it is crashing frequently. An article which might get 14 page views per day should get a couple hundred by being a DYK. To have it drop to zero indicates a problem. But, it's a beta thing, and done gratis, so I can't really complain. - Tim1965 (talk) 15:39, 28 August 2010 (UTC)[reply]

I noticed too and saw a few people inquired on his talk page. It went down a week or two ago and then came back up with the recovered hits.--NortyNort (Holla) 16:24, 28 August 2010 (UTC)[reply]

Adding up stats

I notice that recently some users add up stats of a multiple hook (hook with more than one bold link) and list its total hits here (like this), but this practice poses some problems. For example, this would place hooks with 5-6 bold links on the list even when each of the bold links got barely more than 1,000 hits individually. Any thoughts? --BorgQueen (talk) 16:44, 17 August 2010 (UTC)[reply]

See section "Multis" above. Not counting multiple noms might (and likely did) result in situations when the authors split up and delay parts of their multiple noms into individual noms, which is not productive, especially with the current nomination backlog. Materialscientist (talk) 22:35, 17 August 2010 (UTC)[reply]

Lucan portrait of Leonardo da Vinci

I don't know how to go about adding it to the stats, but it got 5,100 hits.

Amandajm (talk) 22:02, 15 November 2010 (UTC)[reply]

Just a move

See WT:DYK#Wikipedia:DYKSTATS. Simply south (talk) 17:40, 29 November 2010 (UTC)[reply]

This link to this, now archived, is Wikipedia talk:Did you know/Archive 62#Wikipedia:DYKSTATS. SJ Morg (talk) 14:01, 14 August 2011 (UTC)[reply]

Dumb question....

It appears this list is manually created. Is there any reason it's not automated? It seems like that would be simple, no? Maury Markowitz (talk) 19:43, 9 January 2011 (UTC)[reply]

Nobody came up to write a bot/script for that. I myself don't see this as a priority. Materialscientist (talk) 23:30, 9 January 2011 (UTC)[reply]

April 1st 2011

I just added my April 1st hook to the top list, but there are a good number of this years hooks that qualify for top honors. I will start going throiugh and adding them, but if anyone else wants to help, by all means, please do. --Found5dollar (talk) 03:28, 2 April 2011 (UTC)[reply]

On second thought, i don't know how to figure out who nominated the article..... Could someone else please do this, or inform me how to?--Found5dollar (talk) 03:41, 2 April 2011 (UTC)[reply]

I was thinking to put up all 1 April stats anyway, thus if you're in doubt I can do all that. Materialscientist (talk) 04:00, 2 April 2011 (UTC)[reply]

Thank you for compiling all that information Materialsceintist. Will you also be adding the top hooks to the "Lead hooks with over 20,000 views" and the "Non-lead hooks with over 11,000 views" sections?--Found5dollar (talk) 02:54, 3 April 2011 (UTC)[reply]

Yes. I was in and out and thus had to split the updates. Materialscientist (talk) 03:03, 3 April 2011 (UTC)[reply]

Proposal to revise Non-lead hook threshhold

The size of the DYKSTATS page is getting unwieldy at more than 110,000 bytes (and growing). When I started the page, I arbitrarily picked the 20,000- and 11,000-hit threshholds for lead and non-lead hooks to be included on the all-time Hall of Fame lists. We now have > 120 non-lead hooks that have passed that threshhold. I propose revising the threshhold to 13,000 for non-lead hooks. This will help us control the size of the page and bring the number of non-lead hooks qualifying for the all-time HOF into line with the number of lead hooks. If there are objections or alternative solutions, let me know. If not, I'll go ahead and make the switch next week. Cbl62 (talk) 04:59, 8 April 2011 (UTC)[reply]

I agree, the non-lead list is too long and should be reduced. It's just a matter of deciding what the new threshhold should be. Upping it to 12,000 would knock off 25 hooks, and to 13,000, another 25 hooks. I think 12,000 would be fine, but then I suppose there's nothing wrong with 13,000. —Bruce1ee^talk 05:39, 8 April 2011 (UTC)[reply]

Page view stats

One question I'd like to ask is that can a page that has featured on DYK but gained a higher number of views when not on it, can it be included? A case in point I have is one of the articles I created, Baron Carrickfergus gained 5,200 views during DYK but however 6 days before it got 18,600 views so could the higher mention be included in one of the lists on this page or not since it was on DYK? The C of E. God Save The Queen! (talk) 10:21, 8 May 2011 (UTC)[reply]

We decided to look at the stats on the DYK day and subtract a half sum of the views on the days before and after it. In your case, even if the views drop to zero on May 8, the adjusted DYK counts will be below 5k. Materialscientist (talk) 10:25, 8 May 2011 (UTC)[reply]

Image column

Is the image column in "Archive of monthly DYKSTATS leaders" only for images that appeared on the main page or could other images be added there as well? bamse (talk) 09:13, 10 July 2011 (UTC)[reply]

Only for those images which appeared on the main page. Materialscientist (talk) 10:05, 10 July 2011 (UTC)[reply]

OK. bamse (talk) 13:46, 10 July 2011 (UTC)[reply]

Pageview stats tool down

Henrik's pageview stats tool appears to be down and has not been updating article stats since July 12 (see User talk:Henrik). Hopefully it will be up again soon, and will include the missing stats for the past few days. —Bruce1ee^talk 05:36, 14 July 2011 (UTC)[reply]

... and it's up again, with all the missing stats. —Bruce1ee^talk 13:10, 15 July 2011 (UTC)[reply]

article receiving zero views?

Hi, recently Wage reform in the Soviet Union, 1956–1962 was on DYK and according to the stats tool the article got zero views! In fact the article seems to be showing as zero page views at all since the page was moved (a couple of days before the DYK appearance) after it was decided to give the article a more appropriate title. Is there anything I can do to fix this or should I just give it up as a lost cause? cya Coolug (talk) 07:29, 7 September 2011 (UTC)[reply]

Henrik's pageview stats tool is down again, since September 2 (see User talk:Henrik). This happens from time to time, but it normally comes back again with the missing stats. —Bruce1ee^talk 07:54, 7 September 2011 (UTC)[reply]

Hey. Thanks for the reply. Are you sure? There are stats for yesterdays main page featured article, the main page and a few random articles. I'm not saying you are wrong as it could only be a problem that is only affecting some pages, but during my very unscientific experiment I couldn't find any other articles experiencing the same problem. Coolug (talk) 09:14, 7 September 2011 (UTC)[reply]

It seems the tool is working again. Here are the stats for Wage reform in the Soviet Union, 1956–1962. —Bruce1ee^talk 10:16, 7 September 2011 (UTC)[reply]

Thanks! cya Coolug (talk) 12:45, 7 September 2011 (UTC)[reply]

