Want to take part in these discussions? Sign in if you have an account, or apply for one below
"All that matters on the Chessboard is good moves" (Bobby Fischer)
Vanilla 1.1.3 is a product of Lussumo. More Information: Documentation, Community Support.
This was brought up in a conversation with someone about draws in chess. They mentioned that perhaps a draw should cost each player half a point. This is a bit different than draws being worth less, but is the same idea. If whomever I mentioned this reads it, they can feel free to chime in.
Anyhow, what do people think here? I believe the idea has merit. Maybe not lose points, but perhaps for chess, have a win 1 point, a loss 1/2 point, and a draw zero points.
I was thinking more along the lines of win = 1, loss = 0, draw = - 0.1, to encourage people to play hard for the win. I'm sure this will outrage some scoring purists, and people will complain that you can come out ahead by playing to lose, but look at the current FIDE scoring system, and the number of times high-level players collude/have colluded [the Russians had a rep for that in the days of the USSR] to game that system. Yes, this is my lame idea, but at least it would/should encourage playing to a conclusion, even if this, too, can be gamed.
Any thing with a scoring system, and winners and losers will be gamed, because it is a game in and of itself. As for what you wrote, I believe having a tie worth less than a loss does encourage people to play for the win. You also end up having to make a tough decision whether or not to ask for a loss instead of playing for a draw.
To me: win = 1, draw = 0.5, loss = 0
A lot of hard work can go into trying to recover from a lost game and earn a draw. To then be reworded with a "0" when a simple resign would get "0.5" is something I cannot seriously consider. Having a loss and draw both at 0 is also bad. For example, two 2600 rated players in an event out rank their next closest competition by 600 points. They have each defeated all others. Now they get a hard fought draw. With 1/2 point each they tie for first (with each other). But, instead with 0 points they could now be even with players who's only loss was to them. This makes no sense to me.
One way to eliminate draws in chess is to convert 1 of those player's draws to a win (1 point) and the other to a loss (0 point). How would this be done? By a random element (coin toss... dice throw). This would discourage draws. It would mean that after all your hard work on the board you pretty much have to win... otherwise, getting a chess draw gives you a 50% chance of getting a zero. This would discourage draws yet keep the 1 - 0 point system balanced. The masters "gentlemens' draw" would disappear. I would hate taking a draw knowing that it entitled me to a dice roll or a coin toss for a 50% chance of getting a zero.
There's another way to eliminate draws in Chess, but purists would be outraged. In many international variants, draws are impossible. The "stalemate" is pretty much a European variant. In Chinese Chess, if a player has no legal way to move without moving into Check, he loses. If a player cannot move without moving into a repetious pattern, he loses.
However, given that we aren't going to change the rules of Chess, I think it would be a very bad rule. Sure, it would discourage people playing for a draw, but what would it encourage? Playing for a loss? I don't see how that's better.
And as for collusion, it increases the possible issues. I'm about to checkmate an opponent, but I want my buddy to win the tournament. My current opponent will be his competition. So, instead of checkmating him, I do more damage by deliberately moving to a stalemate. I don't see how that has made anything better.
The issues with draws in chess currently is that the chess champion ends up playing defensive and wins. A draw is a draw when the scoring is split evenly. If a draw advanced one side or another to win a match, then it can't be considered a draw. At that point it is a non-checkmate end position, that scored more points for one side or the other.
I personally don't like the idea of a coin flip to decide which side gets the draw advantage. What had possibly been suggested is use of a doubling cube, where one side gets draw advantage. There is also the BAP (I think I recall the name correctly) that breaks the pure draw. The appeal of the doubling cube is that it adds another meta-element that hooks people who don't play, because it gives people something they can follow that is unexpected, but player controlled. It isn't for everyone, however.
As far as not appealing to purists, I don't see IAGO's core appealing to purists. IAGO will partner with them, and can use them to help be contestants in tournaments, but IAGO is all about mixing things up, experimenting, and adding new and novel to the world of abstracts. This means even mucking around with the scoring for a chess tournament. IAGO provides an out for FIDE to experiment, for example. I personally believe enough people are willing to do tweaks and see, that IAGO does serve a useful purpose. Some people are up to major revisions and completely new. Others want a minor tweak. In all this is the ability to find stuff that works better. As of now it is tough for FIDE and others to discover this. IAGO allows them to do it.
Meadmaker, Gary is right - there are "gentlemen's draws" now [thanks for the term, Gary] and people can and do collude now in exactly the way you suggest for the negative draw. I agree that a negative draw system can be gamed, but I think it might be a little more obvious that the game is being "gamed" in this case, though that's only a guess. Gary, your suggestion is a very interesting possibility. Statistically, it averages out, but is does prevent those nasty draws.
