Talk:Shogi opening

New demon killer: last diagram has five knights

New demon killer: last diagram has five knights. 5.18.215.27 (talk) 18:18, 12 November 2016 (UTC)

Thanks. It's always good to have proofreaders/other editors. peace. – ishwar  (speak) 18:49, 12 November 2016 (UTC)

the so-called 'other' openings

Just a note: i've put several openings that can be found in books but arent found in most English language literature. I'm not sure what to do with these. Some can certainly be grouped into static vs ranging. For instance, Ureshino is definitely Static Rook, etc. But, they seem to have a different status as they're not traditional and probably won't be found in professional games. Not sure if that's a valid way to classify them though. In other words, the category is kinda defined mostly by function (i.e. within cultural gameplay usage, etc.) but also by form as well. It's a hodgepodge. – ishwar  (speak) 04:00, 6 December 2016 (UTC)

I think they're fine as "other openings". For sure, they probably could be classified as RR or SR in the most basic sense, but I think they mostly are used for surprise effect to try and gain a quick advantage against an unprepared opponent. You'll probably never see things like "Demon Slayer" used in serious professional games or very high level amateur games when there's something real at stake simply because "hamarite" type openings usually don't fool these types of players. Better chance of seeing these things at local club games or online quick games where winning might not matter as much and there's more room to experiment. A player might be trying to get their opponent into a "Rikisen" type game or maybe is just looking for something to try and keep himself/herself interested. I've won some games playing "Genshi suji chigai kaku" or "Ukiuki Bisha", etc., but best play usually leads to an even game at best. Some like "Right king", however, will pop up every now in pro games usually as the result of transposition. There are similar kinds of openings in chess. Some players just prefer these offbeat openings as a way to keep the game exciting and take their opponent out of book and also possibly to pick up a quick win.

I don't think there's a standardized system for classifying shogi openings like the ones used by the Chess Informants for chess openings, so it's basically RR, SR and everything else. -- Marchjuly (talk) 04:45, 6 December 2016 (UTC)

Ok. I'm conflicted on how to organize things. There's basically two ways to classify openings: functional vs formal. Currently, the article is using a primary functional split into traditional vs nontraditional. And, then the traditional category uses a purely formal classification (by rook position) which is usually how they are presented in books.

However, as you note, putting Right King into that nontraditional category is based more on formal grounds than functional. I'm not even sure whether Right King should be there at all. It could be thought of as a subcomponent of an opening much like Climbing Silver can used as a subcomponent of different openings. It just seemed worth mentioning in a main article since it violates the king-rook proverb. But, since these subcomponents are used in naming the openings (like Bishop Exchange Climbing Silver, Bishop Exchange Reclining Silver, Double Wing Attack Climbing Silver, etc.), perhaps most could be mentioned here? But, maybe that becomes a bit unwieldy? And, where would this information be put into this article?

Japanese editors have come up with a sort of classification: ja:Template:将棋の戦法. But, I'm not sure how much sense it makes.

筋違い角 is a good one. I think Fairbairn (but not Hosking) mentions it, too.

How to translate 浮き浮き飛車? Floating Floating Rook? (It's not common to repeat an verbal adjective like this in a name, but we can't use Floating Rook since that's already a translation of 浮き飛車.) Kawasaki (Hidetchi) called this position (played by Gote) 'a melee' in his Kitao translation. I don't know what the original was, but if it was 浮き浮き飛車, melee seems too vague. And, if Kitao didn't use the name in the original Japanese, it seems like instead we should use the name since that's more specific. (I guess that's obvious – I'm just trying to defer as much as possible to earlier translations.) (Incidentally, there's no 浮き浮き飛車 position in the Kyokumenpedia professional game database, but you see it in a few of the high-ranked player 81dojo data – it's rare there too.)

