Perpetual check in Shogi?

Got this from Wikipedia:

"If the same game position occurs four times with the same player to play, the game is considered a draw...However, if this occurs with one player giving perpetual check, then that player loses."

Couple of questions:

\* Is this implemented here on Little Golem (and in both Shogi and mini-Shogi)?

\* How is the rule traditionally interpreted for the definition of the "same game position"? Can the same board position (including pieces in hand) repeat several moves apart or must the sequence be "in a row"?

\* In tournament play, will the player be given the opportunity to change their move or is it an automatic loss (presumably after the opponent makes a challenge?)?

I know from experience that "perpetual attack" rules can sometimes be contentious if one of the players' chooses to nitpick, partly because it's hard to precisely define the relevant rules. Those who know Chinese Chess but come from a Western Chess background will know what I'm talking about. In my opinion, this rule is relatively cleanly dealt with in Western Chess. But it is most satisfying in Chinese Chess where there are no penalties for attempting to make the 3rd attack/check - you simply are not allowed to and are forced to change move. Note I use "attack" because in Chinese Chess, the piece attacked need not be the king (though usually is), just any piece weaker than its' attacker.

I'm answering the second and third questions.

Can the same board position (including pieces in hand) repeat several moves apart or must the sequence be "in a row"?

The answer is the former.

In tournament play, will the player be given the opportunity to change their move or is it an automatic loss (presumably after the opponent makes a challenge?)?

They shall not be given opportunity to correct their move. Making an illegal move means an immediate loss. The player who made an illegal move usually resigned the game soon after they realized they made it. According to the rules, the other player can ask his or her opponent to confirm if he or she made an illegal move with stopping the game clock if the opponent made it.

What about this game

[game;id:1351448;move:52]

I suppose it should be a win for Aganju. There was a sequence of checks but no move of Aganju was forced he always had 2 or 3 options to chose from. Thats why I dont like the loss for giving check in the game above, as it was not entirely my fault. :)

Now when I look at the game. It is Aganju who repeated the position 4 times (28,36,44,52) but he didnt need to. Is that why it was a draw? If it would be me who repeated the position 4 times (giving check at every move of the sequence) it would be a win for Aganju?

What if not every move of my move sequence is a check but the repeated one is? Is it a loss for me or a draw?

What if because of check (or Perpetual check) player is forced to repeat the position 4th time. Is it still a draw or a loss for a checking player who forced the repetition but not make it himself?

I think kfiecio should have lost this game due to perpetual check four times in this case. The same position occurred four times at the 27th, 35th, 43rd and 51st moves. And all the moves he made from the 27th to 51st move were in check.

Thanks for the excellent response, takodori-san.

Great example game, thanks kfiecio! Looking at that game (#1351448), the first player (kfiecio) repeats on moves 29,37,45, but not move 53 (which never arrives). It is the second player, responding to the checks (Aganju) who repeats on moves 28,36,44,52. takodori is incorrect that move 27 repeats move 35 because the Gold general is on c5 on move 27 and d5 on moves 35,43,51.

Therefore, according to the proper rules, Aganju should lose (even though he is just responding to check). Why is the game a draw? Did it finish automatically?

I think this rule has been incorrectly programmed on Little Golem, possibly following International chess rules. I shall contact Richard and ask him to investigate and direct him to this forum. If you could please look at the game again, takodori - you know the rules better than anybody else!

Re-reading my post, it looks like the game should indeed be a draw! Aganju repeated on moves 28,36,44,52 and was not giving check when repeating, therefore the game is a draw. If move 27 had been a repeat of move 35, then kfiecio would have lost because the repeated position would be giving check.

takodori, is there any relevance that kfiecio made checks for every move from 27 to 51? I thought it was a "repeated position" rule only, not a "perpetual check" rule?

In Western chess: Perpetual check, any number of times = draw; repeated position, 3 times = draw

In Chinese chess: Perpetual check / attack, 3rd time = illegal, irrelevant whether repeated position or not; repeated position (not check / attack), 3 times = draw

In Japanese chess: Perpetual check, repeated position, 4 times = loss; repeated position (not check), 4 times = draw

Is that correctly summarised?

