***
Whenever a kyu player asks how to improve at shogi, they're always met with three answers: play a lot, analyze your games, and solve tsumeshogi. These are not wrong, this is quite possibly the fastest method to improve. But something I thought about many years ago, and that came up this week, is that I think many kyu players whose reading ability comes solely from tsumeshogi pick up some bad habits from these very important training materials. Given the ways tsume puzzles are normally arranged, the "main line" usually has a very specific way of succeeding, which requires exhausting exactly what you have in hand, and often involves dancing around pieces on the board gracefully. I don't deny that this is very useful practice, because it does allow you to sometimes perform magical comebacks in situations that look completely safe.
But if all of your practice with checkmates comes from books, you become conditioned to these graceful dances around pawns and trying to avoid extra pieces. Let me show an example from Madoka Kitao's Ending Attack at a Glance:
Take a few moments to try to solve it before looking at the answer.
The solution to this, of course, is ☗R*82 ☖K-93☗ Rx83+ ☖Kx ☗R*82 ☖K-74 ☗G*75. To an experienced player, this should be obvious. But what about to someone whose primary experience comes from composed puzzles? They may see that line, and have the kneejerk reaction of "Oh, I picked up an extra pawn somewhere, so that's not the main line. Let me try doing it without taking a pawn." Thus, the player subconsciously believes this is not a true solution.
When you solve tsumeshogi, you should not only be solving for the solution provided, but also for all variations. If there are variations you haven't considered, then you have not truly solved the tsume. Furthermore, they are all valid solutions.
Another issue that I often see come up with kyu players is an obsession to empty their hand when performing a checkmate, instead of going for the fastest mate, because they're used to finishing tsumes with an empty hand. This is just a composed example, but this is something I've seen when watching new players' games:
Suppose the king's side has no pieces in hand. In such a situation, I've seen players go for extremely convoluted checkmates such as ☗S*22 ☖Kx ☗Sx43+ ☖K-11 ☗P*12 ☖K-21 ☗+S-32 ☖Kx12 ☗+S-33 ☖K-11 ☗N*23 ☖K-21 ☗R-22+, emptying their hand and perhaps being very proud of themselves for executing a 13 move checkmate. The problem is that a much less flashy ☗Rx43+ ☖K-12 ☗D-23 ☖K-11 ☗D-21 would have sufficed. They might even remark that they wanted to use all the pieces in their hand, but I believe this is a bad habit that will lose them games in critical positions where they may have misread a later move in their flashy line. Imagine, or example, being in 30 second byoyomi and trying to figure out this long checkmate, completely missing the 5 move mate.
So what can players do to combat these habits? Well, the two easiest solutions I can think of are to simply play the game a lot, and also to watch others' games. You can also solve tsumeshogi based on real games rather than composed problems. Such tsumes will rid you if the inherent desire to dance around pieces artistically.
You could also try endgame books that are not strictly tsumeshogi. Of these, I recommend the previously mentioned Ending Attack at a Glance by Madoka Kitao and Which Pieces Do You Need to Mate? by Daisuke Katagami. Kitao's book covers realistic tsume and hisshi positions with no real regards for pieces in hand, while Katagami's book will force you to look at checkmating in a whole new way, by training you to know what pieces you want rather than using the pieces you have. Both books are available in English.
That all being said, this is purely a player issue and is not an inherent fault of tsumeshogi. Tsume is perhaps the most useful tool for getting better at shogi outside of simply playing, so I do not want to discourage anyone from using it. Rather, players should make sure to be using it properly instead of simply confirming that they found the answer that the other wrote on the next page. Used properly, tsumeshogi can carry you so hard in the endgame that you can consistently turn around games that should have been lost in the opening. Just use it responsibly.
I’ve not actually seen this aversion to efficient koma amari variations in person except as a joke. Whenever I teach endgame technique to beginners at the local shogi club, I try to emphasize that they must find the shortest variations possible.
ReplyDeleteThe true sagely advice is hidden a little deeper in the blog post: “You should not only be solving for the solution provided, but also for all variations.”
When a beginner is solving tsumeshogi in front of me, I’ll ask the player questions like “Why can’t your opponent defend if they instead capture with this piece?” or “Why does this pawn need to be here in the position for the mate to work?” or even stuff like “Would this mate still work if you had a different piece in hand? What if you had a silver & knight instead of a gold” etc. By reading out all the variations and thinking deeply about why the solution works, I feel they get more out of the tsumeshogi than if they were only looking for the main line that empties their hand.
Similarly, by fully exploring all the non-working lines, I’ve found this works as very good training for avoiding mates. Doing this deliberate training also turns every tsumeshogi problem into a King escape problem!