For the chesslike classification, what I'm thinking of is definitely not Chess Informants. It's basically just the information you get by seeing the first several moves. Assuming two players will play a traditional opening, by the 4th move, many openings can be identified or at least narrowed down to subset of openings and by, say, the 8th move, most openings' basic form will be largely known. So, I'm thinking specifically of an NHK show where Tanigawa went through the several first 4-8 moves and connected them to traditional openings. And, there's a self-identified 'shogi evangelist' who goes by the moniker yamajunn on the Internets who has written in both Japanese and English a similar thing. There must be some book that does this too but I'm not aware of one. In some ways, this recalls the chess classification system, but it's nowhere near the branch depth and exhaustiveness of the chess analysis. Since this kind of information is well known to shogi players, it seems like it needs to be included in a description of the game. – ishwar  (speak) 16:13, 6 December 2016 (UTC)

I'm not sure if there is a shogi equivalent to the Encyclopaedia of Chess Openings in either Japanese or English where someone has placed pretty much every known opening into certain categories. I think we have to be careful to try and not create our own system of categorization using Wikipedia per WP:NOR, so maybe the best that can be done is to simply follow whatever reliable sources do. So, if a source like Hosking, etc. lists in particular opening as a ranging rook opening, then maybe that's what Wikipedia should do. If you check Amazon for "奇襲戦法", you'll find books like 奇襲大全 which might indicate that a more functional approach to categorizing these types of openings than the simple static rook/ranging rook/everything else approach is being used.

I also am not too sure how to best translate openings into English. Again, I think we probably should stick to what reliable sources are doing (even if that may not be the best possible translation in our personal opinion) whenever possible for Wikipedia's purposes. FWIW, I think I got "ukiukibisha" from a NHK Shogi Koza that Takahiro Toyokawa did back in 1999. It could have just been a name he made up to appeal to beginners and to be funny since he's kind of known for his sense of humor. Expressions like "ukiuki" are used quite a bit in casual Japanese and are probably best translated more contextually than literally, so maybe "The cheerful floating rook" (as lame as that may seem) is a step in the right direction. -- Marchjuly (talk) 01:30, 7 December 2016 (UTC)

I don't think there's any novel categorization here. It's trying to fit the weird ones into the basic categorization used by most sources.

The other issue to what to do with things like Climbing Silver. It's not part of the main categorization according to rook position, but it's part of openings in the sense that the term is used in the names of specific opening. Since it recurs in different names, it's obviously a subcomponent of some openings. What Hosking does is include Climbing Silver in the chapter that discusses the basic opening techniques like exchanging the rook pawn and opening the bishop diagonal as well as some of the well-known tactics (tetsuji). And then, in the main opening chapters, the term is added to the names of openings. So, it appears in both places. Fairbairn even does something unique with Reclining Silver: while he mentions a Bishop Exchange opening, he terms the Bishop Exchange Reclining Silver as Reclining Silver Bishops Off, which seems to promote Reclining Silver up the main categorization term in contrast with other sources. So, that leaves me with uncertainty with where to put these things. It's not an issue with other substrategies that are so specifically termed that they only need to be mentioned within the topic of the main opening. It's just that things like Climbing Silver and Reclining Silver reoccur in different openings. One thing to do is to simply ignore that Climbing Silver is a subcomponent. That's what Kitao does, for example: this is the Yagura opening and this is Yagura Climbing Silver and this is Yagura Waki System... And, this is the Double Wing opening, and this Double Wing Nakahara and this Double Wing Climbing Silver, ....

I don't see ukiukibisha in a Google Books search. But, it's on the web. Maybe you can add a citation? I guess 浮き飛車目くらまし refers to the same thing, right? If so, is that the more common term?

Good point about using 奇襲戦法 for thinking about organizing. – ishwar  (speak) 17:17, 7 December 2016 (UTC)

Tanigawa NHK Koza

Most of the NHK Kozas (before NHK switched to the Shogi Focus) tended to be completely instructional type programs without any of the variety-show type stuff they are currently doing. The information during the actual episode itself often was only the main points the pro was trying to make, and more details could typically be found in the monthly NHK Shogi Koza magazine that NHK published each month. The idea was that a viewer could refer to the text while the topic was being discussed on TV. NHK still publishes this magazine and there is still info about each month's "lessons" in it, but the focus of the program has shifted a bit to try and increase shogi's appeal beyond it's more traditional base. Anyway, back in the say when a certain "season" was finished, the "course" usually ended up published as book by NHK which tended to expand on both the magazine and TV lessons. The book about the Tanigawa course can be found here. I don't own a copy of it, but it might be possible to tweaked the citation a bit from {{cite episode}} to {{cite book}} since I'm pretty sure the same thing discussed on that particular episode would also be discussed in that book.

