How are your results in handicap games ?

[lorill]

I watched my kgs stats yesterday, inspired by the other thread, and I noticed something strange.

This month, i won 65% of my even games, only 50% of my handi games as black, but I won 76% of my handi games as white (with a decent sample size)!
The difference seems huge, so I wonder: do you have the same differences in your results ?

This site will give you these stats.

[Gresil]

I do overwhelmingly better as black, but it's been a long time since I last got a handi game against a stronger person.

[Solomon]

I don't think it's strange at all...at least, it's what I've been noticing for me as well, just some of my accounts (that are still active), ranked games only:

Code: Select all

All   33   26   56%
Black   12   13   48%
White   21   13   62%
Higher   2   9   18%
Even   13   9   59%
Lower   15   8   65%

Code: Select all

All   47   33   59%
Black   18   16   53%
White   29   17   63%
Higher   12   10   55%
Even   15   10   60%
Lower   20   12   63%

Code: Select all

All   27   26   51%
Black   10   13   43%
White   17   13   57%
Higher   5   7   42%
Even   4   2   67%
Lower   12   11   52%

Code: Select all

All   41   19   68%
Black   12   6   67%
White   29   13   69%
Higher   8   6   57%
Even   8   4   67%
Lower   24   9   73%

[freegame]

I win significantly more as white, no matter if it is handy or even games.
My win percentage with white is around 10% better.

I think the reason is partly that with time constraints the stronger player has an advantage.
Also it is easy to forget to take komi into account when evaluating the board position, resulting tin too submissive play by black (in endgame)
in handicap games when there is no komi it might be that people seem to want to try there luck with reduced handicap, but hardly ever want too much handicap.

[quantumf]

I've been playing high handicap games as white almost exclusively on one of my accounts, and my win ratio is 82%. This is not a sandbagging account, my rank is the same as my other accounts, and there is a lot of games being played (about 60 in the sample). I don't know if high handicap games intimidate weaker players, or if KGS's ranking algorithm is broken. But either way, if someone wants to boost an account, I suggest playing a bunch of 4H+ games (as white).

[palapiku]

my rank is the same as my other accounts

if someone wants to boost an account,

Doesn't compute. Looks like the ranking algorithm works.

[dfan]

I think I remember reading that handicap games favor White (that is, in general, a handicap of n stones for an n-rank difference is not as big as it should be for each player to have a 50% chance of winning).

[Laman]

I think I remember reading that handicap games favor White (that is, in general, a handicap of n stones for an n-rank difference is not as big as it should be for each player to have a 50% chance of winning).

yes, you are right. the point is, that in an even game, black starts and white receive komi as half-move compensation. that means that playing without komi is around half stone handicap, but it is played when the difference of strengths is one rank (one stone) -> white should have half stone advantage in most handicap games.

so proper handicap should be two stones for black and komi for white when the difference is one rank

you can read longer elaborate at senseis

on the other hand, you can take it also like a matter of style, because influence and attack oriented players (like me, for example) could be happier playing with handicap stones while territory oriented and agressive players might use more of their skills playing white

[quantumf]

my rank is the same as my other accounts

if someone wants to boost an account,

Doesn't compute. Looks like the ranking algorithm works.

Yeah, it kinda doesn't compute. I'd have been promoted by now, but for KGS rank-lag. The high-handi-as-white is a fairly recent exercise on this account. I stand by my suggestion.

[SpongeBob]

I think I remember reading that handicap games favor White (that is, in general, a handicap of n stones for an n-rank difference is not as big as it should be for each player to have a 50% chance of winning).

yes, you are right. the point is, that in an even game, black starts and white receive komi as half-move compensation. that means that playing without komi is around half stone handicap, but it is played when the difference of strengths is one rank (one stone) -> white should have half stone advantage in most handicap games.

Cool, didn't know that.

I can only underscore that from an empirical viewpoint - I am currently playing a lot of handicap games with white and my rank is at least half a stone better than with even games.

[gaius]

In my experience, white always has the advantage in 19*19 handicap games; typically, this is more than half a stone. Of course, that advantage is greatly dependent on the time settings as well, but even in moderately-timed handicap tournaments, the strongest players usually win quite a bit more often than the weaker ones.

For that and other reasons, handicap games are inherently unsuitable for rated games or serious tournaments.

Unfortunately, KGS is not alone in implementing rated handicap games, the EGF does the same. In fact, I recently heard that even at the Australian national championships, the players below 3 dan have to play handicap!

[quantumf]

In my experience, white always has the advantage in 19*19 handicap games; typically, this is more than half a stone. Of course, that advantage is greatly dependent on the time settings as well, but even in moderately-timed handicap tournaments, the strongest players usually win quite a bit more often than the weaker ones.

For that and other reasons, handicap games are inherently unsuitable for rated games or serious tournaments.

Unfortunately, KGS is not alone in implementing rated handicap games, the EGF does the same. In fact, I recently heard that even at the Australian national championships, the players below 3 dan have to play handicap!

It might be your experience (and it's definitely mine too), but is it really true? Because another arguement I've frequently heard is that ranks (kgs or national or whatever) are unable to keep up with player improvement, so I would expect black to win a disproportionate number of games, even at high handicap settings (since stronger players, while also improving, are improving slower than lower ranked players).

Anyway, in the interests of being more scientific, I've analyzed some records from South Africa. I've got a copy of the results database from a little while back. For reference, these record tournament games as well as most club matches (assuming both players agree that they're club matches, and will thus affect their ranking).

In this dataset, there are 2333 recorded even games, of which white won 1201, a not surprising 51.5%. There are also 2445 recorded uneven games (at the correct handicap), of which white won 1228, a remarkably level 50.2%. Breaking it down further, there are 1660 games with a handicap of 3 stones or more, of which white won 800 (48.2%)

This is not an enormous dataset, but I think it's big enough to conclude with a high level of confidence that in fact handicap games don't favour white (or black).

[gaius]

Interesting! I am apparently mistaken here. Still, I do believe that the result of handicap games does strongly depend on the time setting, so maybe white will be at an advantage in quick games, whereas black has the advantage in long games? Anyway, interesting data!

[quantumf]

Interesting! I am apparently mistaken here. Still, I do believe that the result of handicap games does strongly depend on the time setting, so maybe white will be at an advantage in quick games, whereas black has the advantage in long games? Anyway, interesting data!

I agree with you here, in that I think the time setting is relevant. My recent run of good form on KGS as white in high handicap games (that I alluded to earlier) were all blitz-ish games.

In the dataset I analyzed, the games would all have been club length (or longer), which is at least half an hour each main time, with reasonably generous byoyomi limits.

[Mark356]

I do better in handicapped games, whether I take black or white. You didn't give an option for that.

My theory on that is because people who need to take a handicap against me are enough weaker than me that the handicap stones aren't really enough most of the time. So I win more handicap games as white. On the other hand, I've gotten good enough at using star point stones that by now I win more handicap games as black than I do even games.

(I used to not like playing black in handicap games, because I hated star point stones and preferred third line stones. Thank goodness I've gotten over that!)