Views / hour

In light of the fact that DYK is now on a 12-hour cycle, the issue of "normalization" was raised here (perm.link), and Gatoclass suggested adding a "Views per hour" column. I've taken the bold step of adding such a column to the September table. It's not complete yet, I'll get to it, and I'll also add the column to the other tables. I've put a note at the top of the table explaining how it is calculated. —Bruce1ee^talk 07:17, 29 September 2011 (UTC)[reply]

Lead images with over 20,000 hits

The hook for Island of Lost Men, with a picture of Anna May Wong, leaves me a little confused as to whether it should be included or not in the leader's list. Seeing as the tool does not count the single hits after 1,000, it could theoretically have made 20,001 hits or so. Should the wording be changed to "at least 20,000 hits", and the hook listed at the bottom, or kept the same and the hook left out? Disclosure: It was my hook. Crisco 1492 (talk) 08:14, 15 October 2011 (UTC)[reply]

With 20,000 hits, it appears to me to qualify for the all-time list. Congrats. Cbl62 (talk) 08:23, 15 October 2011 (UTC)[reply]

Thanks. The current wording ("more than 20,000 hits") might be best changed... I notice that the beginning of the page uses "at least 5,000 hits" for the bottom section. Crisco 1492 (talk) 08:36, 15 October 2011 (UTC)[reply]

There is some inconsistency on this page. Rule 4 says "at least 5,000" and later "more than 5,000". The rest of the page uses "over x views". User:Rjanag's Pageview stats page, which is linked to in the user talk page DYK notices, says "over 5,000". I think we need to adjust all these pages and make them say either "at least" or "over/more than". —Bruce1ee^talk 07:18, 16 October 2011 (UTC)[reply]

I'd stick with "at least" if the tool rounds down, "over / more than" if it rounds up. If my math is not mistaken, knowing which way the tool rounds will help us guarantee a firm limit. Crisco 1492 (talk) 14:31, 17 October 2011 (UTC)[reply]

Raise the bar?

Is it time to raise the bar? There are lots of articles reaching the inclusion criteria now and, while that's a good thing, perhaps the list is getting too long. Perhaps we should have 25,000+ for lead items and 15,000+ for non-lead. violet/riga _^[talk] 13:32, 23 October 2011 (UTC)[reply]

I think the bar should be raised. A suggestion was also made earlier this year here, but no decision was made. —Bruce1ee^talk 15:00, 23 October 2011 (UTC)[reply]

want to add Android Lawn Statues

I would like to add Android lawn statues to the list, but i am not sure if the second days views count towards the DYK numbers and i also do not know how many hours it was up for, so i cant do the "per hour" section. any help would be great.--Found5dollar (talk) 05:03, 13 November 2011 (UTC)[reply]

It was on the Main page on November 10 and came off at midnight (UTC), so as per rule 1 the next day's views do count in this case (20.4k). Currently DYK is running on a 12-hour cycle, so the views/hour would be 1,700. Feel free to add it yourself, otherwise I'll add it myself when a get I chance. —Bruce1ee^talk 07:02, 13 November 2011 (UTC)[reply]

i did it , thanks!--Found5dollar (talk) 14:33, 13 November 2011 (UTC)[reply]

Do I do this?

I seem to recall that my last few 5k+ articles showed up here without my intervention, so I assumed someone had BOTted it. But Elan SCX didn't, so now I'm re-evaluating this belief. Should I be adding? Maury Markowitz (talk) 19:31, 27 February 2012 (UTC)[reply]

Sometimes other editors add them when they see them, but sometimes articles are missed. If one qualifies and isn't already listed, feel free to add it in its place. GRAPPLE X 19:35, 27 February 2012 (UTC)[reply]

Got it. Maury Markowitz (talk) 22:41, 27 February 2012 (UTC)[reply]

I've added it here, but as Grapple X says, if it's missing, feel free to add it yourself. I generally update this page every day, but often, because of time contraints, I don't always check every hook. —Bruce1ee^talk 08:36, 28 February 2012 (UTC)[reply]

2nd day viewings

Looking over stats for a few recent articles, I find that there are generally a big number of views on the 2nd day. I understand that articles can span days, but I'm not sure this is really what's happening, perhaps people are using the "archive" link? Any ideas on a way to tell one way or the other? Maury Markowitz (talk) 18:00, 28 February 2012 (UTC)[reply]

Sometimes it's a case of articles spanning two days on the main page; often it happens when articles leave the main page at 00:00 (UTC) and the view stats overflow into the next day (see DYK Stats rule 1); and I'm sure high next-day views also come from people accessing the archives. I don't know of anyway to separate "archive-link" viewing from the first two. —Bruce1ee^talk 05:43, 29 February 2012 (UTC)[reply]

Most viewed DYKs

I will propose creating a list of editors by the numbers of DYK hooks considered the most viewed DYKs. --SupernovaExplosion ^Talk 04:16, 25 March 2012 (UTC)[reply]

Are you referring to the "All-time DYK page view leaders" (top lead and non-lead hooks), or would you also include the monthly over 5,000 view hooks? Remember the monthly hooks don't show the editor's names. Would you append this list to WP:DYKSTATS, or create a new page? —Bruce1ee^talk 07:56, 25 March 2012 (UTC)[reply]

I mean the "All-time DYK page view leaders". It can be added in this page since it is all about stat. Thoughts? --SupernovaExplosion ^Talk 08:21, 25 March 2012 (UTC)[reply]

I think a list would be fine, but perhaps it would be better to add it to Wikipedia:Did you know/Hall of Fame. —Bruce1ee^talk 08:36, 25 March 2012 (UTC)[reply]

I'm not sure, but HOF is for total edits. Anyway are you interested to go ahead to create the list? --SupernovaExplosion ^Talk 16:07, 25 March 2012 (UTC)[reply]

I'm not interested at the moment, but you're welcome to create one yourself. I'd suggest waiting a few more days to see if anyone else has any comments to make. —Bruce1ee^talk 05:22, 26 March 2012 (UTC)[reply]

I'm not convinced that such a list is needed. This page should be focussed on the success of DYK in generating interest in newly-created articles, not a competition about who can get the most most-viewed DYKs. Perhaps a better idea would be just to add an extra column headed "No. of most-viewed DYKs" to the tables of DYK creators and DYK nominators at WP:DYKLIST, and as the tables are sortable you could see who was top of the table (Materialscientist I suspect) by clicking on the column header. This also has the advantage of distinguishing between article creators and nominators, which the most viewed DYKs tables on this page do not (the column heading is "Nominator" only). BabelStone (talk) 20:31, 29 March 2012 (UTC)[reply]