Just wondering if it would be possible to standardise these rules between the chess variants?? Does the character of the game demand the specific rule that we now have?

Personally, I think the Western chess rule is the simplest to implement and understand. The Chinese chess rules are more complex but it is satisfying that a player cannot "spoil" a lost position by finding a perpetual check to draw (others may feel differently, of course!). In Japanese chess, the rules seem like a more confusing version of Western chess - why a loss if its perpetual check? Why is it "4 times"? Surely it would be reasonable to make that "3 times" to bring it more into line with Western and Chinese chess?

Lots of questions...experienced Shogi players probably never have these problems, but you can see how it is confusing for beginners!

Sorry, my lost post has an error here as alain pointed out.

> The same position occurred four times at the 27th, 35th, 43rd and 51st moves. And all the moves he made from the 27th to 51st move were in check.

Yes, at the 27th move, Aganju's Gold was located on 3a. Then, I think that it was correct to end up with draw since the same position appeared four times but Arganju's check was made only three times, not four times. The game is really an interesting case.

@alain - Formally, in western chess there is no "perpetual check" rule. But, in practice, perpetual check will always lead to 3-fold repetition or to 50 moves rule, therefore, leading to a draw. Furthermore, these draws are not automatic but must be claimed by one of the players.

Sorry, I found myself misunderstood the rules of perpetual check and repetition after checking "Shogi Guidebook" published by Japan Shogi Association, which is considered a canon. The attacking side should make a non-checking move before the same position appears four times. So the case of #1351448 kfiecio vs. Aganju should have ended up with kfiecios's victory since Aganju did not make any non-checking move between the 28th move and 52nd move until the same position appeared four times.

I was the one giving checks So I should lose. :) Is this "Shogi Guidebook" only in Japanese or is it an English version too?

The problem with perpetual check is that it is not well defined. After wikipedia "In the game of chess, perpetual check is a situation in which one player can force a draw by an unending series of checks." It is not very formal definition. From takadori post I think that in official shogi rules it is something like "Sequence of checking moves leading to position repetition." If so then It is indeed wrongly implemented here.

Another question If player A forces player B to make the 4th position repetition but not by a perpetual check this is a draw, not a loss for player A?

How about a stalemate, Is it mentioned in "Shogi Guidebook"? I was interested how it is implemented on littlegolem. It is a loss for stalemated player

[game;id:1351470;move:91]

It's only in Japanese. But HIDETCHI translated it on-line here.

http://81dojo.com/manual/index.php?Rules%20and%20Manners%20of%20Shogi

I don't like that an illegal move loses, when the illegal move is only "illegal" because of an arbitrary number of repetitions. I think the server should disallow the move or have the game tell you the number of repetitions, instead of needing to check the game history.

Is it possible to have a scenario where the only playable move would be a check that is a 4th repetition resulting in a loss?

@takodori: Thanks again for that expert opinion. I'm a little confused, though. Why should kfiecio win the game? Aganju did not make any checks between move 28 and 52, so he had no need to "make a non-checking move"?

@MRFvR: Thanks for that clarification. I think you are right - there is no "perpetual check" rule as separate from the "3-fold repetition" rule, more of special case. The rules are more complicated in practical situations, though! Formally, in order to claim a draw under the 3-fold repetition rule, the player making the claim must be notating the game (or else the game must be recorded).

When the game is not recorded, if there is a dispute about the repetition, it is not possible to end the game under the 3-fold repetition rule. Sometimes you can use your opponents scoresheet (if they will let you!). But if a dispute arises [for example, where scoresheets differ] it is usual for the draw claim to be rejected. When a 3-fold draw cannot be claimed you should still be able to claim a draw by perpetual check! Either the opponent will concede that they cannot escape the checks or the player is able to demonstrate to the arbiter's satisfaction that their opponent cannot escape the checks. I've seen situations where the arbiter will actually give the claiming player additional time and then restart the game in order to make them play demonstrate the perpetual check in practice. This is interesting because the player would certainly lose on time nor formally claim a draw by 3-fold repetition because he is not notating...but he is still able to get a draw!

