Wikipedia:WikiProject Video games/Newsletter/20200405/Feature

Feature: High-Level Article trends in the past decade or two

Submitted by Thibbs

Well here we are completing the first quarter of 2020 and I don't know about you, but I feel like I'm getting more gaming time than ever before. Could it be related to the fact that just last week I had to file for unemployment? Could it be because I'm under some sort of government lockdown? Perhaps it's because by playing my heart out I'm finally able to point out to the rest of my family that stockpiling (collecting? hoarding?) video games and their respective consoles makes sense in light of the ever-present danger of zombie apocalypse global pandemic for which I've been preparing these many years? Maybe it's a combination of all of these reasons. Whatever the reason, this first quarter of the new decade has arrived bringing interesting times. And on that note, the WP:VG Newsletter sends out our best hopes and wishes for all who find themselves on the wrong side of the aforementioned interesting times. All levity aside, it is eerie to write an international English-language newsletter report with the understanding that social disruption, danger, and death have become concerns for all readers (even those in non-English areas). Freakin' zombies...

This quarter we will be casting a backward glance at WP:VG's high-level articles: our Featured content, our old A-class efforts, and our Good content. We look back at our efforts through the last decade from our first quarter of 2010 (1Q10) to our current quarter, and we also look back to the very inception of Wikipedia's bedrock "quality" and "importance" institutions nearly two decades ago. We will address historical moments in the two attending graphs, and we will discuss inflection points and their significance in the progress of our efforts. We will also be discussing previously generated graphs and recent discussions from other quarters. Of specific note, we will be relying on much of the information gathered by our primary NAA partner, PresN, without whose assistance the analysis for categories like the A-class would not have been possible. For both graphs that we present in this newsletter, the raw data has been drawn from previously published WP:VG Newsletters with the remainder drawn from WP:VG's Article statistics and counted by hand. Minor discrepancies may result in mismatches between the data presented here and WP:VG's Article statistics.

In the first graph we will restrict our discussion to Featured content. The x-axis uses quarterly tags where 1Q means first quarter (January 1 through March 31), 2Q means second quarter (April 1 through June 30), 3Q means third quarter (July 1 through September 30), and 4Q means fourth quarter (October 1 through December 31). The year associated with each quarter is denoted by the last two digits, so e.g. 4Q15 represents October 1, 2015 through December 31, 2015 and 4Q16 represents October 1, 2016 through December 31, 2016. The y-axis represents the total number of articles recognized by the Wikipedia community in each area (Articles, Lists, Pictures, etc.). Unlike the stacked graphs available in WP:VG's Article statistics, the graphs below overlay each curve, so events like promotion from Good to Featured will have the effect of reducing the Good curve while increasing the Featured curve. In a stacked graph, the overall effect would show a net zero on the y-axis whereas increase and reduction in the overlaid graphs below must both be taken into account simultaneously in order to construe Project-wide progress accurately. In other words, a loss of Good status is not a negative thing if it is to promote it to Featured status.

Historical/contextual notes are presented below both graphs. Speaking prehistorically then, readers should keep in mind that Wikipedia was created on 15 January 2001 (1Q01), that WP:VG was formed about two months later (still in 1Q01), and that by the end of the year (4Q01) the precursor to Featured Articles had been established. The WP:VG Newsletter was established in the first quarter of 2008 (1Q08) more than 6 years after the formation of the WikiProject. Regarding the data, all quarters prior to 1Q08 have been counted by hand during the first quarter of 2020 (1Q20), so the totals may not be strictly accurate. Specifically, some older topics only became "video game topics" after historical events took place. Microsoft would have only become a video game topic after the release of the Xbox, for example. Prior to that it might have only been considered a computer topic and a business topic. In addition, articles like Cortana have been renamed between 1Q08 and today (i.e. Cortana versus Cortana (Halo)). For articles that predate the creation of WikiProject Templates and that predate reporting in the WP:VG Newsletter, new high-quality articles have been counted based on the 1Q20 WP:VG membership. Quarters previously unreported by the WP:VG Newsletter (i.e. 2Q10, 1Q18–1Q19) have also been calculated by hand during 1Q20. All other quarters are drawn from quarterly reports from the WP:VG Newsletter.

