Rule 10.2 (a) AGAIN then

[Jonathan Rogers]

This perhaps prompts even more internet discussion than "who was the greatest world champion?" and "why can't I eat my sandwiches in my anorak in the hotel toilets?". But I just saw the Southend thread in "International" last night and reluctantly decided to open a new thread.

The main bit provides:

10.2 If the player, having the move, has less than two minutes left on his clock, he may claim a draw before his flag falls. He shall summon the arbiter and may stop the clocks. (See Article 6.12.b)

a. If the arbiter agrees the opponent is making no effort to win the game by normal means, or that it is not possible to win by normal means, then he shall declare the game drawn.


In practice it is the second part - that it is not possible to win by normal means - that is quite inconsistently applied. it is rather depressing that it should be so because I think that in most cases there is clearly only right answer. It seems to me that one must always allow the side with rook and knight v rook to at least try to win, and that must apply to R + N v R +P as in Jones v Chernaiev, since the pawn should be won easily enough. Actually Chernaiev was worse off through having the pawn, because that would allow Jones 20 moves to win the pawn before another 50 moves to try to win, but that is neither here and there - the position can still be won normally and the fact that it might be practically impossible to defend with severe time shortage is the responsibility of the player who got so short of time in the first place. 10.2 is only meant to protect those who have such a clear draw that even a beginner would not mess it up; where there are no tricks for the stronger side to try to test his opponent with; where the weaker player could only lose by making a blunder so gross that it would be hard to imagine it happening even with his flag close to falling.

But the arbiter at Southend is not the first to completely misunderstand this. In Sands v Taylor, Essex v Kent 2007, Black stopped the clock when he reached R v R+N, the position was sent to the arbiter and he too gave a draw for the side which had not even attempted to defend it. I also remember a very unpleasant incident in Lancashire v Nottinghamshire 1991 where a junior player was under pressure to agree a draw in this ending on account of his opponent's time shortage - the pressure coming from someone who was an arbiter, though not acting as such on the occasion.

I had hoped, until reading the Southend thread, that such incidents were in the past. In McDonald v Kwiatkowski, Kent v Sussex, just earlier this year, White was correctly given the win when Black stopped the clocks to claim as soon as the ending was reached (where no arbiter is present, a wrong claim must lose the game at once). There was also a correct decision in Reid v Sands, Hackney v Ilford in late 2012 where the following position was reached:

White: king c4, bishop h5, pawns f5, g4
Black: king g5, bishop a7, pawns h6, g7, f6

Despite being two pawns up, it is clearly headed for a draw but white claimed it at once without allowing black even to try for any tricks. And there are things that can be tried. One can transfer the bishop to c3 or maybe f8 and then play ...g6. White then has to decide whether to take on g6 with the bishop and allow Kxg4 and ...h5 etc, or to play fxg6 when Black plays ...f5! with new problems for White to solve. Probably white can draw it without too much trouble but he has to go that trouble, and the arbiter rightly awarded the game to Ilford since the position was not yet exhausted.

But after Southend, it feels like we are at square one again. If the pattern I have observed is at all typical - and if I do say if - does this mean that would-be arbiters and present arbiters at the lower levels may need more training?

[Neill Cooper]

Almost simultaneously I have posted in the Southend thread a copy of the advice written by David Welch (ECF Chief Arbiter) which was lost on the old ECF website. You can find it at http://www.ecforum.org.uk/viewtopic.php ... 45#p112043

[Ray Sayers]

To be perfectly honest, I wish 10.2 didn't exist. It's a right palaver. If you run yourself short of time you should expect to lose on time. It's part of the game (or was).

One day maybe everyone will play on increments and it won't be necessary!