I think such a list would fit in at HOF. However, would the administrative effort be worth the trouble? I would like to see a list of total DYK views per editor including very small results, all added up, but who will do that kind of archival work to unearth the more than 10,000 DYKs from the top 10 percent of editors, the most likely ones? Binksternet (talk) 21:42, 29 March 2012 (UTC)[reply]

Doubt

An article I nominated for DYK, Untitled Tom Cruise project, got over 5,000 views. You can check here. However, I have this doubt: I did not notice at what time the DYK came on the Main page but per my notification, the DYK appeared on March 24, 2012. The 5,000 views came on March 25, 2012. Please advise as to whether the article is eligible to enter DYKSTATS. There is a rule saying that DYK's appearing near 00:00 UTC have some sort of exception, so I'd like to clarify. Thanks. ~*~Ankit Bhatt~*~ 09:23, 27 March 2012 (UTC)[reply]

It is already listed at DYKSTATS with 8,370 views. It hit the Main page here on March 24 at 19:27 (UTC) and came off here on March 25 at 03:12 (UTC). So the page views of both days are added: 3,137 + 5,233 = 8,370. —Bruce1ee^talk 05:18, 30 March 2012 (UTC)[reply]

Oh, thanks a lot for the clarification :) This is my first article that hit the DYKSTATS actually, so I was wondering ... ~*~Ankit Bhatt~*~ 07:23, 31 March 2012 (UTC)[reply]

Prostitution among animals

This should be listed in Wikipedia:Did_you_know/Statistics#April_2012. --SupernovaExplosion ^Talk 04:45, 5 April 2012 (UTC)[reply]

I've added it (24,651 hits). BTW anyone can add to this list – I generally update it every day, but often, because of time contraints, I don't always check every hook. —Bruce1ee^talk 05:35, 5 April 2012 (UTC)[reply]

Page views

I see in the instructions that the stats are suppose to measure the DYK effect. I think we should have two columns: One for the raw number and one for the estimated net number.--TonyTheTiger (T/C/BIO/WP:CHICAGO/WP:FOUR) 03:33, 9 April 2012 (UTC)[reply]

Simpler views per day

Hi. Can someone here kindly explain how to list this article by number of views per day (April 7)? The rules say "Some articles draw additional views unrelated to DYK. Those views are removed by subtracting a half sum on the days before and after, that is V(X) – (V(X+1)+V(X-1))/2. Example [1] V(X)=18,400; V(X+1)=4,400; V(X-1)=2,000; total=18,400-(4,400+2,000)/2=15,200." but I think these rules were intended for a completely different set of articles. Thank you. -SusanLesch (talk) 15:12, 9 April 2012 (UTC)[reply]

In this case the "additional views unrelated to DYK" are negligible and can be ignored. Because this article came off the main page at 00:00 (UTC), the next day's view are added (see DYKSTATS rule 1). —Bruce1ee^talk 15:30, 9 April 2012 (UTC)[reply]

Bruce1ee, thank you for correcting mine. I count a different total both before and after your correction. If I ever list again I will wait for somebody else to do the tally. -SusanLesch (talk) 15:43, 9 April 2012 (UTC)[reply]

Don't let me discourage you from updating this page, but if you'd prefer to wait for someone else to do it, that's ok. Just shout if you see a hook that is missed. —Bruce1ee^talk 06:04, 10 April 2012 (UTC)[reply]

I just tried to update (Edda Göring, not "my" article). The formula was given, thank you, but seems unfair, before 190 - after 1060: obviously much more than half of the second day was not normal traffic but related to DYK. I also wonder about the sorting, should that be by hits/time? --Gerda Arendt (talk) 07:08, 29 December 2012 (UTC)[reply]

I used the combination of 1st and 2nd; I omitted separate stats of 1st and 2nd due to space. --George Ho (talk) 07:25, 29 December 2012 (UTC)[reply]

I don't see it, --Gerda Arendt (talk) 07:28, 29 December 2012 (UTC)[reply]

8,368 = 7308 + 1060 , unless inadequately explained. Using "(7308 + 1060) - (? + ?)/2 = ?" looks awkward. --George Ho (talk) 07:35, 29 December 2012 (UTC)[reply]

More precise: I can follow your calculation, but I don't see it on the page ;) --Gerda Arendt (talk) 07:43, 29 December 2012 (UTC)[reply]

I guess casual viewers do not see it this way nowadays. They love precise stats that are too spacey, yet simplicity overcomes complications. --George Ho (talk) 08:01, 29 December 2012 (UTC)[reply]

Thank you, I will do the same next time, --Gerda Arendt (talk) 08:43, 29 December 2012 (UTC)[reply]

Nominators' names

Why the names of the editors are removed? --SupernovaExplosion ^Talk 04:55, 10 April 2012 (UTC)[reply]

See Wikipedia talk:Did you know#DYK date request. —Bruce1ee^talk 05:56, 10 April 2012 (UTC)[reply]

Humor Styles

Hi all, I'm kind of new to this so I was just hoping someone could fill me in. I created a page on humor styles that received 5,069 views on April 14th when it appeared in the DYK column (just barely made the cut). How do I go about adding it to this to the DYKSTATS list? Is this something I should add myself? Thanks in advance for the help. Majobc11 (talk) 04:11, 17 April 2012 (UTC)[reply]

I've added it. If you see it's not there, feel free to add it yourself – have a look at my edit here. —Bruce1ee^talk 04:54, 17 April 2012 (UTC)[reply]

Great thanks - I appreciate it! In the future I'll add it myself. Majobc11 (talk) 20:16, 17 April 2012 (UTC)[reply]

The Sugar Girls

Can someone explain how to add a page to these stats? A page I created, The Sugar Girls, received by my calculation 5290 hits from DYK. (4763+573 minus half of 46 and 47 for views either side.) I'm not clear how to add it here. Dwab3 (talk) 14:43, 19 April 2012 (UTC)[reply]

I've added it. I ignored the before and after views because they are negligible. —Bruce1ee^talk 06:08, 20 April 2012 (UTC)[reply]

All-time DYK page view leaders (top 50)

Why don't we make both All-time list top 50 lists so that they prune themselves?--TonyTheTiger (T/C/BIO/WP:CHICAGO/WP:FOUR) 05:52, 20 April 2012 (UTC)[reply]

I think that would be a good idea. —Bruce1ee^talk 06:11, 20 April 2012 (UTC)[reply]

Is anyone watching this page. I still think this is a good idea even though I have an article in 53rd of 58 with pictures (Cat Daddy) and 42nd of 79 without pictures (Disappearing Model).--TonyTheTiger (T/C/BIO/WP:CHICAGO/WP:FOUR) 02:23, 30 May 2012 (UTC)[reply]

I'm not sure how many DYK-regulars check this page. Perhaps you should also raise it at WT:DYK. —Bruce1ee^talk 04:48, 30 May 2012 (UTC)[reply]

