10

Chess (often being classified as a sport) stands almost alone among sports in that its results are of a strictly ternary nature that is not quantified any further: The only possible outcomes are a win, a draw or a loss.

While the same is of course true for most team or combat sports, virtually all of them go a lot further by basing the result on some sort of point score, which says a lot more about the match than merely who won: A 6-0 win in association football is not the same as a 2-1 (the first being a "crush" and the second being a rather narrow win), and a 3-3 draw is not the same as a 0-0 one (the first one indicating a wild game, the second a tenacious one). Further in-game metrics such as "ball control" and "shots on goal" are often employed to paint a high-level picture for spectators.

Given that literally hundreds, if not thousands, of variants to chess rules exist, I should be surprised if no one has ever proposed an improvement to how chess is scored, but alas, I have not found any material on that subject. I'm looking for a new angle to see top-level matches from, where, when analyzing games, one often cannot help but feel that some are won by a larger margin than others, and similar insights.

Do any alternatives to ternary (win, draw, loss) chess scoring exist?

6

I think the only alternative would be in match play or comparing a player's score across multiple games against another player or players.

For example, in a 12 game match between two players:

(+3 =8 -1) would be worse than (+5 =6 -1)

even though the first player won both matches.

Scoring by material count or number of moves played would not be beneficial. If the number of moves mattered, with shorter games being considered crushing defeats, players would just refuse to resign and drag the game out. On the other hand, I can lose in two moves without losing any pieces.

The only thing that seems to make sense is to compare score across multiple games against an opponent or field of opponents. You may have a better comparison considering a GM's total score in a tournament with many rounds. In a 9 round tournament, a 9.0 score would be crushing the competition. Winning the tournament with 6.5 would be less of an accomplishment.

| improve this answer | |
4
  1. Sometimes the media would report the number of moves the game lasted as some kind of indication of how badly an opponent was "crushed". If Magnus was to lose in less than 20 moves, that could be considered "badly beaten" by whomever he was playing, and I'm sure anybody reporting on it would mention that fact. To make it more sophisticated, you could count the number of moves until one side had a substantial advantage (i.e. Nakamura was losing by move 17, even if he managed to draw it out until move 63).

  2. Another method I can think of is to use the computer evaluation during and at the end of the game. For example, recently Nakamura was +10 against Carlsen in a game (which Carlsen subsequently won) - that obviously gives an indication of what type of lead Nakamura threw away. By looking at the computer evaluation over the length of the game, you can get some idea of who was winning by when (although how you would coalesce that into a single number is something the statisticians may want to have a go at).

Personally, I'm perfectly fine with chess having only a simple ternary outcome. I am always skeptical of using alternative scoring methods. Does 4-0 really mean Liverpool crushed Arsenal? What if the score was 1-0 with 5 minutes left, and Liverpool's forward scored a hat trick while Arsenal was pushing aggressively? Maybe Arsenal even missed 5 good shots on goal during that last 5 minutes. Is it fair that Arsenal is now eliminated due to a worse goal difference?

| improve this answer | |
2

People have proposed adding a new fourth result for stalemate, where (for example) the stalemated player would get 1/4 point and the stalemating player would get 3/4 points. Unfortunately I don't have a reference at hand.

| improve this answer | |
  • Such a thing isn't going to happen, but if it were I wonder if there should be a distinction made between situations where there would be a legal (losing) move but for the prohibition against entering check, and those where no pieces would be able to move even if entering check were legal. – supercat Oct 10 '14 at 20:32
  • Yeah, it would make the theory nicer, but in practice the number of stalemate positions where you couldn't move even if check were legal is vanishingly small; you really have to work to construct one. – dfan Oct 11 '14 at 0:17
  • I was just thinking of another idea, in case the existing game gets too over-analyzed and needs to be shaken up: what if the game proceeded until the half-move after a player's king was taken. If by the end of that half-move only one king remains, the survivor gets one point. If neither king remains, both players get 1/2. That would mean that a player who was behind could check the opponent's king with a pinned piece; if the opponent takes his king, both players would get a half-point, but the opponent might opt instead to defend his own king. I wonder how that would affect dynamics? – supercat Oct 11 '14 at 3:51
  • @supercat Actually, that's been precisely the case at various points throughout the game's history -- it probably won't swing away from {1/2,1/2} again any time soon, or perhaps ever (though I doubt this), but it's been this way for only about a century and a half (or two centuries?). – Charles Rockafellor Jun 22 '16 at 17:05
1

I know of two.

  1. The ELO rating. Players are awarded points based upon the result of the game versus the expectation. The rating is used to rank players in terms of skill. Higher rating denote higher skill. Higher rating carry more prestige. A higher rating also gives one an advantage in USCF Swiss-style tournaments and higher rated players are paired against lower rated players, giving the former a better chance at a win/loss result that nets prize money.

  2. In some tournaments, draws are awarded .4 points. This makes two draws markedly less valuable than a win and a loss. This has been done to encourage players to play decisive games.

| improve this answer | |
  • 1
    I think the question is after alternative methods to measure the qualitative aspects of a single chess game (i.e. by how much did player A beat player B?). Both your points, although interesting, relate to a series of games. – firtydank Sep 11 '14 at 12:33
  • @firtydank Perhaps. But I offer a different way forward. – Tony Ennis Sep 11 '14 at 12:46
1

The Bilbao 2008 tournament used the FIFA scoring system:

  • 3 for a win
  • 1 for a draw
  • 0 for a loss

It has been used at other tournaments such as the London Chess Classic.

The aim is to encourage 'fighting chess' and discourage early draws as the tournament result favours those who have won most games.

| improve this answer | |
1

You're confusing score with score.

Except for the game score (notation) and commentary, there's no real to show the action on the board. Even a material balance could be misleading, as we'd need to differentiate between a sacrificial attack and a lucky mate when losing badly. We do have an external indicator, ratings*, which, by the difference, can suggest the difficulty of play.

There have been suggestions to alter the score (game result), but these were to reduce the number of (GM) draws. The only accepted suggestion was to give a win more points for a win that two draws, mainly the 3-1-0 scoring system. A favorite was to split the point 0.4-0.6 as black has a harder time earning that draw. Another, misunderstood, system was to split the point based upon the rating difference. A lower rated player would receive more of the point as his effort in drawing the game could be considered greater. (Or the opposite is the the higher rated player has the "duty" to make a more decisive result.) An infeasible system of having a third party judge the game and decide who deserved more points. This could result in the losing player earning more points.

*Rating effects what we assume. In a professional game, a 0-0 score makes us infer that the goalie was exceptional or the players were good at defense. If the game was played by beginners, we might assume that they bumbled every play.

| improve this answer | |
0

Some have tried it.

But it distorts the game and is illogical for game results although useful to help tournaments create more interest for sponsors.

A win and a loss against a player once with white and once with black has to be equal to two draws against the player again with equal colors for each.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy