"Snowballing your advantage" in Go?

[Bill Spight]

OK, let's take two 5 kyu players who are evenly matched from move 51 onward and let them finish a game where White has gained 10 points up through move 50. As 5 kyus they are pretty skillful, but not so that they can consistently break their opponent's plan, protect their own plan, or know how to fight. So Black will win a fair number of games. Komi is 7 points. Do you think that White will win most of the time, and if so, by how much on average (the median)?

This isn't meant to be combative, but your situation/question is sort of a non question.

What does "evenly matched" mean anyways? That they're the same level in attack defense, shape, concepts, endgame, reading ability, etc etc etc? It is such a hypothetical situation that it might as well not exist.
Your rank is an average of all your strengths and weaknesses and is a fluidly changing thing. If someone makes a mistake or perceives a mistake, they might play differently. They might value certain things differently. Their mental state might be completely different. To say they are the "same" doesn't mean anything.

Also, what does a "10 point" lead mean? Does that mean in territory, in general board strength, weak group strong group ratio, thickness?

Komi doesn't matter much either, especially at an amateur level. Huge endgame swings are possible due to misreads, valuing certain areas more than others, or possibly just letting the opponent getting all the big sente moves.

My teacher once said of my game, that when i perceived i was behind i would go crazy and play moves that made no sense. I thought that was what you were supposed to do, pick fights when you're behind and the such. However, my teacher's point was "No, at all times you play proper Go and wait for your opponent to make a mistake. If they don't play a mistake that means they played a good game. But if they are at the same general level then mistakes will happen"

Given the same starting point, board situation, same players, and with white having a lead the result is still a massive toss up. Someone might end up winning 8 out of 10 times, or be generally even.

The thing is, that there are so many variables and things that happen in Go that you can never pin point to just one "thing" that made you win or lose. And especially at an amateur level, there isn't a thing that happens that you can grab and lead with and "snowball" into a landslide victory.

You are overestimating the randomness of amateur play. If what you believe were true, ratings systems would break down. (IMO, they are not worth much below 15 kyu, but work pretty well at the SDK level and above. )

[often]

You are overestimating the randomness of amateur play. If what you believe were true, ratings systems would break down.

No, i'm pretty sure you're underestimating the randomness of amateur play.

If we took two players with solid ranks, it doesn't mean the win loss will be 50/50. One might have a consistent advantage over the other based on where his knowledge lies compared to the other. Rating systems only work for people who play a wide range of people for a decent amount of games.

For example, if we had an aggressive 5k who likes to fight and a passive but solid playing 5k player, the aggressive person might have a hard time playing against the passive player and have a lower win rate. That doesn't mean the "solid" 5k is a 4 kyu or that they aren't evenly matched, it just means he can do well against that one player.

[Bill Spight]

You are overestimating the randomness of amateur play. If what you believe were true, ratings systems would break down.

No, i'm pretty sure you're underestimating the randomness of amateur play.

If we took two players with solid ranks, it doesn't mean the win loss will be 50/50. One might have a consistent advantage over the other based on where his knowledge lies compared to the other. Rating systems only work for people who play a wide range of people for a decent amount of games.

For example, if we had an aggressive 5k who likes to fight and a passive but solid playing 5k player, the aggressive person might have a hard time playing against the passive player and have a lower win rate. That doesn't mean the "solid" 5k is a 4 kyu or that they aren't evenly matched, it just means he can do well against that one player.

You are using the randomness of amateur play to denigrate improvement in the opening. There is such an effect, but it is small. For bots that play completely randomly (as long as moves are legal), for instance, proper komi is 0. The randomness wipes out the advantage of the first move. But for pros the komi is only around 7 points. That is the maximum average penalty for randomness, in practice. It is true that much larger blunders are possible, but even human beginners do not play randomly. The large swings do not eliminate the value of the first play. Nor do they eliminate the value of the other moves in the opening. They reduce it by a few points on average, that is all.

If you are 9 stone stronger than someone and you see them come out of the opening with a 10 point advantage, or even a 15 point advantage (for instance, if their opponent took gote instead of sente), you can easily spot how their opponent could have won the game later on. But that still is not an argument against playing a good opening. Just because the game is not over until it is over does not excuse earlier bad play.

