'Refined Chess' proposal – readers' feedback

Reader Feedback on Refined Chess

The following is a selection of letter we received as a reaction to the article "Refined Chess – a proposal to combat draws", by Ali Ferhat Tamur of Ankara, Turkey. They are roughly in chronological order, and the ones left out were mainly those with rude language, adventurous orthography or ambiguous subject lines (the number of letters we receive make it necessary to group them in a semi-automatic process).

We would like to inform our readers that we will be closing the debate after this feedback report – there are other debates raging and we do not want to burden visitors with too many opinion pages. The subject is bound to come up again with some new, ingenious proposal, so please wait until then to express your views.

Edward Witten, St. Genis, France
I see a problem with this proposal. Consider an endgame such as K+B vs. K+B. We really would like this to be a draw. But, as there are only a finite number of positions on a chessboard, it is inevitable that repetitions will eventually occur as play continues indefinitely. So the endgame will be a forced slight win for one player or the other. However, a chessboard is large enough that play can continue for many thousands of moves before a triple repetition. Do we really want players to play on for thousands of moves in an endgame such as this one in order to secure a slight win? The refined chess proposal could be modified with a 500 move rule (a draw is declared if 500 moves are played without a pawn being moved or a piece captured), but it is inelegant to have such an artificial rule in a position that really is a forced slight win for one of the players.

Jorge Rivera, Grovetown, GA, USA
I would change something in the author's idea: I don't understand why 2/6, 3/6 or 4/6 of a point; it doesn't make sense since there is no situation where a player would get 5/6 of a point or 1/6 of a point. My change would be to give a slight win 3/4 of a point and a slight loss 1/4 of a point. The draw stays at 2/4 or 1/2 of a point. That would be my suggestion, even though I think this idea needs way too much explanation and rules to cover every possible aspect of the game and leaves too much to an arbiter's judgment of a position.

Kajetan Wandowicz, Wroclaw, Poland
I think the rule proposed has several disadvantages. Most notably, there is a risk that complex-but-winnable endgames would no longer be played if the superior side was able and keen to force a slight win. Secondly, draw by threefold repetition is often a form of a silent mutual agreement, i.e. both players are happy with it; it would be unfair not to distribute the points evenly. Moreover, the statement that "every rule is the same except for those concerning the end of the game" is somehow misleading, since changing the rules of finishing the game chances strategy vastly (as, by the way, Mr. Tamur admits later in the article). I also can't agree with the long list of advantages – most of them shouldn't be there, since "grandmasters remain grandmasters" or "current knowledge is still good" are more of obvious, healthy prerequisites to any change of rules, and not some special, unique advantages of Mr. Tamur's proposal.

In my opinion, no rules must ever be changed. Chess is chess. That being said, I still think chess variants are a good thing for the development of the game, but the bottom line is that if any of them proved good enough, the players would simply switch. It is not forbidden to organise a Chess960 tournament, or to organise a pseudo-chess tournament with Mr. Tamur's rules. If this tournament proved successful and organisers started to switch to those rules in other tournaments (which I personally doubt), and the top players still happily attended them, the proposed rules would automatically become main rules. And this would be the right way, by natural evolution followed by FIDE regulations, and not the other way around. Still, I believe those rules would have very little chance of being accepted this way, while they might have a chance to find some support in money-driven FIDE.

Mathematical side-note: it is not true that "the result of a perfect game is believed to be a draw". In fact, it's quite the opposite. Since every game of chess is finished in finitely many moves, mathematics says one of the sides can force at least a draw. This is because of a theorem that states that in this kind of game, either there exists a winning strategy for one side (let's say for White), in which case White forces the win, or not (so Black can obtain at least a draw). This was sub-proved by Ernst Zermelo in 1913 and proved fully by Denes Konig in 1927. This winning strategy (or even the lucky side, for that matter) is simply unknown yet, due to the enormous complexity estimated at 10^120 possible combinations, and will probably remain so at least in our lifetime. Please don't insult mathematics.

