Chu Shogi evert823 - dax00

evert823

Yes already resigned. 

dax00

"まで130手のお持ちまして、dax00の勝ちでございます。"

This brings my online chu shogi record to 99 wins, 1 loss.

I make a few mistakes in 60-minute time control, manually more mistakes in 30 minute time control, but not enough to lose, even against strong competition. It's really hard to play the strong + slow + quiet moves I love while under time pressure. My one loss was in a 20 minute game I took to over 310 moves, in which my opponent ran with his leopards and golds.

In the opening, you really have to make sure all your active pieces have moves. Immobility kills, especially when your opponent has a strong lion. Your pieces got stuck, which gave me plenty of time to slowly pick away at weaknesses. I don't normally try to take that much space, but you have me so much time that it was available safely.

HGMuller
dax00 schreef:

Also, the Edo rules allowed go-between promotion on the 12th rank on a non-capture, which I think is quite elegant and could be extremely beautiful in the rare instance that rule could be useful.

Are you sure about that? It is my intention to have my Chu Shogi engine play by Edo rules, but I never saw this one mentioned anywhere. I know of course that Pawns can always promote on 12th rank, but there is a good reason for that: Pawns are not able to leave the zone once they are in, and there can be a reason to defer promotion in the move where they enter. The Go-Between does not seem special compared to other pieces that do not have an upward-compatible promotion, though.

Do you have access to original Edo-era descriptions for Chu Shogi (or larger variants, such as Tenjiku)? If you have any opinions on the rules of any of those I would love to hear them!

As for repetition rules: Xiangqi formalizes the notion of defending (chased) player, but this gets horrendously complex. Asia rules try to be more explicit, at the expense of sometimes being non-sensical (e.g. considering whether a piece is protected, but not whether it is protected well enough). And when it tries to define subversion of protection as a chase, they even seem ambiguous. Perhaps in Shogi it would be simpler, for lack of Cannons.

dax00
HGMuller wrote:
dax00 schreef:

Also, the Edo rules allowed go-between promotion on the 12th rank on a non-capture, which I think is quite elegant and could be extremely beautiful in the rare instance that rule could be useful.

Are you sure about that? It is my intention to have my Chu Shogi engine play by Edo rules, but I never saw this one mentioned anywhere. I know of course that Pawns can always promote on 12th rank, but there is a good reason for that: Pawns are not able to leave the zone once they are in, and there can be a reason to defer promotion in the move where they enter. The Go-Between does not seem special compared to other pieces that do not have an upward-compatible promotion, though.

Do you have access to original Edo-era descriptions for Chu Shogi (or larger variants, such as Tenjiku)? If you have any opinions on the rules of any of those I would love to hear them!

As for repetition rules: Xiangqi formalizes the notion of defending (chased) player, but this gets horrendously complex. Asia rules try to be more explicit, at the expense of sometimes being non-sensical (e.g. considering whether a piece is protected, but not whether it is protected well enough). And when it tries to define subversion of protection as a chase, they even seem ambiguous. Perhaps in Shogi it would be simpler, for lack of Cannons.

I found the rule about go-between promotion written somewhere in Japanese. I'll update this comment once I find the source. [EDIT: (from 中将棋の指し方 by Okazaki Fumiaki 7段, which predates the JCSA) 仲人 前後一目行く。成れば酔象。仲人が不成で行った場合も歩と同じ。(translated: Go-between - Moves forward or backward one space. Promotes to drunk elephant. In the case of go-between unpromotion, it goes the same as for pawns.)]The reason for this rule seems clear to me. As well as pawns, go-betweens do not facilitate double-capture with a lion. The JCSA got rid of the rule because they can go backwards. [EDIT: The Yokohama Tea Museum runs a classical game club. One of the games available there is chu shogi, played on a very old floor-size (huge) board. They still play by the old rule of go-betweens being allowed to promote on the 12th rank.]

But imagine a scenario where unpromotion is the correct move to block from a ranged attack while not allowing a lion capture, and further advancement of the go-between is necessary (perhaps for "checkmate") while it is being chased from behind (and thus unable to retreat and reenter for promotion) by a piece like a tiger.