On a side note, an interesting thing about this particular koza was that Tanigawa and Saori Shimai did a little doctor-nurse cosplay in one of the episodes. Not sure whose idea it was, but it was a bit "unusual" for the program and seemed pretty awkward. You can find clips of it on YouTube. -- Marchjuly (talk) 00:50, 9 December 2016 (UTC)

Yes. However, it is based on the TV show. And so, we can't know for sure if the same information's in the book. (Weird about the cosplay. I'll try to find it.)

I'm going to try out a new diagram template that is text-based instead of png/svg-based (except for the arrows, which will still be the image file). I think it should help on page load time. (I still have to add all the arrow filenames to the code though.) – ishwar  (speak) 01:26, 9 December 2016 (UTC)

That's fine. I watched that Koza episodes, but don't remember much about it except for the cospay bit. It might help to add the exact date the episode aired since you're referring to a particular episode and tweak the citation template accordingly. That shouldn't be too hard to track down since the kozas typically run for 6 months at a time beginning in April and October. Tanigawa's started in April 2006, so probably the first episode aired on April 2 or April 9. There may even be something on YouTube where an actual quote can be found (just don't link it if it's a copyvio per WP:EL#cite-note_copyvio-exception-2) which might be useful too. -- Marchjuly (talk) 01:42, 9 December 2016 (UTC)

Trends in pro opening usage

Kiyokazu Katsumata 6-dan appeared on the April 16 episode of NHK-E's Shogi Focus [1] and discussed trends in professional shogi opening usage. He focused on three "shogi years" in particular (1983-84, 1999-2000 and 2016-17) presented a breakdown of ranging rook and static rook opening usage in all the official professional games played in those years. The segment was titled "プロの流行戦法" (Puro no Ryūko Senpō") and started at the 15:35 mark of the episode.

Katsumata is well-known among shogi professionals, etc. as a shogi researcher and being involved in computer shogi. So, I think it's fair to consider him an established expert in the field. It's also unlikely he would have annonced these findings on nationwide TV if he did not feel his analysis was reliable. He used pie diagrams on the program, but simple wikitables could be used for the article. Anyway, what he found is as follows:

• Overall breakdown for the shogi year 2016-2017

Opening usage for 2016-2017

Opening name	Usage percentage
Yagura	15%
Kakugawari	10%
Yokofudori	17%
Aigakari	9%
Shikenbisha	13%
Nakabisha	19%
Sangenbisha	10%
Mukaibisha	7%

• Ranging rook opening usage breakdown

Ranging rook opening usage percentages

Opening name	1983-1984	1999-2000	2016-2017
Shikenbisha	55%	68%	26%
Nakabisha	21%	12%	45%
Sangenbisha	21%	13%	23%
Mukaibisha	3%	7%	6%

• Static rook opening usage percentages

Static rook opening usage percentages

Opening name	1983-1984	1999-2000	2016-2017
Yagura	60%	41%	28%
Kakugawari	8%	10%	24%
Yokofudori	2%	34%	32%
Aigakari	30%	15%	16%

If there's someway to incorporate this information into the article, then I think it should be done. There's not a lot of data analysis of opening use available from English sources and much of it probably would just be repeating what pros such as Katsumata has made public. The episode itself could be cited using {{Cite episode}} and there could also be some background given as to why Katsumata decided to focus on these three particular shogi years. -- Marchjuly (talk) 01:02, 23 May 2017 (UTC)

I think a line graph is the best visual representation for this type of data. They can accompany data tables. Pie charts are ill-suited to show changes over time as it's hard to gauge differences in pie slice sizes

Just looking at this, i dont see how aifuri is coded. Is it included within the Ranging Rook labels? (But how?) I guess he just ignored it?