Ralph Dihlmann, Frankfurt
Very interesting propositions, worth trying in pilot tournaments. Maybe there could be an extra room with refinded rules on the Playchess server, to gain experiences with this? I see a slight disadvantage of the refined repetition rule for a player with an active style: after an unsuccessful mate attack, with sacrifice of material, he no longer can reckon with half a point for a “rescue perpetual'. But on the other hand, the player with the superior position would benefit, if the defending player rescues his position with a perpetual.

Nate Plapp, Lemon Grove, USA
There are several issues I have with "Refined Chess". With regard to opening theory: all it does is give advantage to the younger generation as it is not so easy to unlearn what has been learned and burned in with thousands of games. The beginners won't have to. Computers would invent all new theory almost instantly – everything would be called the Fritz 11 Gambit or the Rybka 3 Attack, if they were given credit for their ideas. It also makes all the chess literature, if not obsolete, horribly flawed. This also means that it will likely disappear. There is no way such a fundamental change will be embraced, leading to an almost certain division even if FIDE went along with it.

Chess has had a long road toward global acceptance. Switching to some random variant invites all those countries who have turned their focus from their indigenous chess variants to rethink that decision. This could also make those who have worked hard to master chess in those countries to have no choice but to leave or starve.

Games still will not be played out because it will still be clear who will be doing the stalemating far before amateurs in the audience realize it.

Mr. Tamur's argument has a straw-man element as he ignores several good solutions that were given by earlier respondents that only involve the clock; such as the bidding by time advantage for white and getting draw odds as black as they do in Chinese chess. The bidding process could add undue complications. Just having 1/3 the time when playing black and playing a pair of games, one with white one with black, without the draw odds thing would still be effective. In this situation a draw is the last thing a player wants as white, because it is not so easy to get the draw when he or she has their turn with black. Further, quality should still be there as at least one side has all or most of the time we traditionally have.

Alternatively, there are a number of good reasons for simply having faster games: if we have mini-matches of eight faster games, the mistakes should even out and you can have a separate match and tournament rating. Match rating would see any match result as 1-0, ½-½, 0-1, and tourney rating would be whatever the result was counting each game as a whole game. The fast format also makes cheating virtually impossible because you can only use the restroom between games and must stay at the board while a game is going. A video camera above the board can be used to settle any disputes and substitute temporarily for recording the moves. Held high enough it could cover several boards simultaneously. Even the most drawish players are likely to have at least one decisive speed game in eight; so there is always something for the organizers and audience to appreciate. Also with fast time controls, the clock as a decisive element encourages the side with the worse position to continue, so amateur spectators can see how the game would end. Pairings are also easier as there is no problem with a player getting more whites than another. And more games just gives the final standings more statistical credibility. Ratings also become more accurate and dynamic allowing us to see who is playing well now rather than having a bunch of players whose rating has little to do with their current strength or possessing ratings just reflecting good/bad recent luck because it is based on so few games. Also currently some juniors have ratings that can't keep up with their improvements, making other players want to avoid them because of the likelihood of loosing points. With tournaments consisting of something like 64 games that becomes less likely. Of course there is a balance here and the ratings formulas can be adjusted, but they would only likely be more accurate regardless because they would be based on more information because there were more games to adjust the rating with. One final advantage is that there can be two separate world titles, one for match play and one for tournament play and if they are to be decided by contests then the separate ratings would inform the organizers who to invite to each type of contest.

Estivallet Natan, Porto Alegre, Brazil
It seems strange to tell someone: "in our club tournament last weekend I had three slight wins, two losses and no draws at all". I think most people still think of the result of a game in "binary" terms, 1 or 0, win or loss. The abolishment of agreement draw and more fast games would bring more attention to chess, making it more a "sport", with more human mistakes like all other sports. Other interesting suggestion (David Bronstein) is to play a "mini-match" of four games (15 minutes) instead of a 2hx2h (in "Secret Notes", Olms Ed., pages 221-222). In his words: "I think that GMs would become less tired, and spectators would gain more pleasure, if instead of one long game in an evening GMs were to play four short games, with an additional 15 seconds for each move". With such events and forbidden draws, most of the problems pointed by the author would have been solved.

Mathieu Buard, Nogent, France
What I liked in that proposal is that it echoed an idea I once tried in an off-hand tournament with friends. That idea is about the "slight win". I did not designed it for combative chess (as I'm just an amateur like the friends in my tournament and so no GM draws occur) but because I never liked the stalemate rule : why should a paralysed king earn as many points as the free one? On the other hand, I wanted to keep some kind of saving manoeuvre. Actually, it is a mix between that idea and the Bilbao rule (which rewards a win), here it is :

What I reward is not the victory but the mate

• checkmate is 1.5 - 0 (like in the Bilbao rules)
• dead draw is 0.5 - 0.5 as in usual chess (and Bilbao)
• but a stalemate is 1 - 0.5 because even if the king is not captured, it was paralysed, so the mate bonus is still rewarded.

I'm not sure it really encourages more fighting chess, but I just wanted to give as a feed back that the idea of a "slight win" has already been tried (I'm sure I'm not the only one who already tested it in one form or another).

Andrea Mori, Torino, Italy
I don't think I'm able to provide a precise reference, but the idea of splitting the point differently in the case of a stalemate has been around already for a while. In my opinion, the worst disadvantage is that players will be less likely to take risks (such as sacrificing a pawn for the attack, accepting the eventuality of an inferior – but drawable - endgame) leading to a less brilliant play. Nonetheless, the necessity of rethinking the opening and the endgame theories would be very intriguing consequences. Thus, I believe that it would be very interesting to hold an experimental GM-level tournament with these 'refined' rules to see their practical outcome.

Johnathan Rothwell, Southport, England
I sigh when I see proposals for changing chess, which seems to happen on a regular basis these days. Chess is still the best game. It has stood the test of time and does not need changing. As for the perceived problems:

1. Drawing – simply do not invite players who give short draws on a regular basis.

2. Sponsorship – difficult, especially in the current financial climate. The problem is partly because computers have replaced humans as the best players. We need to get the characters of the game across to the press – people love characters. Unfortunately chess players are regarded as geeks in most media. We also need to wow the public with huge blind fold displays, this can still wow the layman.

3. Computers, cheating and theory – this is the main problem because they effect the game from a sporting point of view, not the game itself (for people like me who just love the game). Ideally I would simply ban the computers, or the software at least, from being produced. That is unfeasible. I don't know the answer. But we must find a way to rise against the machine!

Dominique Pellé, Amsterdam
This is a very nice way to help solve the draw issue. It hardly changes the rules and makes all of a sudden drawish positions more exciting. A player who forces a draw clearly does it to save a lost position, so he/she deserves to slightly lose rather than draw as in the current rules. It makes sense to me, I like it very much. The score for win, slight win, slight loss and loss should perhaps be discussed further. With the proposed scores for win/slight win/slight loss/loss, three slight losses are equivalent to one win + two losses. Three slight win are equivalent to two victories + one loss. I'm wondering whether that's OK. At the very least, it makes it possible to have equal scores at the end of a tournament for different players even if they had different results for their games, which may not be desirable. How about using irrational numbers for scores of slight win and slight loss? For example: one point for a win, PI/4 ~ 0.78539 point for a slight win, (1 - PI/4)=0.214601 points for a slight loss and 0 points. The idea of course is that it reduces the chances of having identical outcomes at the end of a tournament. With these scores, three slight losses (0.643805) are not as good as one win + two losses (1.0). three slight wins (2.356194) are better than two victories + one loss (2.0)

Julian Wan, Ann Arbor, USA
A very interesting idea – I don't agree with all of the notions, but liked how well organized the author arranged his proposal.

