First of all, I'm not trying to push anyone's buttons here. I seek opinions on a way of implementing komi which could be regarded as "future proof." Before the first stone is placed, one player announces what the komi will be, and the other player decides which side to play. I call this komi pie because it is like when two people want to share the last of the pie. One person cuts the pie into two slices, and the other decides which slice to eat. One downside of this protocol is, it places a greater burden on the players. They have to play a sort of meta-game before the real game begins. Perhaps worse than that, there is no established tradition of playing this way. But perhaps if there is interest, some servers might implement it as an option. One possible upside is, no governing body or tournament director would need to announce what komi value will be used. Does anyone agree with me that this might be a good thing?

Thanks for your time.

I can't really get excited about any such proposal. Disagreement about komi is too limited for this to be of real interest.

_________________
Occupy Babel!

How do you deal with information problems? I don't think many casual players have a strong idea about whether the first move is worth five points, six points, or some other number.

I don't think this is a problem. Average players don't have a strong idea what to do at ANY pont of the game, so this one addition does not really change things. One player decides what he thinks the komi should be, and the other agrees or not. Live and learn...

_________________
- Bantari
______________________________________________
WARNING: This post might contain Opinions!!

Where I actually DO see a problems with that proposal is here:

1.
Eventually, optimal amount of Komi will be guessed, and people will just bid it by default. I bet right now most players will say 6.5 or so... out of fashion/inertia, if nothing else. In such case, the person who bids the komi is at a disadvantage because the other player can simply chose the color he prefers...

2.
At lower levels, komi really does not matter that much... I am not sure where the line lies, 5k maybe, or 1d... or 5d... and even then the komi is not very exact. So the decision also boils down to picking the color you like, regardless of komi.

Of course, this can be solved by doing nigiri for who bids the komi, for example. But I think this has to be stated clearly in such proposals, I think.

Another consideration:
At high levels, not knowing the komi and/or the color - it makes it so much harder to prepare for games and matches. Does that mean the standard of play will suffer? Or will it be raised?

_________________
- Bantari
______________________________________________
WARNING: This post might contain Opinions!!

It could encourage posturing. For example, if I get to bid, komi is 35.5. So do you want to take a handicap or do you fancy yourself strong enough to give me one? Are you chicken?

Just a hypothetical flaw....

I don't think it makes very much sense. I general, you don't get to modify the rules of a game just because both players agree.

We have a thread about pie rules in the rules sub-forum, if you want to read whatever said the first time.

Yes, the player who decides which side to play arguably has a theoretical and a practical advantage. So who plays which role must be determined in some fair and impartial manner. The tournament rules could stipulate the methodology that would be involved, for all to understand at the outset. It could also be argued that as things stand, with komi determined by some third party, one side or the other might have an advantage, and this advantage might be greater than with komi pie. For example, perhaps it will be determined some day that komi should be higher than it usually is now. That would imply that black has a greater advantage than players generally recognize now.

Yes, a tendency towards inertia might result in players announcing whatever komi is currently in fashion. Truly playing this meta-game involves extra work. For example, if you prefer black and you are accustomed to 6.5 komi, you must judge whether your preference for black is worth a bump to 7.5 or not. Would players be willing to take on this extra work? I would imagine most would not. But maybe things will slowly change over time.

As to whether this rule change is permissible just because players agree to it or not, I suppose similar arguments were made against the adoption of any komi, when the concept was first introduced. If some server were to implement komi pie as an option, the players could decide for themselves whether to use it or not. If it becomes popular, the resulting game will be played, and it will be called Go, whether it deserves the label or not. I'm not trying to sound confrontational.

I suppose players could announce ridiculous komi values if they wish, but they could also make ridiculous opening moves if they wish. If you make such a weak move, and lose the game as a result, you have only yourself to blame. And if you win, that would indicate you were stronger than your opponent anyway.

Thanks for mentioning the other thread. I have difficulty accessing that page. I keep getting a "server reset" error. I have noticed the same problem with other pages on this site. I will keep trying.

I hate to state the obvious, but if your opponent takes the komi, then you will lose (and feel silly). The komi system isn't flawed; the flaw is in your assumption that your overly macho opponent will give you a free handicap.

(Of course, if we were evenly matched and you did this to me in a tournament game, I would make sure we had a quick and boring game, and probably not play you again.)

I don't think there is a problem here. Hypothetically, if the "optimal" amount of komi is guessed, this means that color does not matter - komi accounts for not playing first perfectly. So the person who bids the komi is not at a disadvantage. If there was real concern, nigiri could be used to decide who bids. So in the worst case, the advantage/disadvantage is the same as it is without the bidding system.

I can only think of one disadvantage with the proposal: if players start to think that whole numbers are best for komi, this could lead to ties. It's common to have a fractional komi to avoid ties. That being said, this disadvantage is also an advantage in that the selected komi is fair to both sides, whereas a fractional komi might not be.

