Life In 19x19

luigi wrote:I've finally settled for komi instead of the pie rule.

Don't buckle under on the pie/komi issue. Komi is an aesthetic Hiroshima.

Pie Go is simply uncharted territory. Not bad, bugged or odd territory. The first move doesn't have to be awful - just bad enough that you don't care if you get pied.

It's just common decency. Black gets to have an extra stone half of the time, thoughout the game. Not an extra, catalog-perfectly placed stone half of the time. That would be hugely inequitable, requiring extreme score jerrymandering in the form of komi.

If the Go tree is as astronomical as it's purported to be, there should still be plenty of game left over after a restricted first move. Now you've got a game of 300 unrestricted placements instead of 301 unrestricted placements.

Go is a lot bigger than the relatively diminuitive Chess, where chopping out a major tree branch would be devastating.

luigi wrote:As for the komi/pie rule discussion, maybe I will opt for komi after all. I've been following your interesting discussion on the pie rule in the other thread and I understand the reasons why you don't like it, namely that it's not actually an even trade and that it disrupts the natural flow of the game. I also understand that it's kind of a waste to use the pie rule if the game already has a scoring system which can be easily adjusted to balance the game with great precission.

The problem with the pie rule in go, IMO, is that there are no viable first moves. Any move on the second line or higher favors black significantly, any move on the first line favors white significantly. Ergo, you cannot compensate for the first move advantage in go with the pie rule, because there is no suitable move to do so. Now you could probably find some balanced position after 3, 4 or 5 moves, but why bother when there is already a perfectly fair and extremely elegant solution in the form of komi?

My only issue with komi is the fact that it should be different for players with different skills. For example, in 5x5 Go, good players should set komi at 25 points (the whole board), but for absolute beginners that would be too much. For this reason, I wouldn't use a fixed komi even if I knew the value which made the game balanced at the highest levels of play. I think it's better to let the players negotiate it before the game in this way: first Black suggests a komi, then White may choose to swap sides.

I've often heard the claim that fair komi (here used in the sense of giving both players equal chances) depends on the playing strength of the players, but I've never seen anyone provide even a shred of evidence for this assertion. Your example doesn't work, because a 25 komi game on 5x5 has only two possible outcomes: it's either a tie, or white wins. Black can never win. Under those conditions, any non-perfect black player is at a disadvantage with 25 komi.

MarkSteere wrote:Don't buckle under on the pie/komi issue. Komi is an aesthetic Hiroshima.

HermanHiddema wrote:The problem with the pie rule in go, IMO, is that there are no viable first moves. Any move on the second line or higher favors black significantly, any move on the first line favors white significantly. Ergo, you cannot compensate for the first move advantage in go with the pie rule, because there is no suitable move to do so. Now you could probably find some balanced position after 3, 4 or 5 moves, but why bother when there is already a perfectly fair and extremely elegant solution in the form of komi?

Even 2-7 or 2-10? There's a thread elsewhere on this topic, btw.

HermanHiddema wrote:The problem with the pie rule in go, IMO, is that there are no viable first moves. Any move on the second line or higher favors black significantly, any move on the first line favors white significantly. Ergo, you cannot compensate for the first move advantage in go with the pie rule, because there is no suitable move to do so. Now you could probably find some balanced position after 3, 4 or 5 moves, but why bother when there is already a perfectly fair and extremely elegant solution in the form of komi?

I agree.

HermanHiddema wrote:

My only issue with komi is the fact that it should be different for players with different skills. For example, in 5x5 Go, good players should set komi at 25 points (the whole board), but for absolute beginners that would be too much. For this reason, I wouldn't use a fixed komi even if I knew the value which made the game balanced at the highest levels of play. I think it's better to let the players negotiate it before the game in this way: first Black suggests a komi, then White may choose to swap sides.

I've often heard the claim that fair komi (here used in the sense of giving both players equal chances) depends on the playing strength of the players, but I've never seen anyone provide even a shred of evidence for this assertion. Your example doesn't work, because a 25 komi game on 5x5 has only two possible outcomes: it's either a tie, or white wins. Black can never win. Under those conditions, any non-perfect black player is at a disadvantage with 25 komi.

You're actually proving me right. You're saying that 25 komi in 5x5 Go is only fair for perfect players, and that everyone else should use a lower komi. Ergo, fair komi (or, more precisely, "balanced komi") depends on the playing strength of the players.

(It shouldn't be that hard to be a perfect 5x5 Go player, by the way.)

hyperpape wrote:

HermanHiddema wrote:The problem with the pie rule in go, IMO, is that there are no viable first moves. Any move on the second line or higher favors black significantly, any move on the first line favors white significantly. Ergo, you cannot compensate for the first move advantage in go with the pie rule, because there is no suitable move to do so. Now you could probably find some balanced position after 3, 4 or 5 moves, but why bother when there is already a perfectly fair and extremely elegant solution in the form of komi?

Even 2-7 or 2-10? There's a thread elsewhere on this topic, btw.

Yes. I would take black in a no komi game with 2-7 or 2-10, or any other 2nd line move as the opening move. Of course, I can offer no proof for this assertion, only that it is how I feel about it at my current playing strength.

MarkSteere wrote:Don't buckle under on the pie/komi issue. Komi is an aesthetic Hiroshima.

Pie Go is simply uncharted territory. Not bad, bugged or odd territory. The first move doesn't have to be awful - just bad enough that you don't care if you get pied.

Mark, I think Herman is right: komi is much more precise than the pie rule. Also, since the winner is already decided on points, isn't it natural enough to make it balanced by giving the second player some extra points? I think this is about as elegant or inelegant as the pie rule.

luigi wrote:You're actually proving me right. You're saying that 25 komi in 5x5 Go is only fair for perfect players, and that everyone else should use a lower komi. Ergo, fair komi (or, more precisely, "balanced komi") depends on the playing strength of the players.