Also, there are some Double Static Rook openings in pro games that dont appear to fit into the traditional openings at least as i understand them. How are they coded? I guess they are new. Maybe they are Aigakari – not sure. For example, openings with no early rook pawn trade or early bishop trade, but they arent Yagura either. (Another example that i at least know the name is Ai-Gangi, it's obsolete now but it wouldnt seem to fit his opening labels.) Anyway, i'm disappointed that there is not 'other' label, which suggests that he is omitting some games. Well, we can't fix his data analysis, but there may be a caveat to mention....

Just a clarification: when you write 1983-1984, you mean 1983 Jan 1 – 1938 Dec 31? – ishwar  (speak) 16:35, 23 May 2017 (UTC)

Responses in random order.

• Line graphs are fine. I used tables just because it was easy for me to show them that way.
• The years are "shogi years" (将棋年度) which is the same as the Japanese fiscal year/school year, etc., etc. and runs from April 1 until March 31 of the following year.
• The opening names and categorizations were the ones used by Katsumata. He was fairly general and did not go into too much detail about specific types of openings and posistion except to provide some background as needed. The enitre Shogi Focus program is only about 30 minutes long and his bit was less than half of that, plus there has to be time for some conversation, etc. between him and the two hosts. He did mention that Hineribisha was used quite a bit back in 1983-84 and was the most popular of the Aigakari openings.

I deleted the program from my DVR, but you can find it on YouTube if you search プロの流行戦法. If you have the time, you should check it out because it is still interesting even if it's not something suitable for Wikipedia. -- Marchjuly (talk) 00:11, 24 May 2017 (UTC)

• File:Furipopchange.png
• File:Ibishapopchange.png
• File:ShogiOpeningPop2016.png

cartoons – ishwar  (speak) 03:48, 26 May 2017 (UTC)

Thanks for thaking the time to do that. The diagrams look fine. Wouldn't the actual data used also need to be added? The diagrams show the general trends, but the actual percentages presented by Katsumata are not clear. -- Marchjuly (talk) 23:09, 26 May 2017 (UTC)

The percent scale on the left axis has labels. Is it unclear because you want more tick marks on the axis? As for as the 'actual data', that is the actual data! It's just a visual representation.

I can actually add years (1976, 1977, 1982, 1983, and i'm still flipping though...) to the graph with the data from Shogi magazine. However, the data for 1983 does not completely match Katsumata. We could use Katsumata as primary in terms of making a graph and mention the discrepancy textually. And, Shogi gives us separate numbers of the amount of hineribisha games, so we can report that, too. It also gives details on furi games broken down by sente vs gote. And, for 1982 & 1983 (but not 1976, 1977), it tells us about the associated castle and the counter-strategy used by the ibisha side. However, we don't have that data for 1999 or 2016. So, we can't do a comparison to see trends. And then, the question becomes do we want to give a detailed report for this earlier period or not? – ishwar  (speak)

Need to check color blindness issues.... – ishwar  (speak) 13:19, 27 May 2017 (UTC)

By "actual data" I meant expressed as numbers and not just a graph. The graphs do make it easier to see the trends over time, but the actual percentages aren't clear (at least not to me). The extra tick marks could help, or maybe there's a way to add more info to the caption or as textual content. In addition, Katsumata only gave his data for those three "shogi years"; the line graphs sort of make in seem as if there were steady increases/decreases over time, but there's no way of knowing whether there were spikes in usage in any of the in-between years. The data you found for the other years may be able to help clarify that. Could the bar graph you created for 2016-17 also work for comparing the other years. -- Marchjuly (talk) 23:06, 27 May 2017 (UTC)

Ok. You are a better person judge than me, i guess. (I look at graphs too much.) There are 2 options: (i) increase info on graph, (ii) increase info outside of graph. It depends on the tradeoff of graph complexity vs simplicity. For the graph, we can add (a) labeled ticks on axis and/or add percents as numbers (like near their square dots on the lines). External to the graph, we can add info to the caption and/or to the paragraph. We can try all of the above and see what works best.

I also thought about making either a Japanese language version of the graphs (for jawiki) or making the graphs free of language so that all lang wikis can use the same image.