[often]

No, no, and no.

I will make this my last post on this subject. You disagree and i don't care, but here is my rebuttal.

You are using the randomness of amateur play to denigrate improvement in the opening.

No, I am using the randomness of play to say that there really isn't a way to "snowball your advantage" as laid claim by the first post.

But for pros the komi is only around 7 points. That is the maximum average penalty for randomness, in practice.

No, komi is not penalty for randomness. Komi exists to even the playing field due to advances in opening theory for pros.

The large swings do not eliminate the value of the first play. Nor do they eliminate the value of the other moves in the opening. They reduce it by a few points on average, that is all.

No, the large swings show that you can't really "snowball your advantage". What good is a 10 point lead in the opening if you play a 40 point mistake in the midgame?

If you are 9 stone stronger than someone and you see them come out of the opening with a 10 point advantage, or even a 15 point advantage (for instance, if their opponent took gote instead of sente), you can easily spot how their opponent could have won the game later on. But that still is not an argument against playing a good opening.

If you're 9 stones stronger than someone, it's a given that you are probably going to win due to being better at aspects of the game, not just because you were ahead by 10 points, otherwise handicap wouldn't make sense.
Again, all of this is being viewed through the perspective of "snowballing your advantage".

Just because the game is not over until it is over does not excuse earlier bad play.

It doesn't, but it shows that just because you have the advantage does not mean you will have a better statistic chance of winning, especially in the game of Go.

[Bill Spight]

No, no, and no.

I will make this my last post on this subject. You disagree and i don't care, but here is my rebuttal.

You are using the randomness of amateur play to denigrate improvement in the opening.

No, I am using the randomness of play to say that there really isn't a way to "snowball your advantage" as laid claim by the first post.

We agree on that.

the simple example is a common phrase that "a 10-30 point in the opening might not exist at all". people make such horrible misplays that they're not aware of in the midgame that the opening advantage/disadvantage rarely carries through.

That's what we disagree about. That does denigrate skill in the opening.

[quantumf]

It doesn't, but it shows that just because you have the advantage does not mean you will have a better statistic chance of winning, especially in the game of Go.

Be interesting to establish whether this opinion, or Bill's reverse position, is actually a fact.

The part that bothers me about your position is that it reduces the game to absurdity. If you can not establish a meaningful lead early on, that you can *usually* maintain to the end, why bother with the opening at all? Why not just slap down your first 50 stones randomly? Your position seems to be that middle game fighting decides the game. Actually, I agree with you on that, which is why the opening is largely, perhaps mostly about positioning yourself to have an advantage in the middle game. This aggregate of firm territory, potential territory and thickness for fighting is what is described as a "lead"

[Bill Spight]

It doesn't, but it shows that just because you have the advantage does not mean you will have a better statistic chance of winning, especially in the game of Go.

Be interesting to establish whether this opinion, or Bill's reverse position, is actually a fact.

At one point often asked what it meant to take a lead of 10 points in the opening. One way that can happen is by making a small play and taking gote unnecessarily, thus losing a move for only a few points. That's almost as bad as giving an extra handicap stone. As I said, if what often believes is true, rating systems don't work. The advantage of an extra handicap stone, which is worth more than 10 points and less than 30 points, and so falls into the range that he indicates, does carry through more often than not, instead of rarely, as he claims.

As for an advantage not giving you a better chance of winning, that's what advantage means.

Your position seems to be that middle game fighting decides the game. Actually, I agree with you on that, which is why the opening is largely, perhaps mostly about positioning yourself to have an advantage in the middle game.

Excellent point.

[hyperpape]

What good is a 10 point lead in the opening if you play a 40 point mistake in the midgame?

It leaves you ahead after the opponent follows it with a 31 point mistake.

[often]

Your position seems to be that middle game fighting decides the game. Actually, I agree with you on that, which is why the opening is largely, perhaps mostly about positioning yourself to have an advantage in the middle game.

And if your middle game sucks despite a good opening, what's the point of a good opening advantage?