(It shouldn't be that hard to be a perfect 5x5 Go player, by the way.)

Although perfect play for 5x5 is known (solved by Erik van der Werf some time ago), there are no human players that can play it. The tree is just too large. Strong players will, no doubt, have no trouble winning by 25 points, bu their play is not perfect.

Using 5x5 to show that komi should be different for different strength players is a bit strange. Komi acts as a balancing tool and doesn't quite work when a game is solved.

I would believe the claim that komi should differ for different strength if you can provide data to support it. http://www.online-go.com/forums/thread. ... eadID=1296 has some data to 2007, and a preliminary inspection suggests that 6.5 isn't very far from even (about 49% winrate for black). Of course, even games between unequal players with randomly assigned colours will drift the winrate towards 50%, so the statistics aren't completely solid.

HermanHiddema wrote:Although perfect play for 5x5 is known (solved by Erik van der Werf some time ago), there are no human players that can play it. The tree is just too large. Strong players will, no doubt, have no trouble winning by 25 points, bu their play is not perfect.

If a player has no trouble winning always by 25 points as Black, then he is displaying perfect play. Every line of play with a forced 25 points win is perfect play.

illluck wrote:Using 5x5 to show that komi should be different for different strength players is a bit strange. Komi acts as a balancing tool and doesn't quite work when a game is solved.

illluck, it doesn't matter if it's solved or not. A 25 komi for 5x5 Go would be balanced for all dan players, regardless of whether they know the solution or not. They would draw every time, whereas if two complete newbies play 5x5 Go, it's obvious that a 25 komi won't balance the game.

luigi wrote:illluck, it doesn't matter if it's solved or not. A 25 komi for 5x5 Go would be balanced for all dan players, regardless of whether they know the solution or not. They would draw every time, whereas if two complete newbies play 5x5 Go, it's obvious that a 25 komi won't balance the game.

I disagree. A game being balanced doesn't mean a draw every time. A game being balanced means both players should draw with perfect play. 25 komi balances the game even for beginners, it's just unlikely that games would finish in a draw. If I play a 3 stone handicap game against a 10k, the game is balanced against me, but I'll still win every time. Balancing in the sense of komi just means that neither player should have an advantage at the beginning, not that the outcome is a foregone conclusion.

FWIW generally, I don't think komi is an unpleasant hack, and I think it has pros and cons just as the pie rule does. Komi isn't an option in a lot of games, such as hex, chess, shogi etc, because the result isn't decided on a score basis - there is no score. Komi only makes sense in score based games.

Mark, I'd really recommend hitting about 10k or so before feeling that the pie rule wouldn't have a fundamental change to the nature of a Go game that may make it less elegant to play. Sure, I have no doubt it could be a balanced game with the pie rule, but the opening would feel very different to play. I wouldn't be keen on that, and that's primarily down to a subjective dislike having gained a reasonable understanding for aiming for optimal moves from the outset. Fiddling my opening to make the most of a single suboptimal stone would feel rather more of a hack to balance the game than a score modification a la komi. I could equally see players advocating the pie rule, particularly those who haven't already been accustomed to Go, it's just not my cup of tea and I don't see anything fundamentally wrong or flawed about using komi (at least, nothing clearly more flawed than other solutions such as the pie rule).

The 5x5 example does not generalize, because with the correct komi, black cannot win. As such, it has no implications for larger sizes.

Lets talk about 5x6 go instead. With perfect play, it is a 4 point win for black. Can you show that a 4 point komi is wrong for weaker players?

topazg wrote:Mark, I'd really recommend hitting about 10k or so before feeling that the pie rule wouldn't have a fundamental change to the nature of a Go game that may make it less elegant to play.

The pie rule is inelegant. A poorly placed first stone is inelegant. Does pie make Go less elegant to play, overall? No. I don't need to play Go 500 times to realize that. Go isn't fragile. It wouldn't have survived the millennia if it were overly sensitive to the opening move. The innocuous pie rule won't harm Go in a general sense. An ugly, though more balanced, first move should actually improve the gameplay overall. Go is every bit as much of a contest with the pie rule as without.

The question is, "Is the pie rule less elegant than komi?" No, it isn't.

1. Komi makes Go an unequal goals game, first off. Black has to totally whomp White, beyond the komi offset, and White has to make that not happen. White has no chance of winning outright. All by itself that makes komi an aesthetic Hiroshima.

2. Komi is variable. The higher the level of play, the closer the final score. Choosing the right komi value is essential to fair play. What if the komi value isn't obvious? How many times have you lost and said, "If only the komi was set right…."?

No rule is uglier than superko, not even komi. But komi holds a distant second place.

MarkSteere wrote:1. Komi makes Go an unequal goals game, first off. Black has to totally whomp White, beyond the komi offset, and White has to make that not happen. White has no chance of winning outright.

Prior to komi, the game was played using alternation of who went first. White often won. Today, White often wins by more than the value of komi.

MarkSteere wrote:Komi is variable. The higher the level of play, the closer the final score. Choosing the right komi value is essential to fair play. What if the komi value isn't obvious? How many times have you lost and said, "If only the komi was set right…."?

Think about this statistically. The variance of the margin of victory may be higher without the advantage for Black being any different. As Herman asks: can anyone actually demonstrate that komi should be higher for weaker players? It's true that there are more games decided by lopsided scores at lower levels, but that includes more games where white wins by 30 points as well. Do they balance out? It's an empirical question, and not one that I'm sure we've properly answered.

MarkSteere wrote:No rule is uglier than superko, not even komi. But komi holds a distant second place.

Yeah, well, y'know, that's just like, your opinion, man.