Support 50. — Crisco 1492 (talk) 02:14, 1 June 2012 (UTC)[reply]
Oppose for two reasons. First reason is that the DYKstats list should not be treated as a competition. This focus on enhancing the exclusivity of the DYKSTATS list is an unpleasant reminder that certain users are motivated more by competition and trophy-collecting than by considerations related to the positive purposes of DYK. I don't like to encourage "look at me, I am the greatest"-type behavior. The second reason is related to the practicalities of maintenance -- it's challenging enough to keep the table entries in proper order without worrying about trimming the last item off the list whenever a new one is added, not to mention mediating between users who feel that "their" personal trophies were prematurely removed from the trophy case. This proposal is a bad idea; it is contrary to the purpose of continuing DKYSTATS as a laid-back list of examples of effective DYKs. --Orlady (talk) 04:05, 1 June 2012 (UTC)[reply]
Oppose per Orlady. I really agree that instead of serving as some sort of leaderboard, the DYKSTATS section showcase all hooks which have reached a certain threshold. I've never seen it as a means of competition, but instead as sort of an interesting statistical look at which type of hooks normally succeed at receiving views more than others. Nomader (talk) 15:35, 1 June 2012 (UTC)[reply]
Oppose per Orlady. Sasata (talk) 16:00, 1 June 2012 (UTC)[reply]
Oppose sorry Tony, but I agree with Orlady who sums it up well. Casliber (talk · contribs) 21:13, 1 June 2012 (UTC)[reply]
Oppose per Oraldy. Cbl62 (talk) 02:54, 13 June 2012 (UTC)[reply]

Splitting up statistics into transclusions

Why not splitting up All-timers into {{WP:did you know/Statistics/Lead hooks}}, {{WP:did you know/Statistics/Non-lead hooks}}, and {{WP:did you know/Statistics/June 2012}}? That way, it is easier to navigate and edit comfortably. --George Ho (talk) 22:59, 4 June 2012 (UTC)[reply]

I don't see a strong need for a split. At ~68K, the page isn't all that long. --Orlady (talk) 02:44, 13 June 2012 (UTC)[reply]

Oh, no, no, no, no, no.... I mean, you can split one section into page and transclude it, as WP:RM is transcluding it. Understand? --George Ho (talk) 04:52, 13 June 2012 (UTC)[reply]

No. I fail to see the point. If it isn't broken, Don't fix it. The C of E. God Save The Queen! (talk) 17:31, 25 June 2012 (UTC)[reply]

Would this qualify or not?

What, if the phrasing of the hook makes a non-target article the main centre of attention? For example, in the hook below the target article, although bolded and easily distinguishable, leaves the impression of being less relevant. The hook was very successful in generating hits to a non-target.

08:42, 25 June 2012 (UTC)

... that the NOT Tower (pictured) in Kraków was nicknamed Skeletor after it had been abandoned by the Polish Federation of Engineering Associations and left unfinished for 30 years?

Polish Federation of Engineering Associations has been viewed only 1,128 on June 25, 2012. [4]
Meanwhile, the NOT Tower has been viewed 9,511 on June 25, 2012, from the average of 50 per day. [5]
Skeletor has been viewed 3,920 on June 25, 2012. Almost six times more than the average. [6]

Should we be proposing hooks that don't steal attention from our targets? Poeticbent talk 03:15, 26 June 2012 (UTC)[reply]

This might be particular to this hook, but generally, yes, the hook should have as its subject its subject.VolunteerMarek 04:08, 26 June 2012 (UTC)[reply]

If I knew that this was going to happen, I would have spend a lot more time expanding the NOT Tower entry using the content of my new DYK article with inline citations, or at least... exclude the internal links from the DYK hook in order to force the reader to start reading at the target article and move from there to all non-target entries, not from the Main page. Poeticbent talk 03:17, 27 June 2012 (UTC)[reply]

This just happened with Peignoir which got over 5000 hits whereas the focus article, Babani, only got about 2000 or so. Whoohoo for ladies' negligees.... Mabalu (talk) 04:34, 28 November 2012 (UTC)[reply]

When does "DYK page view leaders by month (over 5,000 views)" get updated?

There are a few recent ones that qualify but have not yet been listed in the table. Roger (talk) 08:56, 10 July 2012 (UTC)[reply]

I used to update this page regularly, but it became a bit too time-consuming for me. Anyone is free to add qualifying entries. —Bruce1ee^talk 09:30, 10 July 2012 (UTC)[reply]

If any of my own entries qualify, I usually post them. It's updated whenever people put in the effort. Chris857 (talk) 15:03, 10 July 2012 (UTC)[reply]

It's basically up to people who nominate DYKs to watch how they do and update the stats themselves if necessary. BabelStone (talk) 00:49, 11 July 2012 (UTC)[reply]

Miami zombie is qulified, but it had three different names that day (i.e., was moved twice) and I joust haven't got a round tooit. μηδείς (talk) 01:11, 11 July 2012 (UTC)[reply]

October Baby

Based on the main page appearance on 3/26/12 and 3/27/12 and views I calculate 7400. Right? – Lionel ^(talk) 08:15, 25 July 2012 (UTC)[reply]

Queer views at First Motion Picture Unit

FMPU, which is lucky to get 50 views on any given day was on the main page on 7/1/12 and got 2538 views. But then for no apparent reason it gets 4467 views a week later on 7/7/12. Anyone know what happened? – Lionel ^(talk) 08:28, 25 July 2012 (UTC)[reply]

Could have been linked to by an external site, some article get a lot of one-day traffic if a popular site throws up a link to them at some point. Project A119 saw a spike (relatively huge increase, but still only small numbers) when this Cracked.com article linked to it, for example. Could easily be something of a similar nature. GRAPPLE X 21:59, 3 August 2012 (UTC)[reply]

Anaheim police shooting

Anaheim_police_shooting_and_protests got 10k+ hits in July [7]. Please add it.VolunteerMarek 23:01, 2 August 2012 (UTC)[reply]

Only the views for the day on the Main page are counted, which in this case (July 29) is 1,916. Sorry. —Bruce1ee^talk 04:51, 3 August 2012 (UTC)[reply]

Ok, but wait. So what you're saying is that the article got more views while nominated at DYK than while it was featured at DYK?VolunteerMarek 20:16, 3 August 2012 (UTC)[reply]

Yes, it got more hits earlier because it was a hot news story then. DYK/statistics measures only the hits that apparently resulted from an article's appearance in DYK. --Orlady (talk) 20:32, 3 August 2012 (UTC)[reply]