(awaits howls of protest)

Edit - spelling

[Martin Benjamin]

There is an easy way to end 10.2 debates, which is to play with increments. James Toon's excellent Central London League is a good case study. (Highly recommended for any clubs looking for a league). The league is nearing the end of its second season, and unless James tells me otherwise, I don't think there has been a single dispute, and even more importantly, I don't think there has been a single case of a game going beyond closing times for venues, which has been raised as a concern. The time control is 80 minutes per player plus a 10 second increment for each move from move one, which some hasty mental maths tells me guarantees a minimum of 60 moves in a three hour session, and then a minimum of 30 moves in every subsequent 10 minutes (so at least to move 150 after three and a half hours). If the time control is adjusted to ensure (say) at least 150 moves, finishing times are not a problem, and the rules can always stipulate adjudication for extraordinary cases after either x moves or y playing time, whichever is the later. I am surprised that tournaments still use guillotine quickplay, as it would take away a lot of stress for arbiters. I just can’t see any convincing arguments for retaining guillotine finishes which will sometimes lead to 10.2 disputes.

[Graham Borrowdale]

To be fair I have not seen an account of what happened at Southend, but I am assuming that the player with the R & N, being a strong GM, was attempting to win and not just going round in circles. In a lower level league match you might well see the rooks go round in circles and perhaps then an arbiter might be justified in awarding a draw. And how many times can the rooks go round in circles before a draw is awarded (always assuming the 50-move rule can not be invoked)? And does it make a difference if white is a GM and might have a better idea of how to try to win than a mere mortal?

It does sound as though the use of increments could stop these problems, as Martin suggests. Again, I think it makes a difference what level we are talking about. For a professional playing in the Southend Open with a lot of money at stake it is vital, but in local leagues players often 'agree' results because of an implied 10.2. For example, in a game with Q v Q+P and less time I checked the opponents K round the board a couple of times until he was happy I had a good idea how to draw and he agreed a draw before I made a claim (but I did not claim as soon as the ending was reached). With 10 second increments we might have been playing for a long time - with the guillotine we were virtually blitzing the moves.

[Roger de Coverly]

I just can’t see any convincing arguments for retaining guillotine finishes which will sometimes lead to 10.2 disputes.

There are two very obvious ones.

The first is that digital clocks are by no means universal.

The second is that Congresses with multiple rounds in a day have schedules to stick to. e2e4 use 30 second increments so lack of scoring isn't a problem, but about once every five or ten Congresses they have to postpone the next round because the pairings cannot be done because of a game in progress. If you use increments of less than 30 seconds, the tournament arbiters need to know how to stop games where the 50 move rule would apply under normal circumstances.

In e2e4, there's an (informal) rule, that players in a late finishing game can request a modest postponement of their game in the next round.

[Andrew Zigmond]

Interestingly at the 2012 Harrogate Congress a near identical situation occured between two very strong players (both are well known on the congress circuit - I won't name them). With K+R+N vs K+R+P the player with the pawn lost on time having had several draw offers declined. The player with knight objected on the grounds that he was making progress. I believe that the win on time stood. My own last round opponent - the only other player remaining as our game had just finished - said to me he considered the player with the knight to have been quite unsportsmanlike. I disagreed.

In quickplay finishes too many players view the clock as an adjudication device and think `the position is drawn and I'm about to lose on time so my opponent is obliged to accept a draw` or `my position is won and I'm about to lose on time therefore he is obliged to resign`. It isn't. You are allocated a certain amount of time to make your moves - if you overstep the time limit, tough. Obviously it's different if the position is theoretically drawn or if it can be proven within the scope of 10.2 that the player is not making any reasonable effort to win by normal means.

The problem is that what's correct often goes against what is sportsmanlike.

[Roger de Coverly]

In quickplay finishes too many players view the clock as an adjudication device and think `the position is drawn and I'm about to lose on time so my opponent is obliged to accept a draw` or `my position is won and I'm about to lose on time therefore he is obliged to resign`. It isn't.

It would be my expectation that if the position is "obviously" drawn, then the draw should be awarded. Equally the player "obviously" winning should be able to settle for a draw if they don't consider they have enough time to win. As a benchmark, ask what the result would be if a delay or increment clock were substituted. The point being that you try to rule out players continuing where their only realistic expectation of a win is on time.

R+N v R or R+B v R are difficult for the arbiter. The Laws appear not to allow it but consulting the endgame tablebases as to whether the position is drawn or won with best play, would at least inform their decisions. Equally with unlimited time, the player with the material has 50 moves to attempt to win, An opponent's shortage of time ought not to deny this option. The quoted precedent of awarding a draw in K+R v K does muddy the waters somewhat.

[Paul McKeown]

10.2 is clear and entirely fair; arbiters making occasional mistakes is a poor reason to abandon them. I personally loathe increments, essentially you are in time pressure from move 1, and a game may actually not end until one of the players keels over dead, if the increment is too short to allow the recording of moves in order to claim a draw.

[Dan O'Dowd]

what's correct often goes against what is sportsmanlike.

To my mind, this sentence is a total contradiction though quite revealing. What I hear from it is that you feel that the laws and their correct application as it stands, allow some situations where you disagree with the validity of the result (or that a majority of players would do this). If the correct result is derived from the laws, the result cannot be unsportsmanlike, by definition, whether a player feels aggrieved or not, surely. Could you give some concrete examples?

It is correct that a player is responsible for his position and clock. But be clear, an arbiter does not know what a theoretical draw is. He simply sees the position available and uses the play proceeding as evidence for or against the claim. By the way, as you may be aware, if the R+P player in your example was offering multiple draws, I can only presume that his opponent mentioned this to his arbiter since it is forbidden and highly distracting to offer consecutive/repeated draws especially in such a situation.

Roger, it is not correct to use tablebases as an arbiter and among the reasons is the fact that the players have no such access. The claim is not being made to the correct chess - it is being made to a fellow human. At no point should the arbiter treat the claim as an exercise in one player showing his ability to play more correct chess than the other. It is logical to say that since there are GMs who fail to mate with B+N, there will equally be players of all levels who cannot hold R v R+N even if that one is classed by experts as relatively simple. The problem with your expectation that an 'easy' draw be awarded is that an arbiter does not know what the term easy refers to and different players even of the same grade will also here vary. Presuming he has adequate knowledge/understanding, he will observe the play from the claim and should base his decision on that alone. Yes, this is not much to go on, but it is the best way to avoid a bias towards one or other player. As for a player 'winning' (ahead materially/positionally, for clarity) being able to settle for a draw, if by that sentence you mean that a draw should more loosely be awarded I must disagree (though I am not sure which way you mean it). This is unfair to the plaintiff if the position is sufficiently live that play can continue reasonably for some time. Remember, it is expected that a claimant has left himself enough time to play many moves under the arbiter's gaze. Why do you expect that an 'obvious' draw be awarded, and where are your boundaries for this term? I presume we agree up to the trivial K+R v K+R where neither side is in trouble, but where higher would you expect draws?

I was arbiting the Cumbrian Championship in March and a player claimed a draw in a game where he had Q+R+6 pawns against Q+N+5. The game continued for a while before the defender offered a draw himself, but this was only after about six moves each, with no exchange of pieces. The pawn structure asides the extra was symmetrical but had the claimant's flag fallen I would have been forced to award a loss, since I had no positive evidence that the defender was shuffling.

I would also argue against making the claim that we should seek to stop players from having the chance to win on time, provided they are making effort to play the position. It is of course universally known that shuffling the pieces to and fro does not cut ice, but if the player is able to do something or prolong the game (imagine as if an engine was a minor piece down), it is up to the draw claimant to reach a position from which the defendant cannot win by normal means.

Finally I am highly surprised that seemingly none of you are aware of the pending Laws change for July which indeed includes provision for self-determined results in 10.2 claims. The Law will subsequently read:

"...The arbiter may decide that the game shall be continued using a "time delay" or "cumulative time" mode. The extra time shall be 5 seconds added for each move for both players. The clocks shall then be set with the extra time and the opponent shall be awarded two extra minutes."

The order is now:

to accept the claim.
to postpone the decision.
to continue the game using an increment.
to reject the claim.

