Only if you play go for the sole purpose of winning
2 Likes
I have absolutely no problem with komi being set up purposefully to make a draw impossible.
1 Like
I have absolutely no problem with the tie breaker. I see it just as another way to compensate for the first move advantage of black. We might just aswell say ‘White gets 6 Points extra and wins all ties as compensation.’, but just using 6.5 komi seems more elegant.
I really like the decisiveness of having no draws. A game starts simple, gets more and more complex in the midgame and then simpifies again, Finally its either win or loss. Also, tournaments are easier to organize if don’t have to consider three possible outcomes.
4 Likes
I think we need to broaden the debate to consider at least four options:
- Fractional komi (e.g., 6.5 points)
- Integer komi (e.g., 6), white wins all ties
- Integer komi (e.g., 7), black wins all ties
- Integer komi, ties are allowed
Option 1 has sub-options of what the fraction should be, e.g., half? inspired by a multiple of pi?
With options 2 and 3, the result could be W+0 and B+0, respectively.
Options 1, 2, and 3 differ only in style.
Note that Ing rules officially use option 3, by specifying a komi of 8 and that black wins all ties.
1 Like
If two players play a perfect game, the result should be a tie. By saying one colour wins all ties means the other player is required to overplay to have a chance of winning.
Edit note. Please know I’m not implying that all tied games were played perfectly.
4 Likes
…sure…
But I’m pretty sure by the time we discover the perfect game, Go will have lost its mystery, and thus its appeal (that and there’s the whole heat death of the universe thing),
so I find such reasoning irrelevant.
1 Like
That’s not really how perfect play works. In tic-tac-toe, you can easily force a draw, since you can work out all the variations: there’s only three starting moves (due to symmetry) and in the “mid-game” most moves become forcing.
With connect-four, which is also solved (first player can force a win), this becomes a lot harder. The way people solved the game is by brute-force, so there is not a set of rules one can follow to ensure a victory: the only strategy we know is to consider all the possible outcomes. Sure, you can follow the perfect solution, but if your opponent makes a mistake, you still need to know how to take advantage of that.
Go is very much a worse problem, since there is literally not enough time / space in our universe to solve a 19x19 game by brute-force, it has about 10^170 possible positions, whereas our universe only contains about 10^80 number of atoms. And that’s only the number of possible positions, the number of possible games is way larger. Go has been ‘probably weakly solved’ for a 7x7 board (which draws with 9 komi), although this means that there is a strategy that is probably a perfect strategy, but we haven’t proven it is the perfect strategy, nor have we found the perfect strategy for any position, just for the one game tree that has been studied.
It’s similar to AI, now that we have computers beating humans consistently, does that mean that Go has lost its purpose? Sudoku’s are also “boring” puzzles, since you can always just find a solution by trial-and-error, but that doesn’t mean it’s not fun to do sudoku’s.
4 Likes
The first three of your options are identical (i.e., a specific game will score the same winner with all of them,) so I don’t see any value on debating them.
The key thing to realize IMHO is that as it stands, the fractional komi means that white wins ties (assuming a 6 komi is correct in the first place.) I wonder if, instead of having a 6.5 komi, we said “6 komi and white wins ties” people would complain that it’s unfair to whoever plays black? The optics in this case seems to be better with the former notation.
1 Like
Go has been strongly solved on a 5x5 board and the komi is 24 (Japanese.) Given that it is solved and a human with enough skill is capable of playing perfectly, would anybody really think that adding a half-point to white would be in any way fair? I expect not.
3 Likes
Honestly, I don’t really see much value in debating them either.
However, the original post and subsequent discussion is partly a debate over option 1 vs 2/3. I just wanted to poke fun at it a bit by pointing out that there is even a stylistic difference between 2 vs 3 to consider.
Imagine the absurdity of two camps zealously arguing over “komi = 6, white wins ties” versus “komi = 7, black wins ties”.
You say that you see no value in debating this issue, but here you are suggesting that option 1 is better than 2. Tsk tsk tsk.
1 Like
I don’t think it is fair, but it is convenient. As hiryuu mentioned earlier, draws are a serious problem in high level chess competition, so much even that a weaker player can work towards a draw instead of winning.
Another thing to consider, is that historically white is the stronger player (e.g. title holder), and during a game black is contesting that. If black wants to show he is stronger, he has to get strictly more points than his opponent. A draw does not really give him any right to gain the title. If you can’t force your opponent to surrender, you can’t win a war.
But I’m also playing devil’s advocate here, I do agree that allowing draws is more fair. Then the problem becomes different: what is the right komi for 19x19?
1 Like
That is not what I am saying.
What I am saying is that once humans have discerned a ‘perfect play’ for Go, people are going to begin to follow it, which leads to games becoming more and more repetitive (like what happens very often with amateur fusekis), leading the game to become less interesting.
if we have a ‘perfect solution’ then the refutations to mistakes would be available, and without such we have not discovered a perfect solution, and thus if someone plays with ‘perfect play’ no one will know if they have, and so I see this as irrelevant.
I’m pretty certain that the perfect solution for go is complex enough that no human can ever learn it. Or the perfect solution to chess (which is a lot “easier” to remember) for that matter. The closest we can get is achieving locally perfect play, or in other words, joseki.
2 Likes
And that’s my point. The ‘tie for perfect play’ argument seems moot to me, since we can’t know the solution, and even if we did, Go would become repetitive and boring to the point where people won’t care.
It is a nice sentiment, but very often ties are annoying when you are either competitive or in tournament, so I’m incredibly comfortable with a tiebreaker. In casual play, however, I am all for ties, since that might make ratings more accurate.
1 Like
but how do you know that if two players played perfectly it would be a draw… 
no matter how good you are, you can’t pick a middle one out of a row of 4 objects… If you know what I mean 
or maybe a better example if you have two players taking stones one by one from a bowl of 5, no matter how good they are, one will have more
I do not know what my point it, but I felt like sharing my deep thoughts
5 Likes
Well, the scoring in Go makes sure that both players have an integer number of points, so you can set the komi to make their score equal, which would be a draw. Or you set the komi .5 point below or above that ‘perfect score’ to define the perfect game as won by black or won by white.
1 Like
The idea is that it should be a tie. like, if we have two players taking stones one by one from a bowl of five, you should give an extra stone to the second player to tie the game.
3 Likes
Ah, that makes sense
I may have something to add here. The New Zealand Rules were used in the artificial game example given above by mekriff - https://online-go.com/game/3442486 . Here is a real example of an online game using this rule set : https://online-go.com/game/8870838 (Im glad a bug report has been filed as I won this game unfairly I thought).
These rules were devised in the 1980’s in New Zealand specifically to allow for draws in Tournaments where the NZ rules are used - ie a komi of 7 points. There is no problem organizing tournaments with the occurrence of draws as suggested by xeldrak - if you use software like “McMahon” it handles it all for you.
There were two draws in the 2017 NZ Open, I was involved in one and it had a seki - it was a “two common point seki” - if it had been a “one point common seki” - a draw would not have been possible.
My personal opinion - I like the possibility of a draw, it is still rare and usually i feel quite good about it afterwards if it happens to me.
4 Likes
What ties? 