@kingofthebesI: There are, unfortunately, situations where an illegal move loses. For example, in rapidplay western chess, leaving your king in check (after pressing your clock) is usually a loss or other severe penalty. I agree with you its unfortunate, but it is probably necessary in a practical sense. If someone wanted, they could continuously play illegal moves and there has to be some sanction in place to prevent this. As long as the rule is the same and is known by both players, there shouldn't be a problem. Of course, if the rule is bad, then it should be improved.

Regarding the situation where a player may unwittingly enter a sequence where they are forced to repeat the position for the 4th time, and be giving check each time, then I suppose its fair that they lose if those are the rules. They would just need to be wary of allowing themselves to fall into that trap. It would be very rare, though - I cannot construct a position where this occurs, though no doubt it is possible.

@alain - It's my job to be right as far as chess rules are concerned (obviously, it doesn't mean I'll always be, though). As regarding your answer to KoT, although I'd agree with you for OTB games, I believe KoT was referring himself to games played over here. In this context, I do believe server should prevent players from making a 4fold repetition losing move.

@KoT - I believe one can construct such an example, but it will probably be far-fetched and involving 'stupid' game play. I may get back to you on that, but I'm thinking on the lines of a player being obliged to drop a bishop to cover a rook check.

@Alain

True leaving the king under attack CAN lose in chess, I very aware of how chess works, however online you wouldn't be able to leave the king en prise and would never be deliberately leaving your king under attack.

Just looking at the board should tell you, if you left your king under attack. The only way to know if a shogi check is safe to make is by looking at the game history carefully if game is reasonably long!

The issue of verifying Repetition in OTB chess would apply to OTB Shogi illegal check, only it could be more contentious since the swing in result is potentially the full point.

@MRFvR Yes I'm referring to the server, OTB tournament you have an inkling how many times you've repeated and have one game in progress.

I think I envisage the scenario which would be to a forced losing check but it would be implausible to think that it would be repeated.

If the 4th repetition with check was prevented by the server and would leave a player no move then resignation could be the only option as happens when victory can't be broken at WYPS!

We fortunately don't implement touch move for games that have that as the rule!

@MRTvr and kingofthebesI: In OTB play where a time scramble occurs, it is quite common for one player to want (and often to try) to claim a draw by repetition. Unless the game is being observed by the arbiter or recorded this never succeeds. This is an example where "perpetual check" is your friend! :o) I realise that this does not apply to games played here at LG where there is no "time scramble" and the game is always recorded.

@MRTvr and kingofthebesI: Thanks for the opinions. I'm not convinced that the server should prevent you making a losing move. If you went down that road, should it also prevent you playing really, really bad moves as well? ;o) That comment is a bit tongue-in-cheek, but my point is that if the rule states that the 4th repetition is a draw (if not giving check) or a loss (if you give check), then its up to you to decide whether you play that 4th repetition. It should not be up to some external party (the server, the arbiter, the spectators, etc). If your opponent has cleverly constructed a position where he is able to force you to repeat 4 times while giving check, then credit to them. In this case, why shouldn't they win the game?

After reading all the responses I feel the rule that "you lose if the 4th move repetition is giving check" is strange (even confusing), but would be rare in practical play. Not really worth bothering about unless you decide to take Shogi more seriously. Just one of those quirky rules, like the "en passant" rule, that you have to live with if you want to play the game.

The confusion that exists over the Shogi repetition rule is also in Xiangqi. I've known several strong Xiangqi players who are unable to satisfactorily explain the "perpetual attack" (usually "perpetual check") rule. For those unfamiliar with Xiangqi, in most circumstances you can distil the rule down to: "You are not allowed to give check more than 3 moves in a row". This will cover almost all practical situations. The "draw after 3rd repetition" rule is identical in Xiangqi and western Chess (if the 3rd repetition is not an attack on a more vulnerable piece, of course!).

Perhaps because I grew up with western Chess, I find the "draw after 3rd repetition" to be relatively easy to understand and therefore explain. That doesn't mean its a "better" rule, though. In my opinion, the Xiangqi rules on this matter are the most satisfying.