Incidentally a full review of changes can be found here: http://www.londonchessclassic.com/fidel ... s_2012.htm

[Richard Bates]

There was also a correct decision in Reid v Sands, Hackney v Ilford in late 2012 where the following position was reached:

White: king c4, bishop h5, pawns f5, g4
Black: king g5, bishop a7, pawns h6, g7, f6

Despite being two pawns up, it is clearly headed for a draw ...

I think this is a bad example because i think that you've got the position wrong? Are you sure Black didn't have another pawn on the Qside or something (probably on c5)? I remember we looked at it after the match and black kept winning! Whereas this position just looks like a dead draw if white runs the K to h1? (although i may be missing something obvious in my head) - to the extent that a draw award would probably be fully justified IMO. The plan with g6 is just nonsense because of the wrong coloured rook's pawn.

[Richard Bates]

More generally i think that people citing the 50 move rule in relation to 10.2 claims are creating a bit of a red herring. The 50 move rule is an artificial rule created to stop games going on for ever, they should have no bearing on decisions under 10.2. If it was the case that anyone with R+N v R had 50 moves to try whatever they wanted then no 10.2 draw claim would ever be given in these circumstances. Because the 50 move rule would kick in concurrently (assuming anyone was even recording the moves). The question is where you draw the line, on how long to have to play to demonstrate that you can draw it. R+N v R is not in the same league as R+B v R. In most normal circumstances I think grandmasters would agree draws immediately, or at most after half a dozen moves. I think it should be only mildly difficult to draw if the K is cut off on the edge of the board. (he says, fully expecting to lose it comfortably next time he gets it...) Ideally there probably is a case for taking the respective and relative strength of the players into account, but of course it is very difficult to write this into the rules, and puts pressure on the chess ability and knowledge of the arbiter.

[Chris Rice]

"There was one case of ‘…making no effort…’ when black with lone K on e5 claimed a draw against a K on f2 and a rook on a4. Play continued 1. Rb4 Kd5 2. Ra4 Ke5 3. Rb4 and black’s flag fell. The arbiter correctly awarded a draw. Note that the artificial prolonging of the game is the key point of the decision."

In my opinion this is an incorrect interpretation. Their is no artificial extending of the game when players repeat the position to gain time on the clock, work out the winning method etc so if they do are they immediately to be penalised by their opponents claiming a draw? Further I don't doubt in that endgame that the guy was trying to win on time, so what? They have handled their clock better so why should the opponent be allowed to run themselves out of time and yet be rewarded for it?

Suppose it was K+N+B v K and I play two repeat moves with the N & B because I'm not sure how to win it, is the arbiter expected to know how to mate with N&B before he can decide whether I'm legitimately playing for a win?

I'm just about to play IM Tania Sachdev in Rd 3 the Dubai Open. This is the third time I've played this tournament and they always play 90 mins + 30secs increment from move 1 and I can't ever remember a problem like 10.2 occurring so that's obviously the way forward IMO, finances permitting of course.

[Andrew Bak]

My view on all this is very simple - if you lose on time that's your own fault, no-one else's!

10.2 should only be invoked as an exception to the above rule, where a draw is only possible result that the opponent can obtain, or if it is absolutely clear that the player winning on time is making no attempt to win over the board.

I recently won a time scramble where I had K and R vs K, N, B and P, I saw chances for swindling and fortunately everything worked for me - strange things can happen in chess sometimes! I only played on from the initial position because I had to win to draw the match for our team.

[Roger de Coverly]

Suppose it was K+N+B v K and I play two repeat moves with the N & B because I'm not sure how to win it, is the arbiter expected to know how to mate with N&B before he can decide whether I'm legitimately playing for a win?

The arbiter should only need to know that the position can be won. Counting moves to 50 is sensible and an entirely rational way of awarding a draw for progress not being made.

K+R+B v K+R or K+R+f+h v K+R are more difficult, since while there exist drawn positions, not present in K+N+B v K, playing them in practice is non trivial.

The addition of a definition of normal means has been proposed. That is that the player seeking to win has a reasonable expectation of being able to win other than on time.