All in all, I think the "Komi pie" is a great idea.

_________________
be immersed

There is a slight problem, though, and it takes into account the fact that "optimal" komi varies from one player to another.

The "I cut, you choose" problem basically says that I should attempt to cut a pie into two pieces I would be equally happy with, and let you choose which you prefer, so we both end up happy. But (this sounds like a silly thing to say) we both understand the pie equally well, and I don't think we both understand komi equally well. Let me attempt to do some sleepy maths...

Suppose that you and I are, on average, the same strength. I am about "10 points better" at playing with black than with white. (I'll call this my "relative colour strength", for the sake of a name.) That is, the choice "take black and give 5 points' komi" or "take white and take 5 points' komi" is a completely fair one for me to make against an average opponent - I'll win or lose 50% of my (non-jigo) games by taking or giving 5 komi if I play enough opponents enough times. But no single opponent is average: you are especially strong with black / weak with white compared to me, so your relative colour strength is 20.

If I don't know you at all, and I bid a komi of 5, what happens? Of course, you decline - you would normally give someone like me 10 komi. I think I've got a fair result, but you're taking the colour you prefer and giving 5 points less compensation than you normally do. You're essentially at a 5-point advantage.

If I know your relative colour strength but you don't know mine, I should bid 8 or 9. You happily take black (because giving 8 points' komi is worth 2 points more than even to you), and I happily take white (because taking 8 points' komi is worth 3 points more than even to me), and I have essentially a 1-point advantage.

The "correct" komi for a game between you and me (whatever that means) is actually 7.5, but usually neither of us has access to enough information to work that out, giving lots of scope for mind games.

(Of course, this is a problem with normal komi too (and it's also nowhere near the biggest problem in the world at my level anyway), but it's an interesting point - this isn't just a standard "I cut, you choose" problem.)

I really like the idea for club games.

In the past I would have stated 8.5 because I strongly preferred black.

Now, without knowledge of my opponent's preferences, I would bid 7.5 because I slightly prefer black.

I don't think this is a good idea for rated/tournament games, however, because it modifies the winning percentages, and could affect players not involved with this game.

_________________
- Brady
Want to see videos of low-dan mistakes and what to learn from them? Brady's Blunders

I recall an earlier discussion where professional players with black/white splits were discussed. If I recall correctly most had roughly even results, and more excelled at white than black, particularly a cadre of excellent players. These players were adept at neutralizing black's advantage and simplifying the game into a small win with the komi.

What I don't like about komi pie is that if an even split player faces a black/white specialist (and specialist we're maybe talking 2 points komi, if not 1) they'll offer an undersized/oversized komi to counteract that. That color now offers no advantage for the specialist, so strategically they seem best served by giving the opponent the disadvantaged side. A white specialist would end up playing black against a smaller-then-usual komi more often than otherwise.

Aesthetically, that strikes me as wrong. If a player excels at white, don't we want them playing all the games as white they can? If everybody could use 6.5 komi to win disproportionately as white, we'd adjust the komi. But when just some players do, it's a demonstration of their talents, an argument that 6.5 (or the way their opponents play) is too much compensation but that most players aren't good enough to use the advantage.

I think these quirks of strengths give a little spice to Go. It's harder to judge Shusaku's strength because he played black no-komi so much, but isn't it great we've got these masterpieces of black no-komi play to look back at now? I think the same is true for any player who excels at one aspect of the game, be it a color, moyo-play, yose or what have you.

I think a blind auction would make more sense. Each player places how many stones on the board that they would like to play, which ever player places the smallest, gets white. The problem is for ties. Also for multigame matches best of 3 best of 5, you won't have necessarily have alternating games.

I think some sort of "floating" komi is more likely to be adopted in amateur ranks. I for one would love the opportunity to alway play my favorite fuseki, and might even give a stone or two to do it.

I'm becoming more and more of the opinion that for club play, or any situation where you expect to play someone repeatedly, rather than playing around with komi, it's better to use a kadoban system, where there is no komi but the handicap changes every time someone gets some number of wins ahead.

I don't care for the kadoban system. It only works if you play frequently with the same people and only infrequently with other people, and it skews the game towards hoshi-handicap style play. I play mostly even games online, so it isn't unlikely for me to show up and be a stone stronger and to be better at playing without handicaps.

That wasn't the thread about the pie rule. We started arguing about it in this thread:

http://lifein19x19.com/forum/viewtopic.php?f=45&t=5481

Then we moved into a thread called, "The pie rule":

viewtopic.php?f=45&t=5501

"Komi pie: will it ever become widely used?"

It will probably not be widely used. It seems to be a solution in search of a problem.

I agree with oren.

Komi was calculated statiscally from game results with players of same strength. It is usefull for tournaments and rated games.
Apart from that, you either give your oponent a handicap or a reverse komi, which is something both must agree on and the purpose is solely to make an interesting game.