Finally, readers should take note that other potential issues including historical merges, deletions, splits, and renamings may all effect the final tallies. In all cases, a best effort was made to determine the intent of the (sub)topic for each article, and Good/Featured counts were applied based on 1Q20 designations. For a good up to date cross-Project comparison dating back to the 1Q08 dawn of the WP:VG Newsletter, please see SandyGeorgia's FAC talk thread "Where we write (and where we don't)".

Dates to note:

• March 2001 (1Q01) - WP:VG is formed
• November 2001 (4Q01) - FAs are invented (originally referred to as "The Excellent Prose List" and treated similarly to Barnstars with community input but no nomination process)
• May 2003‎ (2Q03) - FPs are invented (originally referred to as "Brilliant pictures" and allowing image files and charts)
• June 2003 (2Q03) - FAs now must be nominated via "Brilliant prose candidates" (today WP:FAC)
• November 2003 (4Q03) - FPs now must be nominated via "Brilliant pictures candidates" (today WP:FPC)
• May 2005‎ (2Q05) - FLs are invented
• November 2005‎ (4Q05) - FTs are invented
• December 2005 (4Q05) - FPOs are invented
• March 2008 (1Q08) - WP:VG/NL begins to track and report new FAs, FLs, and FTs monthly
• April 2009 (2Q09) - WP:VG/NL tracks and reports new FAs, FLs, and FTs quarterly. Only a few quarters go unreported (i.e. 2Q10, 1Q18, 2Q18, 3Q18, 4Q18, 1Q19)
• March 2012 (1Q12) - WP:VG/NL begins to track and report new FPs quarterly
• March 2016 (1Q16) - WP:VG/NL begins to track and report new FPOs quarterly
• March 2017 (1Q17) - FPO designation is marked "historical". Then-existing FPOs are allowed to maintain their status by replacing their coveted "Featured" bronze star () with an empty star (). As a historical designation, new FPOs cannot be nominated, but they can be lost if the portal is deleted.

In the second graph we cover A-class and Good content. Although the data presentation in this second graph only begins in 3Q05, the x-axis has been extended back to 4Q01 matching the scale from the previous graph. This facilitates comparison between the curves and perhaps allows explanation for significant promotion-heavy quarters which may be flattened or even reduced in the second graph as good status is lost and featured status is gained to increase the curve on the first graph above. A good example of this can be seen at 1Q15 when the community decided to deprecate the A-class category. Comparison of the curves demonstrate that although the FA designations didn't see an appreciable increase, the GA designations took a sudden jump. This shows that rather than converting A-class to FA, they were instead converted to GA en masse.

Categorically generated quarters for designations like the A-Class figures (subsequently deleted) do not leave a historical footprint if not recorded, so the figures for 2Q06–2Q08 and 2Q10 are drawn from WP:VG's Article statistics. All other quarters are drawn from quarterly reports from the WP:VG Newsletter. For the purpose of comparing the Good/A-Class graph below with the Featured graph above, please note that although the absolute height of the y-axis matches that of the first graph, the scale is considerably different.

Dates to note:

• September 2005 (3Q05) - A-Class designation are invented by the third quarter of 2005 or earlier
• October 2005 (4Q05) - GAs are invented
• March 2008 (1Q08) - WP:VG/NL begins to track and report new GAs monthly
• July 2008 (3Q08) - WP:VG/NL begins to track and report new A-Class designations monthly
• August 2008 (3Q08) - GTs are invented
• October 2008 (4Q08) - WP:VG/NL begins to track and report new GTs monthly
• April 2009 (2Q09) - WP:VG/NL tracks and reports new GAs, GTs, and A-Class designations quarterly. Only a few quarters go unreported (i.e. 2Q10, 1Q18, 2Q18, 3Q18, 4Q18, 1Q19)
• January 2015 (1Q15) - A-Class designation is deprecated

In reviewing the two graphs above, it is clear that in the last decade there has been a trend toward increase in all non-deprecated designations (with the possible exceptions of FTs and FPs). Tracking the history of the FAs we see an initial exponential burst of growth (4Q01-1Q09), followed by a period of linear growth (1Q09-3Q12) at a rate of 1.8 articles per quarter, followed by a more variable period of generally linear growth (3Q12-2Q16; possibly influenced after 4Q14 by the deprecation of A-class) at an increased slope of 4.1 articles per quarter, and finally a return to 1.8 articles per quarter linear growth (2Q16-4Q19). This can perhaps be best compared against the contemporaneous GAs and A-Class periods within their respective histories. Tracking the GAs we see considerably stronger growth curves than with the FAs and could be approximated to linear growth at a rate of 25.5 articles per quarter. Looking more closely we can identify three separate periods of GA growth starting with a period of very steady linear growth (4Q05-2Q11) at a rate of 22.5 articles per quarter, turning later to a slower and more variable linear growth (2Q11-3Q15; possibly influenced by the conversion of A-class to GA in 4Q14 and corrections from GA to FA in the following 3 quarters) at a rate of 19.1 articles per quarter, and finally entering a period of mild third degree polynomial growth (3Q15-4Q19). Tracking the A-Class category shows two distinct periods of growth initially displaying variable linear growth (3Q06-4Q09) at a rate of 2.1 articles per quarter and then proceeding at a much slower (though less variable) linear growth (4Q09-4Q14) at a rate of 0.35 articles per quarter until it was deprecated. Alongside FAs, GAs, and A-Class articles we can also profitably take into account the FLs as stand-alone list articles are in fact a featured form of articles and thus could be considered as an artificially altered totals for the FAs. Here we begin with a period of linear growth (3Q05-4Q07) at a rate of 0.7 lists per quarter and then developing at a fairly constant logarithmic growth (4Q07-4Q19).

A comparison can also be made between the FTs (above) and the GTs (immediately above). Tracking the history of the FTs we see that this is one of only two high quality content histories (the other being FPOs) that displays a non-increasing overall trend. Tracking the history of the FTs shows three periods of development starting with a linear growth (3Q06-3Q09) rate of 1.3 topics per quarter, reversing direction to become a variable linear growth in a negative direction (3Q09-4Q10; i.e. a reduction) at a rate of -2 topics per quarter, and then rallying with very very slow linear growth (4Q10-4Q19) at a rate of 0.1 topics per quarter. The history of the GTs is considerably simpler with slow linear growth (2Q08-4Q19) at a rate of 0.6 topics per quarter. The last two curves above are those of the FPOs and the FPs which are probably best compared with other featured growth curves. As with the history of the FTs, the FPOs also displays steady non-growth (4Q15-3Q19) and, like the A-class designation, comes to a premature end as FPOs are marked historical. Finally tracking the history of the FPs, growth appears to develop in two phases first displaying slow linear growth (4Q11-4Q13) at a rate of 0.9 pictures per quarter and then changing to variable logarithmic growth (4Q13-4Q19).

Thanks to GamerPro64 for pointing out SandyGeorgia's cross-Project comparison. Thanks to Salavat, MrKIA11, MuZemike, and everyone else who helped generate the very gnomish raw data represented in the graphs above. Thanks again to PresN for historical assistance with the A-Class category and remember dear readers to stay safe, wash hands before sharing controllers, practice safe spacing, and ultimately when it comes to the virus: Just Survive!

+ Add a commentDiscuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try .

Empty star icon at COM:RFD

The empty star icon has been nominated for deletion: https://commons.wikimedia.org/wiki/Commons:Deletion_requests/Files_found_with_insource:flaticon «“I'm Aya Syameimaru!”I„文々。新聞“I„userbako”» 19:24, 15 April 2020 (UTC)

Interesting discussion. At this point it looks like it will be closing without deletion. Sadly WP:VG has only ever had one of these, but it seems like a useful symbol. Presumably the same sort of thing might have been awarded for A-Class articles after the deprecation if they had been treated the same as Good and Featured content. I wonder what the next group that might plausibly be marked historical. Maybe the "Topics" (FTs and GTs) whose growth has basically flatlined. If that was to happen I'd hope the empty star was available to designate our former achievements. -Thibbs (talk) 15:35, 16 April 2020 (UTC)

Hungry for further analysis?

I feel like I didn't do enough to analyze the data above. I recognize that the text may be a bit dry, and for some this may be a "Thank you Alex, you've done quite enough" scenario so I will collapse the extended analysis (i.e. my own notes on how I interpret the curves above). It's nothing spectacular, but for those who want to read more, please read on.