Philip Hindes listing needs an asterisk

The DYKSTATS listing for Philip Hindes needs some sort of an asterisk on it. Since he won an Olympic gold medal -- and generated some controversy in the process -- on the same day that the hook appeared in DYK, only a fraction of the 59,000 hits that day can be attributed to DYK. The article creator deserves kudos for creating an article about somebody who went on to win a gold medal and DYK should be proud of featuring it, but we can't claim that the DYK hook was absurdly effective. --Orlady (talk) 14:52, 5 August 2012 (UTC)[reply]

A well timed DYK, I say. Hawkeye7 (talk) 19:41, 8 September 2012 (UTC)[reply]

U.S. Presidents by net worth

This article needs to be on this page, since it got more than 18,000 views in August. Futurist110 (talk) 01:31, 1 September 2012 (UTC)[reply]

I take it you're referring to List of United States Presidents by net worth. I've added it, but it got 15,789 views (10,613 + 5,176) – only the views while on the Main page count (August 29 + the next overflow day). —Bruce1ee^talk 07:16, 1 September 2012 (UTC)[reply]

Traffic statistics hosted by stats.grok.se abandoned as of September 3, 2012

Does anybody know if this invaluable and highly inspirational service is ever going to come back? Poeticbent talk 18:32, 8 September 2012 (UTC)[reply]

I don't think it's been abandoned. From time to time it has stopped updating for a few days, but it has always come back again. See here and here above. —Bruce1ee^talk 07:46, 9 September 2012 (UTC)[reply]

We'll see, but I was also wondering. With all the amazing infrastructure at our disposal, why can't we host our own stats for the DYK column directry on wiki? After all, we already host numerous other data collection points such as

{{NUMBEROFARTICLES}} → 6,915,004
{{NUMBEROFEDITS:R}} → 1254445816
{{NUMBEROFPAGES:R}} → 61896831
{{NUMBEROFPAGES}} → 61,896,831 and so on.

Such a bot, if possible at all, could for example collect the Wikipedia front page data for all DYKs, by their given placement, rather than individual article's monthly views... or something similar for a quick and easy access. I'm not a programmer though, so I have no way of knowning if this service would ever be possible. Poeticbent talk 14:07, 9 September 2012 (UTC)[reply]

Well it's working again. Regarding taking over this service, similar sentiments have been expressed at User talk:Henrik, the author of this tool. —Bruce1ee^talk 06:43, 11 September 2012 (UTC)[reply]

Hydrophis donaldi

Why did Hydrophis donaldi not make the list? It received over 7,000 views. Abyssal (talk) 23:07, 1 October 2012 (UTC)[reply]

There is no official system for updating this page, it happens as people put in the effort. Chris857 (talk) 23:41, 1 October 2012 (UTC)[reply]

~~==ERROR on the DYK Recent Additions archive page==~~

Any DYK reported on 0:00 of a day is listed under the incorrect date in Wikipedia:Recent_additions. For instance, Hubur is listed as June 17, but on both the talk:Hubur page and the DYKSTATS page, it's June 16. Same with Aspen Community Church, Talk:William_P._Cronan or, well, any Midnight DYK. The Steve 03:12, 8 February 2013 (UTC) Never mind, that's when it was removed from the main page. I was confused, because the link in the DYK template sent me to the wrong date...[reply]

Help with 54-article hook

I think I don't have enough time to make drafts of inserting the hook into the stats. The hook relates to palentology and fossils in the United States. --George Ho (talk) 03:31, 12 February 2013 (UTC)[reply]

Fish soup bee hoon

I'm not sure how it's checked. Does fish soup bee hoon pass? ☯ Bonkers The Clown \(^_^)/ Nonsensical Babble ☯ 13:49, 13 June 2013 (UTC)[reply]

No, it is under 5,000 views. It is not qualified to be part of the charts. --George Ho (talk) 15:54, 13 June 2013 (UTC)[reply]

Potato production in North Korea definitely passes. Can someone add it in please? ☯ Bonkers The Clown \(^_^)/ Nonsensical Babble ☯ 13:48, 14 June 2013 (UTC)[reply]

Dam

Shouldn't be Dam in the list? --Tito☸Dutta 23:47, 23 June 2013 (UTC)[reply]

Question

I have noticed that several articles that has been mentioned at the DYK section has reached the threshold of 5000 views but is not mentioned at the DYKstats page. I have fixed those that I have found but as most articles mentioned at DYK does not reach the views needed wouldn't it be nice to have some better way of keeping track of those few articles that does so they can be mentioned at DYKstats?.--BabbaQ (talk) 14:41, 12 July 2013 (UTC)[reply]

Make a template for calculating hourly and total hits?

My impression is that for DYK hooks logged as highly viewed (>5000 hits), the logging of hourly rates and even total hits is error prone. People don't notice the instructions or else they don't take the time to do all the calculations, which can be slightly tedious - particularly when the number of hours is shifting around and nonstandardized. I noticed 4 hooks in August 2013 (including, initially, my own) that had done incorrect hourly calculations. Some also failed to subtract off the estimate of background hit rate. It strikes me that one way to improve and standardize the process might be to create a template that does the calculations. Assuming that Wikipedia templates are capable of doing calculations, the appropriate entries might be:

1=hits on day when appeared on main page,

2=hits day before,

3=hits day after,

4=hour posted on main page (translated into 24-hour decimal, e.g., 1:15pm -> 13.25; extract this info from DYK archive),

5=hour removed from main page (if next day at 1:15am, could designate as 25.25)

The data extraction would still be mildly tedious but a template would remove the need to do calculations, and it would also make it easier to check if the data was complete and was being properly deployed on the DYK stats page. But since I don't know how to build such templates in Wikipedia, this is just a suggestion, based on what appears, at least at first glance, to be the high prevalence of errors in the DYK stats. -- Presearch (talk) 17:06, 17 August 2013 (UTC)[reply]

I agree with Presearch. Also, shouldn't the time DYKs are actually featured on the main page be standardized to reflect a minimum of 12 hrs? An article with 8 hours of exposure VS an article with 12 hours would typically have fewer views if both were featured using the same time frame. Atsme☯^Consult 22:20, 23 July 2014 (UTC)[reply]

I've often thought of doing a statistical study of factors affecting DYK click-through, but dreaded gathering the raw data. The above specification makes too many assumptions e.g. that the hook appears no more than 24 hours. If I could have my wish here's what I'd want:

A vector of the dates representing the N days up to the hook's appearance, and M days after (you'd have to wait M days, obviously, for this to be fully populated)
A vector of the click counts for each of these days
The date-and-time the hook first appeared, and date-and-time it disappeared.
Slot #, with or w/o image, etc.
Text of the hook.