Ramón Jiménez, Santo Domingo, Dom. Rep.
After watching the NBC chess promo (it's great IMHO), I guess it would be interesting to try out a chess version similar to basketball regarding the play timer: 1) Each player gets 24 seconds to make his move; 2) If a player doesn't make a move after 24 seconds, he doesn't lose the game, instead the opponent gets to move again! 3) All other rules remain the same. In a sense this is similar to Rapid Transit, but you don't lose the game on time. And with the potential of someone making two moves in a row, the concepts of threat and combination take on a whole new meaning!

Luis Moser, Costa Rica
My friends, the draw problem has a very simple solution, and I am at a loss to understand why nobody proposes it. The solution: a win is worth two points and a loss one point, a draw is zero points. But why something so simple when there are more complex ways?

Denis Bucher, Switzerland
Another solution: if a player offers a draw and ends up losing the game, he/she receives a -0.5 penalty. This will make draw offers a rather risky strategy!

Rajesh Parvathini, Hyderabad, India
There is a way to force players to play for a win in super-GM tournaments. Last year, there was a tournament which used Bilbao scoring system (3-1-0 system). I propose a slight modification. We know that winning with black is very difficult at GM level. So we should give more points for a win with black because Black has to take more risks than white if he is fighting for a win. Because we do not want to encourage draws, draw with any colour gets same points. So:

This gives more incentive to players with the black pieces (who wants a draw more than White) to play for win.

Dr. Will Denayer, Cork, Ireland
I read Ali Ferhat Tamur's proposal with a lot of interest, but I do not agree with his proposal. I do not think that classical chess is suffering from a number of lethal problems, as Ali asserts. To me, the drawing tendency is not a big issue. Look at the ongoing tournament in Nalchik where some very interesting draws have been played so far, so what is the problem exactly? Lasker went off to play bridge for a while, Capablanca proposed the introduction of new pieces and Fischer proposed 960 Chess, but since then many great and outstanding games have been played, so I guess all three were wrong in their assessment of chess being a draw. Capablanca believed that he was playing perfect chess, but as we know now, even he made mistakes, even in his most famous endgames. All of this was fine until a certain Alekhine came along. Perhaps Fischer wanted to introduce 960 Chess, so that in his mind he could stay world champion forever. As to Ali's fourth objection - opening theory has become too advanced thereby stifling creativity - I do not understand why a N on move 25 is less creative than one on move 5. I find Ali's suggestion of Refined Chess interesting enough, but this is not going to help attracting sponsors as it's going to be difficult to explain that White won a little or that Black lost slightly. However, it is perhaps worth a try and perhaps someone can organize a tournament in which these rules hold so that we can see the results.

William Plants, Campbell, CA, USA
I think the "Refined Chess" idea is simply brilliant! Given that there has been much heat with very little light in this whole area of discussion, I think it's important to acknowledge when someone actually does come up with something refreshingly new and potentially important. Like any new rule, a substantial amount of play testing would be required to further refine the idea. For example:

• To eliminate grandmaster draws of the 12 moves in an uneventful opening variation (where three-fold repetition isn't an issue) followed by a handshake, the Sophia rule would probably need to be included so the game gets to the point where it is clear what the final result will be. Just because something is a draw with perfect play, doesn't mean both players will play perfectly all the way to the end even at the super-GM level.

• Splitting the point 2/3 to 1/3 may not be the right number. For everyone's sanity, I would suggest sticking to decimal values between 0.00 and 1.00 with a resolution of either 0.01 or 0.05 or 0.10 so a full win is still worth 1.00 and a full loss 0.00 just like always.