Or you might have a 5 minute blitz game of Los Alamos Chess, or any other 6x6 variant - if that ends in a draw, run another game at 4 minutes, then 3, 2 and 1. [Okay, you could do a blitz FIDE game, or an even larger one but this wouldn't work on tv so well.]
joejoyce, I believe the Chess960 championship used a blitz match to resolve and end-game tie in scoring. As far as I am concerned here, I feel an overwhelming amount of draws, with balanced scoring, doesn't make for an event that draws new people in. This is why I would like to see alternatives produced here. The idea of the draw being worth less than a win is a start. I believe also a doubling cube can help. So long as there is a strategic/tactical way for people to change what a match is worth, it lends to interesting metaplay, which gets people who are new interested.
Ok, maybe I am a bit biased with an American perspective, but I believe my concerns are valid.
Meadmaker who told you about Chinese Chess?
http://www.clubxiangqi.com/rules/asiarule.htm#c4 seems to be quite different.
When you give draws lower score than losses, what you really do is change the losing conditions. It will lead to situations where a player can make a single losing move, or a single, drawing move, and it's rational of him to choose the "loss"!
No, if you want to discourage draws, look at how football (european) does it: Three points for a win, one point for a tie, no points for a loss. With that system, it's never in the interest of a team to do something they know for certain will lose, but it becomes more profitable to take chances to turn a tie into a win.
If we used the negative points for draws principle suggested here, there would be a lot more last minute own goals in football!
Perhaps IAGO can run some game theory classes for the benefit of its members? ;-)
omweso,
I was oversimplifying to the point of error. Yes, draws are possible in Xiangqi. What I meant was that stalemates were impossible. Also, in Chess, a threefold repetition can happen, and a draw is automatic if it does. In Xiangqi, one of the players is obligated to break the repetitious pattern if possible to do so. If, on the other hand, there's no way to break the repetitious pattern in a reasonable way, (see rules for definitions) it's a draw.
As a consequences of these rules, draws are less common in Xiangqi than in Chess, but they do still happen.
vinterman, what are your thoughts in regards to the BAP system?
Hey, Vinterman, I'm a retired postal employee; I just design 'em, I don't understand 'em! :-)
The point about the negative score for a draw is exactly what you said - it pays more to lose than draw, thus giving a winner and a loser. I would think the 3-1-0 system encourages people to draw rather than go all out for a win, because if there's a "whoops", then they lose, the worst outcome.
My logic [such as it is] in proposing a loss of 1/10 point for a draw is that draws become the worst outcome, the one to be most avoided. Any system that rewards a draw more than a loss will encourage draws over losses. The worst outcome is the one to be most avoided, no? Better outcomes are more to be sought after. That's my rationale. IAGO wants games to have winners and losers, not an endless series of draws. Punish the draw. Game theory only works when all players are rational.
Well, a draw is a draw if scoring is balanced for the draw. This means both players are affected equally. This is particularly true in the case of where the tournament just has two players in it. In the case where scoring is not balanced, and one side has an advantage for drawing, then a draw position ends up being a win for someone. The game in this case, takes on a unisided escape type mode, with one side as an attacker and one side as a defender. IAGO should have games like this in it, like games in the Tablut family. When scoring goes into this mode, someone is then playing the defender and trying to cause their opponent to run out of gas.
However, I do believe there is merit to considering having scoring less for a draw than a loss. Players still want to go for a win, but then have to decide when to call it quits. If you are going to have it where it is like -.1 for a draw, then I seriously suggest scoring for a match be multiplied by 10, so you don't deal with decimals or fractions. Make a win 10 points, a loss 0 points and a draw -1 point, if you will do this. This way, no fraction. Of course, this scoring only works in a Swiss match. It is zero sum in 2 player only. Now if say you give -1 to only one side, and not both, that is a different thing.
Having draws worth less than a loss, encourages players to do riskier lines, which I consider a good thing. Riskier lines make for more dramatic games, which is good. Same would go with the use of a doubling cube.
You're opening up a Pandora's box of weird strategies with this nonsense. Yes, it is nonsense!
Consider that a player who is in a bad position - or, what would be considered a bad position in regular chess. He can draw, but he can't win. Now he can threaten the winning player. "Hey! I'm going to play for a draw unless you let me win!" Let's see at the "winning" player's option then:
1. He refuses. His opponent ties the game, leaving him with -1 point.
2. He caves in. Now he at least doesn't have negative points.
What you have here is a classic game of chicken. The player who can force a draw hurts himself if he does it, but if he can convince his opponent that he's irrational enough to do it (and it might even occur situations where it's rational! In these cases, the "winner" is in even worse trouble.), he can walk away with 10 points.
Just look up the game of chicken in wikipedia, OK? If you're going to mess around with alternate payoff systems, read up on some basic game theory!!!