The reason we want post-appearance data is that it I believe there are may be interesting effects if the hook is picked up by e.g. a popular blog. And to really do this right we should have the same data for everything linked from the hook, not just the "primary" article. BTW, I suspect hourly (or better) click counts are actually available somewhere if we just knew where to look, and that would sharpen the data tremendously, for obvious reasons. Finally, redirects muddle everything and we should be sure that's handled properly. EEng (talk) 22:46, 23 July 2014 (UTC) In the foregoing, "we" means someone not me, since I have no idea how to go about getting all this. But I'd do a great job on the analysis, if I had the data![reply]

P.S. I see a good deal of discussion here about whether DTK "championship" records need to be adjusted for various reasons, wondering whether it's "fair" that this or that condition may have affected the counts. Is getting on some scoreboard really what people are interesting in? I've had several DYKs over the years that I think were over the cutoff, and I don't think I even put them on the board. I'm much more interested in figuring out what makes people find a hook interesting enough to click on. EEng (talk) 22:53, 23 July 2014 (UTC)[reply]

EEng, I agree with much of what you've stated, but don't quite understand what you mean by we should have the same data for everything linked from the hook, not just the "primary" article. I think the hook to a specific article is important as long as the hook itself does not delve into "sensationalism", and remains within the realm of factuality. I don't think one hook/one article should equate into stats on numerous articles as each article will have its own stats. Equal timing is of the essence, therefore I would not object to an established 24 hr. in/out time for each DYK feature, even if it means expanding the DYK section on the main page to accommodate overflow. JMO - Atsme☯^Consult 03:30, 24 July 2014 (UTC)[reply]

I don't know what you mean about sensationalism and so on. Just to be clear, let's suppose we have this DYK:

Did you know ... that George Washington was a farmer and general before he was president of the US?

Here we might call George Washington the "primary" link and the others "secondary". It is commonly thought, and my limited experienced seems to confirm, that secondary links draw clicks away from the primary, and any investigation would need to look into that.

As to the length of time a hook should be on MP, that's out of our hands -- subject to change for all kinds of intentional and accidental reasons. The way we gather stats should give the maximum reasonable amount of information in a way that can be sensibly interpreted under the fewest assumptions (such as any assumption about how long hooks will be on MP). EEng (talk) 03:58, 24 July 2014 (UTC)[reply]

May I note that, as per here, I think the times are when the set comes off the main page, so what Atsme is reporting as errors are not. Thanks, Mat ty.007 06:56, 24 July 2014 (UTC)[reply]

An American Hippie in Israel

Much appreciated if someone were to help me include this on this month's list. ☯ Bonkers The Clown \(^_^)/ Nonsensical Babble ☯ 13:52, 20 August 2013 (UTC)[reply]

Pneumatic tube mail in New York City

I believe that my hook from yesterday passes - the stats page lists the pageviews for the day at 6078 [8]. Is that correct and the final number? Witty lama 01:54, 22 October 2013 (UTC)[reply]

Addition

Hi, did I add things correctly in this edit? Thanks, Mat ty.007 19:42, 26 November 2013 (UTC)[reply]

Something wrong with page view stats?

Since the new year, counts dropped on most pages, either zero or some low counts. Z22 (talk) 19:35, 6 January 2014 (UTC)[reply]

Yes, see the bottom few sections at User talk:Henrik. Mat ty.007 19:44, 6 January 2014 (UTC)[reply]

Pageview stats off

I have started up a thread about how all the page view stats being used for WP:DYKSTATS, WP:TFLSTATS and WP:TFASTATS are using 23:00 to 22:59 (UTC) rather than 00:00 to 23:59 (UTC) as the day at Wikipedia:Village_pump_(technical)#wikiviewstats_vs._stats.grok.se.2Fen_time_ranges_.2823:00_to_22:59_.28UTC.29_rather_than_00:00_to_23:59_.28UTC.29.29.--TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 19:18, 10 August 2014 (UTC)[reply]

Thanks for helping out

Thanks to all the discussants at Template:Did you know nominations/Dylan Penn, (especially EEng, HelenOnline, Martinevans123, BlueMoonset, Bali88, Jakec, and AndyTheGrump) who put up with all my pushing and pulling to help us get to final hook that seemed to entice the readers enough to make it a highly viewed page.--TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 01:16, 11 September 2014 (UTC)[reply]

Anyone named after a rabbit gets my vote. Martinevans123 (talk) 07:24, 11 September 2014 (UTC)[reply]

Infant swimming

I'm a bit unclear as to how to check the exact times infant swimming was on the Main Page, and how to split the views between them. There are over 300,000, so it might or might not be an all-time #1 depending on this. Oreo Priest ^talk 17:23, 25 September 2014 (UTC)[reply]

What happened to wikiviewstats

It seems like https://tools.wmflabs.org/wikiviewstats/? is no longer working. Does anyone know what is going on with it?--TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 20:39, 8 October 2014 (UTC)[reply]

Some suggestions

At ~126.8k bytes (>50k over the suggested 75k max from WP:TALKCOND), this talk page needs an archive.
I have noticed the same thing as Presearch from the above thread, #Make a template for calculating hourly and total hits?. Nearly all of the October 2014 DYK stats are off. Mostly the errors represent a failure to subtract off the background noise (i.e. V(X) rather than V(X)-(V(X-1)+V(X+1))/2), but in some cases this is compounded by the user simply adding the totals from days X and X+1 (i.e. V(X)+V(X+1) rather than V(X)-(V(X-1)+V(X+1))/2). In looking into it more closely I believe this is the result of the DYK notification template which states "(here's how, live views, daily totals)". Of these three links, the middle one ("live views") is broken and the first one ("here's how") links to User:Rjanag/Pageview stats, a page whose instructions differ dramatically from Wikipedia:Did you know/Statistics#Rules. In fact it is the User:Rjanag subpage that recommends using V(X)+V(X+1) in step 6. I recommend that the DYK template be updated to delink the middle ("live views") link and that the first ("here's how") link be relinked to here. Does that make sense?
I considered going back and revising the DYK calculations for October, but I didn't want to step on any toes and I didn't know how far back I would have to check. Have the stats ever been checked for accuracy in the past? I recommend that some of the regular page watchers here should make an effort to go through them all and then post the corrections here for oversight by others. This can be done all at once, year by year, or month by month.