Further analysis

• In general I would interpret the growth curves in the following manner:
  • Growth vs reduction - This is pretty basic, but if we see growth that means that work is being done. Articles are being passed through the quality processes and articles are being advanced to high quality articles. Reduction or "negative growth" (which only occurs as a trend for FTs during 3Q09-4Q10) indicates either that previous assessments are being re-evaluated, or that articles are being lost (deletion, merging, etc.)
  • Steady vs variable - Steady growth indicates that quarterly growth remains relatively consistent from quarter to quarter. If you look the rate of growth from 1Q15 to 2Q15, the same rate of growth would closely match the expected rate of growth between 1Q15 to 1Q20. Variable growth indicates that there is instability in the expected rate of growth such that a quarter-by-quarter comparison would produce surprising bi-quaterly, tri-quarterly, or yearly results, but that over longer periods a general trend would be observable
  • Exponential growth - Growth at an increasingly rapid pace over a constant period of time. The only portion of the above graphs that demonstrate exponential growth is in the FAs during 4Q01-1Q09. Growth like this seems to indicate an explosion of interest in a topic.
  • Third degree polynomial growth - Rapid growth that halts or reduces before returning to rapid growth. This kind of growth appears to be roughly demonstrated by the GAs during 3Q15-4Q19, and suggests an explosion of interest, interrupted for a period, and then returning to explosive growth. The interruption period may either reflect a period of lack of interest, instability (as from policy changes, community discussions about standards, etc.), or re-evaluation.
  • Linear growth (<1) - Slow but steady growth that maintains a constant rate. This may suggest some degree of disinterest in the category by the community.
  • Linear growth (>1) - Fast and steady growth that maintains a constant rate. This suggests that interest is steady but not explosive. The greater the rate above 1 per quarter, the greater the level of interest. This tends to be typical of well established categories that are not yet approaching the maximum of the category.
  • Logarithmic growth - May indicate either general disinterest or maximization of the category over time. Investment in the growth of these curves show that there was a period of great growth at first, but that the rate of growth has slowed and the curve has been flattened to remain at steadily slower increase.
  • Non-growth - A flat curve indicated disinterest or maximization of the category. Non-growth at a low level tends to suggest disinterest whereas maintenance of non-growth at a higher level suggests maximization or saturation of the category.
  • Negative growth - A reduction of the number of catalogued entries. The only period demonstrating negative growth is with the FTs during 3Q09-4Q10. This kind of curve most likely suggests a period of re-evaluation either by stricter monitoring or the development of stricter criteria.
  Regarding the most recent year or so: FAs show linear >1 growth which suggests that growth in this category is generally important to the Project and that the maximum is not yet in sight. GAs show third degree polynomial growth which suggests that there is great interest in increasing this category. The reason for this may be attributed to zealous GA creation by the community (or individual Project members) or re-evaluations leading to promotion. Standards may also play a large part especially if linked to re-evaluations. FLs are growing logarithmically so interest seems to be either waning or approaching the categorical maximum. FTs and GTs both demonstrate linear <1 growth suggesting possible disinterest in the category, and the logarithmic nature of the FPs suggest that the category has reached a maximum (which seems extremely unlikely) or that there is general disinterest in the topic (which is more likely but still seems implausible). An alternative explanation is also available in this category when the individual images are regarded as a gallery. The vast majority of FPs were created by a single editor, User:Evan-Amos (see 2015 Interview from WP:VG Newsletter (Volume 8, No. 2)) who seems to have self-imposed limits for his own contributions in this category to video game consoles. It is possible that Evan-Amos is losing interest in the subtopic, the category as a whole, or that he is beginning to reach a categorical maximum (either limited by the number available as a whole or by the number available to him) and this illustrates a very interesting phenomenon: that individual editors can have a significant impact on the curves we track simply by participating zealously. Although most of my analysis attributes growth changes to the community, a more thinly-sliced analysis may show that individuals are more significant contributor than the contributions of the community and although this claim has been called into question for articles, for images this in fact supports Jimmy Wales' 2010-era claim that 2% of editors have created 3/4 the all of Wikipedia's content. For non-collaborative works like images, Wales' claims may be true.

Additional further analysis to cover the past decade (i.e. the original idea that prompted this feature) could also be forthcoming if I find the time. Please read or ignore at your leisure per your preference. -Thibbs (talk) 15:35, 16 April 2020 (UTC)