While on the topic of Edo rules, they most likely never played with the "baring rule". I dislike the baring rule for games other than shatranj. Either you can capture the enemy king, or you can't - simple. The baring rule is probably meant to save time in the case of someone playing on in a hopeless situation. But king capture is normally quite simple. If indeed drawn, and the player with an extra piece continues on hopelessly, there is a modern "local" rule whereby a player can claim a draw after 199 consecutive moves without either capture or promotion.

HGMuller

I completely agree with your stance on the baring rule. As far as I could trace the idea that Chu Shogi had a baring rule to complete twisting of the meaning of a sentence meant as strategic advice in one of the Edo manuscripts. It seems that Sho Shogi had a baring rule, though. But perhaps this is also made up, as it seems to need one even less.

About the modern draw rule you mention: it seems obvious that a rule of this kind is very desirable. But what I dislike about it is that some end-games, such as 2 Tigers vs Queen, which in unlimited play would be wins because of the ban on perpetual checking, will turn into draws. Because a Queen can check in so many different ways on a near-empty board it can take hundreds of moves before the defending player has to allow a single Tiger move, though, especially if you only judge the perpetual after a 4th repeat, so that it can take thousands of moves to force the win. So putting a reasonable limit to the number of non-progress moves will be tantamount to allowing draw by perpetual checking. A less disruptive rule would be to combine a maximum number of non-progress moves with a maximum number of checks in any non-progress period. This should make KTT vs KQ winnable in a reasonable number of moves, as there will be only 2 promotions in the winning line replenishing the 'checking budget'.

About Go-Between promotion: the inability to act as a 'capture bridge' is indeed what makes a Tokin not upward-compatible with Pawn, and thus provide reasons for non-promotion. But there are many other pieces that do not have upward-compatible promotion, and the argument you give could be applied just as easily to those. E.g. you might want to block a check by a Queen on 8b by moving a Copper from 7e to 8d, without promoting, in order to keep something on 9c protected. There will certainly be cases where it would come in handy when the Copper then is allowed to steam on to 12th rank to promote there. Even a Lion is not upward compatible with a Kirin.

dax00

Clearly 4-fold repetition is way too lenient, an extrapolation from modern shogi's repetition rule and not from history, where any single repeated check position would be illegal. With single repetition forbidden, I really can't envision it taking more than 199 moves to safely guide a tiger to promotion in a [K+BT+BT vs K+FK] endgame. I'd like to play this endgame out (playing with the tigers) to see how long it would take. It shouldn't be difficult to block most annoying checks by advancement of a tiger or running the king forward.

Why is a pawn allowed to promote on the 12th rank? Because of its general lack of mobility and special ability to not facilitate double capture. Go-betweens suffer from similar lack of mobility, and share the special ability of pawns. Although able to retreat, this is rarely a good option in the endgame (quite useful in the opening), and so for me it makes complete sense that they promote on the 12th rank as well. Stronger pieces could potentially (and much more likely) do defensive work undefended, so it would be unnecessary to grant them another way to promote. Go-betweens are more or less classed together with pawns. I feel they should be treated the same in regards to rules. On the board is a different story...

Go-betweens and pawns are the only two pieces that can never avoid hit-and-run capture on an open board by a lion that moves adjacent to them.

dax00

So, I played through a K+BT+BT vs K+FK endgame, starting from the worst conceivable position (for the tigers side) that a good player might have reached. I tried my best to play the best moves possible for both sides, although there surely are at least a couple errors. I assumed that no single repetition was allowed. In the end, it took 261 moves to get the first tiger promoting to stag.

Starting position


Even a minor improvement in starting position easily cuts off 20-30 moves, so I have no problem with the 199-move rule as is. If you get stuck with such a terrible position when needing to promote in an otherwise winning endgame, it's your fault if it's drawn. I'll test out K+FS+FS vs K+FK to see how long it takes to win. Edit: Comfortably within the move limit - just get to the edge of the board and it runs like a conveyor belt.