Please share your opinion on any of these suggestions. -Thibbs (talk) 23:06, 16 October 2014 (UTC)[reply]

@Rjanag: - I should have pinged you earlier. Do you have any opinion on the matter? Because the other option for suggestion #2 if nobody is interested in linking to Wikipedia:Did you know/Statistics#Rules is to update User:Rjanag/Pageview stats to match the official rules. And then you'd probably want to watchlist "Wikipedia:Did you know/Statistics" so that you'd notice if there were any more changes to the official rules. What do you think? -Thibbs (talk) 12:33, 20 October 2014 (UTC)[reply]

I'm not up to speed anymore on what the standard practices are for DYK views; I made that page as a how-to guide for people who don't really know any technical stuff, rather than as a way to prescribe rules. So if the instructions there don't match what is currently standard practice, I certainly don't mind if someone wants to update that to explain how people can get these counts in whatever way is appropriate for how the rules are now. rʨanaɢ (talk) 18:57, 20 October 2014 (UTC)[reply]

You mean third party updates to the User:Rjanag/Pageview subpage? I hadn't thought about that. It's probably better than nothing. Are we mostly linking to your user subpage for historical reasons, though, or are there additional reasons? I'll ping Shubinator for this discussion as well since he operates DYKUpdateBot (the bot responsible for placing the DYK notification templates). Any thoughts on how best to link DYK participants to an accurate version of the official rules for calculation? -Thibbs (talk) 13:01, 21 October 2014 (UTC)[reply]

The bot uses {{DYKmake/DYKmakecredit}} and {{DYKnom/DYKnomcredit}}, so once those are updated, the bot will automatically pick up the latest notification templates. Shubinator (talk) 01:09, 22 October 2014 (UTC)[reply]

Thanks Shubinator. I've just updated the User:Rjanag/Pageview subpage and due to the general lack of interest here I will cross-posted my question regarding switching to the official rules at the village pump tomorrow. -Thibbs (talk) 15:23, 22 October 2014 (UTC)[reply]

Just an update here: Nobody at the village pump had any interest in this proposal and since there seems to be no interest here either I will just give up on the suggestion. Please try to keep in mind that User:Rjanag/Pageview stats needs to be updated in parallel with the rules here or else we'll see the problem arise again. In fact I'll go ahead and add a note to the rules section to remind future updaters. -Thibbs (talk) 17:57, 9 November 2014 (UTC)[reply]

Relevance of DYK stats

I like DYK stats and find it to be an interesting and funny way to see what articles gets the most views. But I think however that the relevance of the DYK stats lists becomes low when considering that almost half or more of the articles that reaches the 5,000 views threshold are never added to the list. I am not sure if there are a similar "problem" with the DYK stat lists with higher view counts above. Is it possible to create a bot who can detect this and add it to the DYK stats? Because if not all or atleast a huge majority of the articles that makes the threshold are added I really personally find it to be a irrelevant page. This needs more input. Cheers.--BabbaQ (talk) 11:42, 26 October 2014 (UTC)[reply]

BabbaQ, you don't seem to be netting pageviews. Are you sure that many are missing? Can you adjust your statistics and remove the articles that don't meet the threshold after netting?--TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 12:32, 26 October 2014 (UTC)[reply]

BabbaQ, come in.--TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 02:50, 28 October 2014 (UTC)[reply]

Netting pageviews

Why are the majority of the statistics being added without netting pageviews. Do we care about this anymore?--TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 12:29, 26 October 2014 (UTC)[reply]

Adding to list

I was pleasantly surprised to notice that a DYK that I proposed, Alvin "Shipwreck" Kelly, received in excess of 11,000 pageviews on Dec. 29. This is the first time any DYK I've written has ever gotten such readership. I went over to the DYKSTATS page to add that, and I immediately observed a couple of things that confounded me. First, there had been no DYKs added to the page since October. Why is that? Secondly, part of the chart calls for a "views per hour" listing and I can't find any way to determine how long my item was on the main page. Can someone please clarify this? Thanks. Coretheapple (talk) 17:10, 1 January 2015 (UTC)[reply]

Never mind, I located the repository of the duration data and have added the late Mr. Kelly to the stats page. Coretheapple (talk) 00:18, 2 January 2015 (UTC)[reply]

Also I was looking at the archives page - there are many listings for November and December elsewhere. duh.... Coretheapple (talk) 17:21, 2 January 2015 (UTC)[reply]

Violet (opera)

I have submitted many DYKs in the past with views ranging from a couple of thousand to over ten thousand, but I've never yet see a result of zero, as for Violet. Shurely shome mishtake?--Smerus (talk) 07:12, 6 February 2015 (UTC)[reply]

I have seen it, but for a whole day, not a single article, --Gerda Arendt (talk) 07:52, 6 February 2015 (UTC)-[reply]

I'm having the same problem with Jill Valentine, which was featured at DYK of 5 February. It was my 31st article featured at DYK, but is the only one of my DYKs that supposedly has no page views the day is was featured. My 32nd nomination, which was featured at DYK later that same day, has over 5,000 page views. I'm guessing the counter stopped working for most of the 5th? Freikorp (talk) 07:51, 8 February 2015 (UTC)[reply]

Looks like it actually stopped working for a few days, appears to be back online now. Freikorp (talk) 12:43, 12 February 2015 (UTC)[reply]

DYK views

The DYK views counter has stopped for 4 February and no further days can be seen. Just an example, [9].--BabbaQ (talk) 19:16, 11 February 2015 (UTC)[reply]

Yep, it appears to be back online as of the 11th but no pages have any views between the 5th and the 10th. A real shame, since I had three DYKs during that time. :( Freikorp (talk) 12:42, 12 February 2015 (UTC)[reply]

How to add entry

Not sure how to add an entry here...is it populated automatically. I am looking to add Dyslexia 6000 views -- Moxy (talk) 17:31, 7 April 2015 (UTC)[reply]

Moxy I don't believe that entry would qualify. You take the day's count and subtract the previous day's. In this case, 6,334 - 2103 = 4,231 which would fall below the 5,000 requirement. Mkdw ^talk 18:36, 5 June 2015 (UTC)[reply]

New Count Tool

So; I debuted this at the the DTYK talk page and it seemed to go down well. Using the new WMF Page Visits data I've created a DYKSTATS count tool. Here is an example of it in action. Does anyone object to me adding it to the DYKSTATS page in the intro as an alternative to the existing tool? Benefits are: it does the calculation for you, leading to consistency in results AND it covers all page views (including mobile) which so far looks to increase view counts by about 10%. The drawback is that the new API only has data from Oct 2015 onwards. --Errant ^(chat!) 14:03, 17 November 2015 (UTC)[reply]

The count tool is down since January?