I would think the 3-1-0 system encourages people to draw rather than go all out for a win, because if there's a "whoops", then they lose, the worst outcome.
You may think so, but you're wrong! Since a win is worth three times more than a tie, you can justify taking pretty big chances. Imagine that they are tied, and one team estimates that agressive play will make them lose slightly less than 2/3 of the time, and win slightly more than 1/3 of the time. With the system used in chess today, they would be fools to play offensively. With the system actually used in football today, they should play offensively. Football organisations (unlike football fans *cough*) aren't stupid. They know agressive play and fewer ties means more money, and they have found a solution that works.
Don't think it's enough incentive? Raise the win value to 4, or 5. But don't make ties worth less than losses! You should see that it fundamentally changes the game when the terms "tie" and "loss" aren't even appropriate any longer.
Actually, adding a game of chicken to the metagame might be an improvement to make an abstract strategy tournament more watchable by spectators who don't know the game. Throw in a doubling cube also, and you could be on to something. Vintermann, do you believe that draws should score the same for both sides? I personally think this makes both sides also play defensively.
Man, I hit the wrong key and just lost a long comment... curses!
Vinterman, I'm willing to be convinced, but I haven't been yet. After reading "chicken" 1.5 times, and following some of the references, I see that my proposal changes a 2-player zero-sum game into a 2-player non-zero-sum game. Apparently, I have to learn game theory to follow your argument - been almost 40 years since my last math course, linear diff eq, and matrices and I never saw eye-to-eye, so help me out here. I don't see the structure of the game changing at all; the rules, win, loss, and draw conditions are all still the same, only the payoff is different for losses and draws. Yes, one's strategy will be different, but the means to execute that strategy is the same: the same sequence of moves will lead to the same outcome regardless of payoff. Your proposed payoff numbers are 3-1-0, whereas the [normalized] FIDE payoff is 2-1-0. As I see it, your argument is that you only have to win just over 1/3 of the time with your system to make aggressive play worthwhile, whereas FIDE requires you to win just over 1/2 the time, a more difficult task. This will then reduce the number of draws. But by how much? If a player believes [s]he is losing by going all out, then it becomes worthwhile to go for the easier goal of a draw, gaining for oneself and costing the opponent. If your goal is eliminating draws, how far toward that goal does the football scoring system take you?
I do see part of your argument, but before I agree, show me that the incidence of draws will really drop precipitously, which is IAGO's goal, and that penalizing players for draws won't effectively reduce draws.
I'm not looking to rile people up here, just looking for a decent understanding of why I'm wrong. Or get everyone to say I'm wrong, and I'll go along without the understanding it part; I'm easy.
Joe
Richarhutnik, adding "chicken" to the metagame is roughly equivalent to adding "rock-papers-scissors" to the metagame. The only strategy viable in a tournament contest isn't stable - you have to randomly decide whether to yield or push, biased according to the payoffs.
Joejoyce, the football scoring system does not eliminate draws, but it probably eliminates deliberate draws. But remember, this is also a change to the game. There's probably a reason Chess federations don't do it that way - you're encouraging the players to increase the variation in the outcomes, which makes it more entertaining, but less an accurate measure of skill (for a single game). The football method also changes it into a non-zero sum game, but one with a reasonable strategy: Go for the win!
If players think they will lose outright more than 2/3 of the time, it may become profitable to go for a draw. But in such a situation, they should rather be given a more equally matched opponent (or a handicap, for those games that support it).
Also note that since it is non-zero sum, equally matched players may cooperate by arranging to win one game and lose the other. That gives more points than a tie in both... but this presumes that they can make binding commitments (who says the player they let win first will follow through and lose the second?) and it requires a lack of ambition uncharacteristic of competitive players. Surely they have a chance of winning at least one of the matches, gettin 4 points instead of 3?
vintermann, as I see it, the "chicken" element ends up adding a bit of "reading other players" to a game, to see if they would fold or not. It isn't random, but based upon knowing whether or not your opponent fold. It also requires you know your game, and the state, and work to avoid a draw position. If you feel you can fight for a win, you do it, or else your resign. Throw in a doubling cube on top of that, and you provide a way to force people to make decisions. It makes it something people who don't follow a game can follow.
Also, I am curious here if anyone thinks a draw should score evenly for both sides. I personally feel this causes draws to become stalling points and are counter productive. In other words, they result in things that happen that don't advance the match in any way. You have a problem, when if you review a tournament, out of 16 games played, only 2 or 3 are notable, while the rest are forgettable. That is not a good thing. I remember seeing how the recent chess championship in Mexico had the current world champion (Ananda?), say how it was a great match, and he spoke only of 2 or 3 games being interesting.