Your network blocks the Lichess assets!

lichess.org
Donate

World Champion Tiebreaks: A Counter-Intuitive Proposal.

There's a problem with the engine evaluation. New versions of the engine could change the evaluation of certain positions and therefore, for this purpose, engine evaluations are unreliable. Using the best available engine with the best setup and using previous versions of the engine or different engines to consider margin of error is simply not enough. Remember Alpha Zero, it was way stronger than Stockfish (that version of Stockfish), it would be impossible to calculate a reliable margin of error using previous versions of stockfish or different engines at that time.

Another problem is that the method punishes big blunders again. For example, if one of the players blunders a piece, his percentage is alrealy killing him, even if he wins another game. If player A blunders a piece, but the game after that he plays a positional chess masterpiece, official score is 1-1, but in reality he would be still losing because of his first game. This could make the player with the better "tiebreak percentage" to play very solid, drawing chess, to keep the match even, knowing that if they draw the match, he wins the championship. It's like Carlsen-Caruana situation, but worse, because the player trying to draw the match wouldn't have to win the tiebreaks, he would be already champion if he draws.

This tiebreak method also punishes sub-optimal openings. It is said that choosing something like the Pirc of King's Indian doesn't hurt the player because his opponent is playing the same position and it's complicated for both. Consider this situation: Both players play 1.e4 as white in all the games. Player A, plays Caro-Kann defense as black, and Player B, the Pirc. All of the games are won by white.
When player B is white, his opponent plays the Caro-Kann, the game is hard and player B has to play for positional advantages, eventually winning his games because of slight positional mistakes by his opponent. Player A lost, but playing close to the engine moves, because his defense was ideal for that.
When player A is white, his opponent plays the Pirc, the game is very sharp and player A goes for a direct attack against the king, eventually winning his games because of a tactical combination. Player B lost, but since the position was very sharp and his mistakes were tactical, not merely strateginal as in the other case, his positions were very different than those of the engine.
What's the lesson here, winning by tactics is more valuable than winning exploiting positional advantages. Possibly, players who are very good calculators will be in a better spot than positional masters.

There's a problem with the engine evaluation. New versions of the engine could change the evaluation of certain positions and therefore, for this purpose, engine evaluations are unreliable. Using the best available engine with the best setup and using previous versions of the engine or different engines to consider margin of error is simply not enough. Remember Alpha Zero, it was way stronger than Stockfish (that version of Stockfish), it would be impossible to calculate a reliable margin of error using previous versions of stockfish or different engines at that time. Another problem is that the method punishes big blunders again. For example, if one of the players blunders a piece, his percentage is alrealy killing him, even if he wins another game. If player A blunders a piece, but the game after that he plays a positional chess masterpiece, official score is 1-1, but in reality he would be still losing because of his first game. This could make the player with the better "tiebreak percentage" to play very solid, drawing chess, to keep the match even, knowing that if they draw the match, he wins the championship. It's like Carlsen-Caruana situation, but worse, because the player trying to draw the match wouldn't have to win the tiebreaks, he would be already champion if he draws. This tiebreak method also punishes sub-optimal openings. It is said that choosing something like the Pirc of King's Indian doesn't hurt the player because his opponent is playing the same position and it's complicated for both. Consider this situation: Both players play 1.e4 as white in all the games. Player A, plays Caro-Kann defense as black, and Player B, the Pirc. All of the games are won by white. When player B is white, his opponent plays the Caro-Kann, the game is hard and player B has to play for positional advantages, eventually winning his games because of slight positional mistakes by his opponent. Player A lost, but playing close to the engine moves, because his defense was ideal for that. When player A is white, his opponent plays the Pirc, the game is very sharp and player A goes for a direct attack against the king, eventually winning his games because of a tactical combination. Player B lost, but since the position was very sharp and his mistakes were tactical, not merely strateginal as in the other case, his positions were very different than those of the engine. What's the lesson here, winning by tactics is more valuable than winning exploiting positional advantages. Possibly, players who are very good calculators will be in a better spot than positional masters.

With all due respect, this is as TERRIBLE as an idea can be.

Engine evals absolutely do not always translate to human play. There are human sacrifices that engines won't agree with but are completely fine in practice.

What happens when an engine improvement/depth changes the eval?

Honestly, I don't think you could come up with a worse way if you tried.

With all due respect, this is as TERRIBLE as an idea can be. Engine evals absolutely do not always translate to human play. There are human sacrifices that engines won't agree with but are completely fine in practice. What happens when an engine improvement/depth changes the eval? Honestly, I don't think you could come up with a worse way if you tried.

@RuyLopez1000 said ^

Evals do correlate with winning chance tho right? ,so it's not arbitrary?

I mean It's not random. People talk about evals to indicate their 'advantage'. People would not talk about Evals if they are arbitrary. Philosophically a position is either a win, draw or loss. But the eval isn't a random number.

I will stand on this position: if we talk about objectivity in chess, our beloved game is determinated and there's no room for probabilities. It doesn't really mater whether you have +2 or +12 advantage if you able to convert your advantage. Both evals on big enough depth will transform into checkmate in X moves eval; if you don't have 'winning advantage', you objectively don't have any advantage, at best your opponent may be forced to make 'only moves' in perfect (sufficient to maintain a draw) play.

@RuyLopez1000 said [^](/forum/redirect/post/wLcTtmKK) > Evals do correlate with winning chance tho right? ,so it's not arbitrary? > > I mean It's not random. People talk about evals to indicate their 'advantage'. People would not talk about Evals if they are arbitrary. Philosophically a position is either a win, draw or loss. But the eval isn't a random number. I will stand on this position: if we talk about objectivity in chess, our beloved game is determinated and there's no room for probabilities. It doesn't really mater whether you have +2 or +12 advantage if you able to convert your advantage. Both evals on big enough depth will transform into checkmate in X moves eval; if you don't have 'winning advantage', you objectively don't have any advantage, at best your opponent may be forced to make 'only moves' in perfect (sufficient to maintain a draw) play.

@Tactical-Attack said ^

But even in a classical time control game, you can't give more credit to a solid game by Anish Giri, who doesn't take risks and waits for his opponent to kill himself, than to a game by Alireza Firoujza or Richard Rapport, who complicate every game in pursuit of a full point from the start. Even in a classical time control game, mistakes will happen, and that's part of chess; it doesn't diminish the merit of the fight. Furthermore, many moves that a human wouldn't consider mistakes, even if the computer says they are, since we're playing against real people, not stockfish. Besides, you prevent them from making "human" poisonous preparations.

I agree as in chess the main motto is to never give up . Anyways we all are humans not STOCKFISH !

@Tactical-Attack said [^](/forum/redirect/post/wAUZOMeg) > But even in a classical time control game, you can't give more credit to a solid game by Anish Giri, who doesn't take risks and waits for his opponent to kill himself, than to a game by Alireza Firoujza or Richard Rapport, who complicate every game in pursuit of a full point from the start. Even in a classical time control game, mistakes will happen, and that's part of chess; it doesn't diminish the merit of the fight. Furthermore, many moves that a human wouldn't consider mistakes, even if the computer says they are, since we're playing against real people, not stockfish. Besides, you prevent them from making "human" poisonous preparations. I agree as in chess the main motto is to never give up . Anyways we all are humans not STOCKFISH !

@DaBassie

Thought above what you said. You are indeed right.

A draw should have the same TWPL as the win percentage loss should be 0%. A draw should not have different TWPL's. This is only a concern if players draw all the games. Otherwise won games can be used to differentiate the more clinical player. Players drawing all the games is rare (only Carlsen-Caruana).

If players draw all their games then the Classical World Champion will be decided by a coin flip. This is because if two players play equally well then a 50% chance is more fair then Rapid and Blitz where the better player in that format has an advantage.

The above scenario is unlikely both statistically and also because the contenders want to win the championship outright. Having a lead in the points is a greater advantage than a coin toss. The Classical World Championship should be based on Classic chess. Rapid and Blitz have their own World Championships.

Ultimately, this tiebreak mechanism is a means to an end. It does not change the fact that players will want to win by points, in contrast to situations like the 2018 World Championship. The Championship will ideally be decided by a player winning more games. In the event of a tie the Winner is based on Classical Chess games, as opposed to Rapid and Blitz games. But we want to see a win by most Classical wins as opposed to a tiebreak win. This tiebreak system incentivizes wins in Classical Chess compared to the current Rapid and Blitz playoff.

@DaBassie Thought above what you said. You are indeed right. A draw should have the same TWPL as the win percentage loss should be 0%. A draw should not have different TWPL's. This is only a concern if players draw all the games. Otherwise won games can be used to differentiate the more clinical player. Players drawing all the games is rare (only Carlsen-Caruana). If players draw all their games then the Classical World Champion will be decided by a **coin flip**. This is because if two players play equally well then a 50% chance is more fair then Rapid and Blitz where the better player in that format has an advantage. The above scenario is unlikely both statistically and also because the contenders want to win the championship outright. Having a lead in the points is a greater advantage than a coin toss. The Classical World Championship should be based on Classic chess. Rapid and Blitz have their own World Championships. Ultimately, this tiebreak mechanism is a means to an end. It does not change the fact that players will want to win by points, in contrast to situations like the 2018 World Championship. The Championship will ideally be decided by a player winning more games. In the event of a tie the Winner is based on Classical Chess games, as opposed to Rapid and Blitz games. But we want to see a win by most Classical wins as opposed to a tiebreak win. This tiebreak system incentivizes wins in Classical Chess compared to the current Rapid and Blitz playoff.

@ChesslyChessdriller said ^

When player B is white, his opponent plays the Caro-Kann, the game is hard and player B has to play for positional advantages, eventually winning his games because of slight positional mistakes by his opponent. Player A lost, but playing close to the engine moves, because his defense was ideal for that. When player A is white, his opponent plays the Pirc, the game is very sharp and player A goes for a direct attack against the king, eventually winning his games because of a tactical combination. Player B lost, but since the position was very sharp and his mistakes were tactical, not merely strateginal as in the other case, his positions were very different than those of the engine. What's the lesson here, winning by tactics is more valuable than winning exploiting positional advantages. Possibly, players who are very good calculators will be in a better spot than positional masters.

A very good point. Thank you.

This TWPL cannot work. It won't work.

@ChesslyChessdriller said [^](/forum/redirect/post/VqPkEv5R) >When player B is white, his opponent plays the Caro-Kann, the game is hard and player B has to play for positional advantages, eventually winning his games because of slight positional mistakes by his opponent. Player A lost, but playing close to the engine moves, because his defense was ideal for that. When player A is white, his opponent plays the Pirc, the game is very sharp and player A goes for a direct attack against the king, eventually winning his games because of a tactical combination. Player B lost, but since the position was very sharp and his mistakes were tactical, not merely strateginal as in the other case, his positions were very different than those of the engine. What's the lesson here, winning by tactics is more valuable than winning exploiting positional advantages. Possibly, players who are very good calculators will be in a better spot than positional masters. A very good point. Thank you. This TWPL cannot work. It won't work.

@tpr @PRIYAMVAD @Akavall @Tactical-Attack @elkress @shikharnanavati @farista1 @solarglitch @DaBassie @zqx2009 @GennadyBukin @MillenniumBug @mrgwbland @alijeba @aanellien @SimplyBubbles014 @ChesslyChessdriller @maroon_marine @PRESHAKS_2017

TWPL as suggested by me cannot work.

Because logically the TWPL should be the same for draws. And because when considering won games, tactical players may have an advantage as the relative TWPL difference would be greater for tactical players as blunders weigh heavily, compared to positional players where the opponent may not make a blunder, but many inaccuracies. This pointed out by @DaBassie and @ChesslyChessdriller.

The new counter-intuitive proposal is that the tiebreak shall be a coin toss.

  1. This is because if two players play equally well in Classical then a 50% chance is more fair then Rapid and Blitz. The better player in those formats has an advantage which is unfair as this is the Classical World Championship. Rapid and Blitz have their own Championships.

  2. This tiebreak incentivizes fighting chess because the contenders want to win the championship outright. Having a lead in the points is a greater advantage than a coin toss.

@tpr @PRIYAMVAD @Akavall @Tactical-Attack @elkress @shikharnanavati @farista1 @solarglitch @DaBassie @zqx2009 @GennadyBukin @MillenniumBug @mrgwbland @alijeba @aanellien @SimplyBubbles014 @ChesslyChessdriller @maroon_marine @PRESHAKS_2017 **TWPL as suggested by me cannot work.** Because logically the TWPL should be the same for draws. And because when considering won games, tactical players may have an advantage as the relative TWPL difference would be greater for tactical players as blunders weigh heavily, compared to positional players where the opponent may not make a blunder, but many inaccuracies. This pointed out by @DaBassie and @ChesslyChessdriller. ---- The **new** counter-intuitive proposal is that the tiebreak shall be a **coin toss.** 1. This is because if two players play equally well in Classical then a 50% chance is more fair then Rapid and Blitz. The better player in those formats has an advantage which is unfair as this is the Classical World Championship. Rapid and Blitz have their own Championships. 2. This tiebreak incentivizes fighting chess because the contenders want to win the championship outright. Having a lead in the points is a greater advantage than a coin toss.

I solved this problem 6-7 years ago

Play the match in reverse order - Armageddon, blitz, rapid then classical

Random reddit app, please remind me in 10 years about this post

I solved this problem 6-7 years ago Play the match in reverse order - Armageddon, blitz, rapid then classical Random reddit app, please remind me in 10 years about this post

Huh?? Your argument defeats the point of chess. Chess is a strategy game with no luck involved.A coin toss is not a measure of skill, and then the players may both feel that they are not good enough and draw everything just for the coin toss.I have. a counter proposal.Make them play an odd number of armageddon classical ,with slightly less time.3 games would be the best number.

Huh?? Your argument defeats the point of chess. Chess is a strategy game with no luck involved.A coin toss is not a measure of skill, and then the players may both feel that they are not good enough and draw everything just for the coin toss.I have. a counter proposal.Make them play an odd number of armageddon classical ,with slightly less time.3 games would be the best number.

@SSB2013GEM said ^

Huh?? Your argument defeats the point of chess. Chess is a strategy game with no luck involved.

The coin toss only applies if its drawn. The players will still try to win by getting the most wins.

Don't forget that the coin toss makes players more motivated to win Classical games because the contenders want to win the championship outright. Having a lead in the points is a greater advantage than a coin toss.

A coin toss is not a measure of skill

Correct. But Rapid and Blitz is also not a measure of Classical skill.

What do you think about this line of reasoning?:

If two players play equally well in Classical then a 50% chance is more fair then Rapid and Blitz. The better player in those formats has an advantage which is unfair as this is the Classical World Championship. Rapid and Blitz have their own Championships.

, and then the players may both feel that they are not good enough and draw everything just for the coin toss.

How could the players feel they are 'not good enough'? These are Super GMs. They are serious people who have worked all their life to try to become champion. Both contenders are not going to just want their chance at become champion be decided by a coin toss.

I have. a counter proposal. Make them play an odd number of armageddon classical ,with slightly less time.3 games would be the best number.

It's a good idea.

@SSB2013GEM said [^](/forum/redirect/post/qUwGnGZl) > Huh?? Your argument defeats the point of chess. Chess is a strategy game with no luck involved. The coin toss only applies if its drawn. The players will still try to win by getting the most wins. Don't forget that the coin toss makes players more motivated to win Classical games because the contenders want to win the championship outright. Having a lead in the points is a greater advantage than a coin toss. >A coin toss is not a measure of skill Correct. But Rapid and Blitz is also not a measure of Classical skill. What do you think about this line of reasoning?: *If two players play equally well in Classical then a 50% chance is more fair then Rapid and Blitz. The better player in those formats has an advantage which is unfair as this is the Classical World Championship. Rapid and Blitz have their own Championships.* >, and then the players may both feel that they are not good enough and draw everything just for the coin toss. How could the players feel they are 'not good enough'? These are Super GMs. They are serious people who have worked all their life to try to become champion. Both contenders are not going to just want their chance at become champion be decided by a coin toss. >I have. a counter proposal. Make them play an odd number of armageddon classical ,with slightly less time.3 games would be the best number. It's a good idea.