This statement only proves my position. It still comes down to the middle game, be it from a deficit, advantage, or neutral standing after the opening.

[Kirby]

I think this discussion is too generalized, because there are many different types of "opening scenarios" you might encounter.

If black and white play a calm opening, and white made more mistakes than black - maybe black has an advantage, but it might be not that significant.

If the opening involves some large-scale joseki, it's quite possible for one of the players to make a mistake that's almost irrecoverable in a game between two players of the same rank (e.g. taisha or magic sword joseki screw ups).

So I think it's only safe to make specific assessments, such as:
* In this particular game, white's opening was so bad, he's at a serious disadvantage.
* White had a better opening this game, but there's still a lot of chance for black to make a comeback.

---

If we want to go back to generalized statements, I'd say, middle game fighting skill is necessary, but not always sufficient.

[Kirby]

It still comes down to the middle game, be it from a deficit, advantage, or neutral standing after the opening.

I personally think the statement is a little extreme, and not always true. But probably true more often than the converse (i.e. if you have to choose middle game skill or opening game skill, I'd think middle game skill will get you more bang for your buck).

Actually, I think there's somewhat of a consensus here: the game is largely decided in the middle game, but the opening gives you either an advantage or a disadvantage going into it.

[Uberdude]

Often, let's say we pick 200 random KGS 5ks. I then construct a board position after the opening in which black is around 10 points ahead (according to me or Bill or some pro or you or a monte carlo bot or whoever). We then pair up those 200 5ks and get them to play 100 games and record how many black wins. Is your position that the opening is so irrelevant that the answer will follow a binomial distribution with mean 50, in other words the same as if they just played all those 100 games from a blank board? My guessestimate is the mean would be perhaps 55-60. Probably rather similar to if 100 5ks were playing as black against 100 6ks. If we did the same experiment with 200 3ds I expect black would win perhaps a mean of 65 due to the higher skill and consistency of those dan players and their ability to maintain a lead throughout the game and the smaller sizes and frequencies of their blunders. With pros, perhaps 90+?

An aside:

Perhaps I am overestimating my ability (and their weren't really enough 6d+ players to test this theory well), but I feel back when I played on OGS if I had a 10 point lead after the opening you'd need to be a 6d+ to beat me as lots of time means far fewer mistakes and blunders. For example this game below I feel by move 7 white set off down a losing path, and I did not let him leave it for the rest of the game. A poor opening lead to a loss despite his best middlegame efforts. Now perhaps you could say I played a good middlegame too, and better than his, but that good middelgame was built on the foundation of an opening lead. On the other hand I can lose to a 1d even after a 10 point opening lead in a face-to-face tournament game with reasonable time limits like one hour each, though it's quite likely to happen in overtime which one could say is poor time management on my part.

https://online-go.com/api/v1/games/136960/sgf

[Kirby]

It's a good point, Uberdude, that the opening might have a more significant effect on the result for higher level players. For pros, it might be very hard to come back from a bad opening, whereas for someone just learning the game - well, anything can happen.

Another argument for studying the opening, which I want to bring up, is that knowing an opening well can be helpful in tournaments.

If we assume that the middle game is the most important part for deciding the game outcome, it might be good to spend a lot of thinking time on that part. If you have 40 minutes of main time, it would be a pity to have only, say 10 minutes, to think during the middle game.

Therefore, if you know the opening very well, you can breeze through that part in a short amount of time, and spend a lot of time thinking on the middle game.

If you don't know the opening well, at higher levels, you either spend too much time on the opening during a tournament, or you are at a disadvantage that you have to push past in the middle game in order to win.

[hyperpape]

Uberdude, it's really unsporting to be precise about what we mean. This is the sort of argument that only works if you let people make vague generalizations and inconsistent claims.

[Loons]

The original idea often seems to occur for me like this:


The opponent takes territory for thickness

The opponent takes territory for thickness

The opponent takes territory for thickness

The opponent tries to reduce

The thick player isolates the stones and asks the opponent to show their eyes

The opponent tries to reduce

The thick player isolates the stones and asks the opponent to show their eyes

The opponent tries to reduce

The thick player isolates the stones and asks the opponent to show their eyes