• I could see a 0.51/0.49 split working just fine (since those extra 0.01's may be sufficient working merely as tie breakers), or maybe something larger like 0.60/0.40 or 0.65/0.35 or 0.70/0.30 so that a number of slight wins can add up significantly in a long event. My gut feeling is that 0.75/0.25 is probably too big; a "slight win" isn't really all that "slight" at that point (since at 0.75/0.25 two slight wins equal a win and a draw -- and making two classical draws the equivalent of a win and a draw isn't good if you want to solve the draw problem). My intuition says that most likely something in the 0.51/0.49 to 0.60/0.40 range is probably right, but the best thing is to empirically determine a value with play testing.

• An interesting option might be to let the players agree on how "slight" the win is, with a fixed value for things like three-fold repetition or stalemate. (Maybe even a different value for different classical draws would be reasonable – but that's probably getting too complicated without adding any value.) Wouldn't it be fun listening to two top GMs bargaining over how to split the point? ("I'll concede a slight win at 0.60." "No, I want 0.70." "How about 0.65?" "Done!")

Albert Frank, Brussels, Belgium
This is another completely absurd idea. For instance, stalemate is an important part of chess, and it ends the game as a draw (nobody should receive "more points" than the opponent).

John Miller, Springfield Township
Just leave everything as it should be, except count wins as two points. When professional and amateur chess players cannot draw their way into a respectable result, norm, or place for the money, there will be more exciting results.

Ricardo Rodulfo, Puerto Ordaz, Venezuela
In football years ago when the game was a tie they gave one point to each team and two point for a win. Then games between strong teams became boring. FIFA then gave three points for a win and only one for a tie. Why FIDE does not do something similar?

Tom Braunlich,Tulsa, OK, USA
I think it's an interesting idea and should be tried with some experimentation. I like that it is a subtle change that doesn't change the "flavor" of chess much, but yet should have significant effects. The terminology needs to be improved. "Slight Win" sounds funny, at least in English. Perhaps the five results could be called: Victory, Win, Draw, Loss, Defeat.

Kung-Ming Tiong, Semenyih, Malaysia
This latest proposal on overcoming some of the weaknesses of chess is just "not so refined". The terminology of slightly win/lose and the conditions attached to it are just mere paraphrasing of previous proposals on counting stalemate and three-fold repetition as "not a win". The only difference now is instead of giving a half point each, now the stalemating player or the player to "force" a three-fold repetition now achieves a slight win (with a larger slice of the points).

As such, the proposal do not solve much, as most unexciting or uneventful draws are not obtained through stalemating or three-fold repetition. Also, there is this peculiar proposal on an opponent conceding a slight win – well, who would want to concede a slight win and end up a fraction of a point poorer?

Again, it must be emphasized that it is a fact acknowledged by strong/elite players (and many players not in those categories) that White enjoys a slight advantage, and those who play white are motivated to convert that advantage in either match-play or tournament play. Failure to capitalize makes it perfectly logical and sensible to penalize White and award Black with a "stronger" draw / win (thus the White 0.4 and Black 0.6 for draw, and White 1.0 Black 0 for a White win and White -0.1 and Black 1.1 for a Black win proposal). Added with the Sofia rule, much of the concern for short uneventful draws could be diminished.

Louis Morin, Montreal, Canada
It is not that clear that such a rule would diminish the number of short uneventful games. For example, with the normal rules, if Black is much stronger than White he will never allow White to force a three-fold repetition in the Zaitsev variation after 10.d4 Re8 11.Ng5 Rf8 12.Nf3 Re8 13.Ng5 Rf8 14.Nf3. However, if Black wins slighly and White loses slightly after these moves, both White and Black might be tempted to play them.

Chris Burns, Wanganui
Finally a sensible approach to the stalemate rule. I have always disliked the stalemate rule when one player enjoys a nice advantage due to playing well in the early stages, but can't convert because the blind bishop or some other silly reason. This change alone will see more games played to the end or players having to risk more if they want more than a half point. Good for tournaments and also sudden death games. Should create new life to certain openings and to chess engine programming.