As for Katsumata's data choices, it's a good point. When quantize (coarsely sample) the data like this, we don't what devil may lie in the details. It can be dishonest. Maybe he did choose exemplary data but it could also have been cherrypicked ones to show extreme values. Unfortunately, Shogi is not that much help. It doesn't show detailed opening stats for all years. And, it died abruptly in 1987. (G Hodges only got back about half of what he spent on popularizing shogi even after over a decade. He spent about the British equivalent of $50,000 – not adjusting for inflation. I guess no one has worked as much to popularize shogi than him – not even the Hidetchi guy.) But, Shogi will add more information. I'll post a more info some time/days later. I guess this info is somewhere though. I don't have the detail for the 2016–2017 year for Katsumata's 1999 year. But, Shogi does have it for 1983–1984 (although we can't completely reconcile with Katsumata's 1983 data). – ishwar  (speak) 15:46, 28 May 2017 (UTC)

Shogi data

So, i've gone through all the Shogi issues. Unfortunately, it only gives detailed opening stats for the 4 years already mentioned above. The other years' stats have info on total number of games, winning ratio of Sente/Gote, etc. but nothing on openings.

Here is the magazine's 1983/1984 raw data:

1983-1984

Opening name	n	%
fortress	518	24.1
irregular fortress	112	5.2
side pawn black	27	1.3
side pawn white	3	0.1
side pawn both	0	0.0
twisting rook black	194	9.0
twisting rook white	6	0.3
twisting rook both	1	0.0
climbing silver rook pawn off black	5	0.2
climbing silver rook pawn off white	1	0.0
climbing silver rook pawn off both	0	0.0
reclining silver rook pawn off black	8	0.4
reclining silver rook pawn off white	12	0.6
reclining silver rook pawn off both	5	0.2
climbing silver bishop off black	37	1.7
climbing silver bishop off white	12	0.6
climbing silver bishop off both	0	0.0
reclining silver bishop off black	9	0.4
reclining silver bishop off white	12	0.6
reclining silver bishop off both	13	0.6
wrong bishop diagonal black	4	0.2
wrong bishop diagonal black	3	0.1
other static rook vs static rook	292	13.6
central rook mino black	26	1.2
central rook anaguma black	6	0.3
central rook other castle black	27	1.3
central rook mino white	63	2.9
central rook anaguma white	17	0.8
central rook other castle white	43	2.0
4th rook mino black	134	6.2
4th rook anaguma black	14	0.7
4th rook other castle black	3	0.1
4th rook mino white	237	11.0
4th rook anaguma white	31	1.4
4th rook other castle white	6	0.3
3rd rook mino black	70	3.3
3rd rook anaguma black	8	0.4
3rd rook other castle black	2	0.1
3rd rook mino white	108	5.0
3rd rook anaguma white	3	0.1
3rd rook other castle white	9	0.4
opposing rook mino black	5	0.2
opposing rook anaguma black	0	0.0
opposing rook other castle black	6	0.3
opposing rook mino white	12	0.6
opposing rook anaguma white	1	0.0
opposing rook other castle white	25	1.2
double ranging rook	23	1.1

As you can see, there is an issue with the categorization here that makes it hard to reconcile with Katsumata. The aiibisha games are that don't fit in the yagura or side pawn categories are categorized differently. For example, Katsumata uses a label of aigakari while Shogi uses a categorization based on (a) climbing silver, (b) reclining silver, and (c) twisting rook which means there may be some of Katsumata's aigakari category games within in Shogi's (d) other label. Similarly, Katsumata uses kakugawari while these games in Shogi would be in variously as (i) climbing silver bishops off, (ii) reclining silver bishops off, (iii) wrong diagonal bishop and (iv) other. Undoubtably, there are a number of kakugawari hayakurigin (bishop exchange rushing/rapid advancing silver) games in Shogi's other (although i'm not really clear on the popularity of hayakurigin at this time). Because of this complicated categorization differences we can't even take the average of both sources' numbers. The only thing to do really is report data from both. It's too bad. Makes for a more complicated explication in the article.

In an effort to make the data more easily comparable, i've collapsed the yagura & yokofu into a single label. And, i've defined a new labels for Shogi:

• aigakari = twisting rook + climbing silver rook pawns off + reclining silver rook pawns off
• kakugawari = climbing silver bishops off + reclining silver bishops off + wrong bishop diagonal

1983-1984 (double) static rook

Opening name	Shogi n	Shogi %	Katsumata NHK %
yagura	630	49.5	60
yokofudori	30	2.4	2
aigakari	232	18.2	30
kakugawari	90	7.1	8
other	292	22.9	???

As you can see, the very large number other games makes the comparision problematic.

We can also ignore the other label to look at the discrepancies this way:

1983-1984 (double) static rook [other ignored]

Opening name	Shogi n	Shogi %	Katsumata NHK
yagura	630	64.2	60
yokofudori	30	3.1	2
aigakari	232	23.6	30
kakugawari	90	9.2	8

I'd guess that most of the problem is actually that several of Katsumata's aigakari games are inside of other since they were not twisting rook, climbing silver or reclining silver aigakari types.

The furi games are more comparable putting aside the question of how the aifuri games are treated by Katsumata. – ishwar  (speak) 17:50, 5 June 2017 (UTC)

Thanks for taking the time to gather all that data up and post it. Katsumata may have done things in such detail privately or for another book, but I think he realized that something that detailed would pretty much be impossible for "Shogi Focus". The JSA puts out yearbooks like this each year. I've only casually looked at them in the past, but I wonder if they have some more detailed analysis in there about the opening usage during a particular year. Maybe Shogi Sekai does something as well in some kind of "end-of-year" issue. I think that, like Katsumata, anything about this in this article needs to try and find a balance between being too simple and overwhelming the reader with too much data. If both Katsumata and Shogi can be cited and used to show general trends, then maybe that's enough. A simple basic sentence could be added to the article supported by an {{efn}} which can contain more specific detail with supporting citations. That way the readability to the article body is not interrupted by too much detail and reader's who want to know more can see the notes and citations. Just a suggestion. -- Marchjuly (talk) 21:47, 5 June 2017 (UTC)

new graphs

added text labels near the line points. Maybe all text needs to be bigger for viewing at thumb size. Clearer? – ishwar  (speak) 17:50, 8 June 2017 (UTC)

Possible Error

In the 'Identification by initial moves' section, in the paragraph beginning with 'Besides pushing only their rook pawn...' the paragraph gives a starting sequence of 1. P-76 P-34 2. P-26 P-84 and then says 'This 4-move sequence will usually be followed by 3. G-78 G-32 4. P-24 Px24 5. Rx24'. Here, move 4 seems to be impossible since the advanced black pawn is on 26 not 25. Is this an error or have I misunderstood something? Feanaro9 (talk) 22:51, 17 May 2018 (UTC)

@Ish ishwar: It does seem that the move "P-25" for black is missing from the sequence. I'm not sure how you want to add it. It seems that 4. P-25 P-85 should work, but 3. P-25 P-85 4. G-78 G-32 would also work. Not sure which move order you want to use or which is the more common these days. -- Marchjuly (talk) 00:29, 18 May 2018 (UTC)

Hi. Thanks. It's definitely an error, your understanding is perfect.

I intended: 1.P-76 P-34 2.P-26 P-84 3.P-25 P-85 4.G-78 G-32 5.P-24 Px24 6.Rx24.

3.P-25 after 1.P-76 P-34 2.P-26 P-84 has the highest probability in game records. And, i think it's often the only sequence mentioned in many Japanese joseki books (which often ignore transposition issues for reasons unknown to me). And, from memory, that's the order that Tanigawa presented on TV.

3.G-78 instead does logically give the White the opportunity to reject Side Pawn by closing their bishop diagonal with 3...P-44 leading to Yagura or Gangi or some other off-beat Static Rook form since 3.G-78 doesnt threaten the immediate rook pawn trade-off like 3.P-25 does. But, i dont know if any book has ever discussed this. It's just something that can be seen from the pro game records, which are always more varied than joseki books.

I'll fix it. It's always good to have other pairs of eyes improving what i wrote! – ishwar  (speak) 04:20, 18 May 2018 (UTC)

Thanks ishwar for fixing it and thanks Feanaro9 for pointing this out. FWIW, Gangi has had a rebirth in the last year and is often used in professional games these days, but back in 2006 it wasn't used a lot; so, you're probably right about the move order Tanigawa would've used. -- Marchjuly (talk) 04:53, 18 May 2018 (UTC)