Seems like the count tool has been down since the middle of January and so far for the whole of February. See [10] [11] for example. It seems that an alternative exists here, but not sure how we can integrate that to this page. HaEr48 (talk) 18:31, 4 February 2016 (UTC)[reply]

@HaEr48: My stats calculator uses that API to calculate the scores. Feel free to use :) --Errant ^(chat!) 20:14, 4 February 2016 (UTC)[reply]

@ErrantX: I don't know how this page works in the first place, so let's wait for the next person :) HaEr48 (talk) 22:10, 4 February 2016 (UTC)[reply]

It mostly works by people manually adding their DYK'a. A few people have been using my tool so you can use that as a reference :) --Errant ^(chat!) 00:03, 5 February 2016 (UTC)[reply]

Ah I see. Initially I thought it was added by bot which parses the stats.grok.se tool. HaEr48 (talk) 01:49, 5 February 2016 (UTC)[reply]

That would be a good approach; if I get some time it might be possible for me to write something like that, just needs to load in lists of DYK's (not impossible, but sufficiently complex to make it more than a 10 minute job). --Errant ^(chat!) 08:11, 5 February 2016 (UTC)[reply]

@ErrantX: Any luck with the automatic approach. Anyway, if you haven't found one already, This (and its sibling pages) seems to be a good possible source to load all the DYKs? HaEr48 (talk)

I came here to say the same things as ErrantX. The new API also accounts for mobile page views and not just desktop, so I guess the DYK project will have to close the record book and create a new one entirely? Just know that moving forward you should not compare stats with what we have recorded in the past — MusikAnimal ^talk 07:55, 6 February 2016 (UTC)[reply]

Please see Wikipedia:Web statistics tool for a collection of questions and possible new solutions. --.js^{[democracy needed]} 14:57, 6 February 2016 (UTC)[reply]

If there is a new stats tool in use, then that obe should be the tool linked to in the DYK template placed at the talk pages of the articles appearing on DYK. If the stats tool is down forever I mean.--BabbaQ (talk) 19:33, 19 February 2016 (UTC)[reply]

BabbaQ, the new tool doesn't go back far enough, so it will affect older nominations. The issue is in discussion at Wikipedia talk:Did you know#Checking daily views. Jolly Ω Janner 19:46, 19 February 2016 (UTC)[reply]

Got over 5000, can I add it?

One of my creations, Imperator (horse) got 5485 views yesterday.[12] is it eligible to add? This is by far the most views I've ever had. White Arabian Filly ^Neigh 22:39, 12 April 2016 (UTC)[reply]

It looks very likely that it will meet the criteria for monthly inclusion. I would advise waiting at least 24 hours for the DYK stat counter to give a more accurate figure on its page views. Also note that it was displayed for 11.83333 hours unlike most hooks. Jolly Ω Janner 01:48, 13 April 2016 (UTC)[reply]

It still says it got over 5,200. How can I add it? White Arabian Filly ^Neigh 20:10, 14 April 2016 (UTC)[reply]

Got over 5000, what next?

Education of the British Royal Family has 6K [13]. Now what? LavaBaron (talk) 18:07, 20 April 2016 (UTC)[reply]

Add it to the monthly list. Just copy the syntax of the ones around it in edit mode and preview your edit to ensure you've done it right. The C of E God Save the Queen! (talk) 18:58, 20 April 2016 (UTC)[reply]

Do I get a medal or barnstar for this? LavaBaron (talk) 02:40, 21 April 2016 (UTC)[reply]

I can't tell if you're being sarcastic. Jolly Ω Janner 19:13, 21 April 2016 (UTC)[reply]

Problem with dykstat tool?

Operation Leader recently (15 June) ran as a DYK. I can't really make sense of the numbers I get from the Dykstats tool. Anyone got an idea why I get "-4216 Views"? Negative views is a bit surprising. Pageviews Analysis gives a much higher number. Manxruler (talk) 18:57, 17 June 2016 (UTC)[reply]

Bit late @Maxruler: but I purged the cache and the numbers look much healthier :) I suspect a bug, will look into it! --Errant ^(chat!) 13:39, 4 October 2016 (UTC)[reply]

Stuffing the Ballot Box

It appears the ballot box has been stuffed to combine views from two different articles that appeared in the same hook so that they, collectively, form the most popular non-lead hook on WP, specifically Blood Moon Prophecy and something else. Should Tiffany Trump actually be the most popular non-lead hook? Proposal: Because there are two articles involved, a divisor by two should be applied. LavaBaron (talk) 22:20, 8 August 2016 (UTC)[reply]

Support Tiff should be the most popular LavaBaron (talk) 22:20, 8 August 2016 (UTC)[reply]
- Tell me what's the purpose of this? Is it of a serious concern that the stats are being recorded incorrectly or that you just want to be able to claim that you wrote the most viewed non-lead hook? Not trying to be nasty or cynical in any way but it does seem that this wasn't an issue until you brought it up and you have been using your sig recently to make big claims about it. I think that it would be beneficial if this could be toned down a bit because it will only cause trouble later if this path continues to be trod. The C of E God Save the Queen! (talk) 22:46, 8 August 2016 (UTC)[reply]

Serious concern. Shouldn't our records be accurate? LavaBaron (talk) 23:12, 8 August 2016 (UTC)[reply]

OK. Thank you for clarifying. I'll wait and see what others think. The C of E God Save the Queen! (talk) 23:19, 8 August 2016 (UTC)[reply]

You're welcome. LavaBaron (talk) 23:49, 8 August 2016 (UTC)[reply]

Frankly I don't understand how half of these view statistics are calculated... some are two day, some are averages, some look like a high school algebra problem. To me, page views should be for the new/improved article for the day (00:00:00 - 23:59:59 UTC) it was on the main page. If there are two new/improved articles, each can have its own entry for each page's views, but just have the same hook. EvergreenFir (talk) 00:48, 9 August 2016 (UTC)[reply]

@EvergreenFir: That's the way it should work. In practice from time to time, queues aren't filled so sets overrun which puts the whole thing out of whack. Gradually it does get back to the 12-12 system but during that time you do have hooks that spill over to the next day. It's only fair to allow them time because a hook that's on for 6 hours a day should be given the same time to gain views for a hook that's on for 12. That's how the 2 day thing works. As for the algebra, I don't agree with it personally but I believe the theory is that it stops inflated views from popular articles where clicks have comes from places other than the DYK link. So it's sort of a filter to make the views more reasonable. The C of E God Save the Queen! (talk) 06:36, 9 August 2016 (UTC)[reply]

@The C of E: Thank you for the explanation. The multi-day part makes total sense from your explanation. I'll trust the DYK regulars know what they're doing with the algebra stuff. :) EvergreenFir (talk) 06:40, 9 August 2016 (UTC)[reply]

Comment It seems like there's a consensus to move Tiffany Trump into the number one spot. Last call for objections. LavaBaron (talk) 19:57, 2 November 2016 (UTC)[reply]
I'm confused; multi-article DYK's are common enough and we usually record the combined pageviews of them for the overall hook "score" --Errant ^(chat!) 08:23, 7 November 2016 (UTC)[reply]

That's a good point, Errant. I guess the question is, though we've done it in the past, should we? LavaBaron (talk) 21:29, 3 December 2016 (UTC)[reply]