Sundar Iyer, Palo Alto, CA
I haven't thought through all the ramifications in detail, but my first thought is (pardon the borrowed phrase) –bloody brilliant! It's a really fresh and unique way to look at the problem, very different from many of the solutions presented in the past. One thing that may get refined is that there is a difference between a player forcing an opponent to repeat moves (e.g. say the endgame KNN vs K), and another where the repeating of moves is really because neither player has a better move. In the former case it should be a "slight win", while the latter should really be adjudicated a draw irrespective of who first made the third repetition. I wonder if there is any easy way to distinguish the two? But, Ali's idea is definitely in my opinion a beautiful suggestion.

Valentin Paunescu, Limassol, Cyprus
This is the only interesting idea to combat draws (apart of Fischer Random chess and Sofia rule) in my opinion. Maybe you should give it a try and ask the super-GMs about their opinions on this solution.

Sinan Sarac, Dresden, Germay
I think this is a great idea. Many of the theoretical draws will vanish and players will try to squeeze out tiny advantages instead of agreeing to a draw.

Ozan Ayhan, Istanbul, Turkey
I think the drawback of the proposed solution is that it gives the players more chances of an agreement. What should be aimed to avoid is these agreements between players; the method does not inhibit this, in fact it gives another means of an agreement, agreement to a slight win. The players will have many chances to an agreement at every stage of the game, and this could make them accountants, not chess players. I mean, the game would then revolve around the scores, not the game itself.

Geoff Chandler, Edinburgh, Scotland
Excellent idea. I'm fed up showing possible and plausible finishes in the games I annotate. The fans want to see blood. If a player wants to resign to end their misery then let them construct a sui-mate. This adds something new to the game: imaginative resignations! Yes, good idea, the game ends with checkmate and not when some player cannot see any counter-play or does not have the stomach for the fight.

Michael Kimsns, Georgia, USA
Chess will never attract the level of sponsorship that other sports do simply because chess will never be as popular as other sports are. Face it, chess is a tough, intellectual game; something the masses will never be attracted to. Chess players represent a small fraction of the worldwide population.

The problems spoke of in the article, "Refined Chess – a new proposal to combat draws", if indeed they really are problems, are the problems of the chess elite. Problems that affect only a small percentage of the chess loving population. There are millions of us lower rated chess lovers (say those who play below the 2200 level) who are happily playing in our tournaments, reading our chess books, and gradually unravelling the mysteries of chess; not some dreadful variant cooked up because the chess elite can't attract enough money to the sport.

A radical change to the rules of chess destroys centuries of discovery and accumulated knowledge. It would destroy the very heritage of chess. Chess is chess. Anything else is a different game or at the most a variant. The game is tough enough without the ground moving beneath us. There is a beautiful and I believe a delicate balance of positional and tactical elements that exists on the board as well as a justice. There can be no improvement on this game.

I suspect this talk of changing the game has more to do with saving/creating the jobs, positions, and/or cash prizes of those involved - players, organizers, promoters, authors, and no doubt a host of others – than it does with the noble intention of saving the game. If all of these kinds of people were to walk away from chess and the various associations ceased to exist (yes, even FIDE) chess would still be there because people who truly love chess and its heritage would continue to study and play. And a new chess elite would rise that would actually love the game more than its own comfort and ego.

The game is fine. It has survived for centuries without your help. This is about the pursuit of money and/or the hatred of Real Chess.

Benjamin Austin, Jukjeon, Korea
I enjoy your coverage of new and imaginary chess variants, but I was put off by Refined Chess, and I was wondering if anyone has simply let pawns move one square sideways and backwards?? If this has never been done could it?? I played a few games like this and enjoyed it. I hope your resources can find out.

Shaun Graham-Bowcaster
The new variant is "a single backward pawn move or backward pawn capture". This simple move does not require new learning, new pieces, or a new board. However it introduces new opening ideas and new tactics and in many cases increases defensive possibilities. This move is called the Bowcaster – when you execute the backward pawn move or backward pawn capture, you say “Bowcaster”.

Previous discussion on the subject

ChessBase articles on unfought draws