@alain: The thing is, making the 4th repetition with checks is not a bad move, it is an illegal move. As illegal as moving your piece the wrong way or twice in a turn. Since the server doesn't let you do those moves, I don't see why it should let you do the 4th repetition check.

@JeanHerbert: OK, thanks for that clarification. I see the error in my understanding now. Do we have a contradiction, though? The Wikipedia article states that the 4th repetition with check "is a loss" (no mention of illegality) but you're saying is a loss "because its illegal". Maybe the wiki needs clarifying? One reading of the wiki rules is that you should allow the move to be played on LG, but then the player loses (which is how I read the rule). Given what you and takodori say, the move should not be allowed to be played on LG at all! It really would be interesting to see how the server handled kingofthebesI's example where the player is forced to make the 4th repeat with check...but can't, leaving him no legal moves at all!

In the western Chess tournaments I've played in, a first instance of an illegal move is never considered a loss. Instead the player gets a penalty, for example, a warning and 2 minutes added to the opponents' time. Subsequent illegal moves either get additional warnings and time added or can result in a loss.

Maybe thats the way it should be handled on LG - let the player click any piece and place it absolutely anywhere they like. After they click "Move", if it is illegal, they get a warning. After two warnings, they lose! :o)

Sometimes practical OTB play compared to play on servers like LG can be quite different!

@alain - in that case I really would have to read Tori's rules. And just for the record, in Blitz games one does lose after first illegal move. As a principle, I would see as a good idea that in any serious (pro or semi-pro) one illegal would mean a defeat regardless the rate of play. But, when novices are involved (specially kids) that sound rather harsh.

@MRFvR: True, in blitz games an illegal move will usually result in a loss. I suppose with very serious or professional players, maybe an illegal move in slow-play games will be a loss, I just don't see illegal moves often enough not know the precise FIDE rules to know either way. Typical I witness an illegal moves when the player(s) are involved in a time scrambles to get to the time control or towards the end of the game. In those circumstances, accidentally touching a piece that is pinned is not uncommon - now you've got to move the piece but cannot...in the events I've played in, some quite serious, that is usually not a losing offence. Typically time is added to the opponents' clock.

I've seen some unsporting behaviour over the years, though. Once I was playing a young player who played fine for most of the game, but in the endgame started leaving his rook on the line between two squares (and immediately pressing his clock). Now you have to clarify, "Is that on f3 or f4" on your own time. If you complain, are you "distracting" your opponent? Is what he doing actually illegal? Etc. An interesting situation for the arbiter!

I definitely agree that in casual games or ones involving kids an illegal move should NOT result in a loss!

The Shogi example is interesting. I'm hoping someone will construct a games where the one player is forced to repeat, with check, and therefore cannot actually play! :o)

Touch move does not apply, if it would force you to make an illegal move at chess (I'd hope the same would apply in shogi/xiangqi), although if are continually touching pieces that can't move it might be seen as unsporting and offputting.

Are the pinned pieces pinned against the king in the scenario you describe or are you saying people aren't abiding by touch move because the realise they would lose a major piece by moving?

1st illegal move time penalty, 2nd one time penalty and 3rd illegal move is a loss for classical chess!

If you have a dispute you should stop the clock and say you are getting the arbiter or ask the opponent to adjust the piece if it is a one off you. While you are always entitled to call an arbiter, it can leave a bad taste to be too pedantic! One thing to note you aren't supposed to adjust an opponent's pieces however off centre they are! I heard firsthand, that GM Jonathan Rowson (then IM) got the arbiter once at the British Championships about 14-16 years ago over an opponent placing pieces too off centre.

I was awarded extra time last year as a result of continual sloppily placed pieces when the arbiter stepped in after my opponent got too sloppy in a time scramble!

It should be doable to construct an example of enforced suicide by check, what would be of interest to find one scenario that made sense to happen!

When I played blitz chess as a child it wasn't a problem for an illegal move losing, we knew the the rule and accepted it. The 4th repetition being a loss is more understandably upsetting as it brings in a disputable game history. What make the shogi scenario so tricky is that you have a move that 99.9% of the time is perfectly legal being penalised with an immediate loss.

A time penalty for a reasonable mistake seems more fair than immediate death!