Chess Forum
[ mod this thread ] << - < - > - >>
| From | Message |
| leo_london 10/20/2004 11:04:42 [ report this post ] |
 Subject: White/BlackMessage: Something has puzzled me for a while. I was always led to believe that playing "white" (making first move in other words) was a distinct advantage. I am sure 20 or more years ago, when I used to follow some of the major championships, a draw playing black was considered an achievement*, whereas a draw with white was a setback*. I have noticed on GK that there seems little difference in the stats for white/black wins even amongst the top players. Does the white advantage still apply, but just at IM and GM level ? *maybe could have worded that better, but you get my drift. |
| chrisp 10/20/2004 11:56:16 [ report this post ] | Well Message: There is still no doubt that playing white is an advantage - having the first move inevitably leads to more centre control, more rapid development of pieces and generally the initiative (this is all obviously dependant on the opening played, but applies to most of the major opening systems). However, the simple fact of having the first move is not sufficient to provide white with a winning position. IMO, the advantage of being white is mainly that your position can often become a winning position with only a relatively small error by black, whereas white must make a more significant error to provide black a winning position - a small error often simply allows black to fully equalise the position. I think that years ago, and still today, that white has and will always have a slight advantage in the overall statistics, but that this advantage will always be small - anything else, and the game of chess would probably have lost it's appeal years ago. I don't know if these ramblings are of any assistance to you, so I'll shut up now and go back to playing some chess :-) chrisp |
| chrisp 10/20/2004 12:00:20 [ report this post ] | Stats Message: I've just had a quick look at my own stats at GK, and I actually lose more games with the white pieces than with black!! Although I also have more draws with black and less wins. Whatever all this means, i haven't a clue!! chrisp |
| ormus 10/20/2004 12:22:42 [ report this post ] | Black, White, who cares??? Message: I think the lower the rating class, the lower the initial advantage to White. The tempo advantage gained from having the first move can easily be squandered away by something as simple as putting one of your pieces on a bad square in the opening - finding good squares for pieces (with either color) isn't easy, even for GMs. It really doesn't matter what color you play with at the lower rating levels because too many mistakes are made to cash in with the extra tempo. an interesting, nonsense game... 1.e3 e5 2.e4! (now a KP opening with Black to make all decisions) ...Nf6 3.Nf3 (A reverse Petroff opening, until) ...Ng8 4.Bb5 Nc6 (Ruy Lopez with an extra tempo!) 5.Bxc6 (The rest is a matter of style ~Capablanca) ~Chess Life, 1995 Yearbook |
| fmgaijin 10/20/2004 14:34:00 [ report this post ] | GK Stats DO Support White Advantage Message: If you take even a quick look at the GK Games Database, the data there DO support the "small advantage to White" theory. All studies at the IM/GM level show about a 5% edge to White (55-45). The GK stats for 1900+ players using NORMAL openings (e4, d4, Nf3, c4) all range around this figure. The only time the White edge begins to disappear comes with less-common openings (f4, b4, etc.). Either the earlier posters misread the stats or expected a LARGER edge for White than the "small" but steady and significant 5%. |
| i_play_slowly 10/20/2004 18:10:35 [ report this post ] | It's all in the mind Message: I disagree with the notion, "There is still no doubt that playing white is an advantage." Fischer, said, "I think it's almost definite that the game is a draw theoretically." His theory is supported by the fact that a computer playing against itself will draw. I have seen it argued that White's advantage is, consequently, entirely psychological. Black, believing himself to have the disadvantage, loses--it's a self-fulfilling prophecy. Computers draw because psychology is not part of their game. They do not know they are supposed to lose. * Supposing a class of children, while being taught chess, were told that Black usually wins because White, having the first move, usually makes the first mistake. As long as these children believed the fallacy, they might tend to win more with Black. Perhaps Black's victories at Gameknot can be similarly explained by the fact that its members, like the children in my example, or like computers, by and large play in blissful ignorance that they are 'supposed' to lose. |
| fmgaijin 10/20/2004 19:43:35 [ report this post ] | But Fischer . . . Message: . . . had a solidly better score with White over his career than with Black. Notably, Fischer won more with Black than OTHER GM's of his time (similarly, Kasparov in his prime could say the same), but overall he won more and lost less with White than with Black. BTW, computers do NOT always draw when playing themselves or other computers. In fact, studies of computer self-play as well as computer vs. computer results show that White wins more . . . . |
| bonsai 10/21/2004 02:40:22 [ report this post ] | to i_play_slowly Message: I do not think it's all in the mind. Being a tempo up is useful. I certainly do not claim that the initial position is anything but a draw with best play by both sides (in fact I'm convinced it is a draw with *best* play), but it is (often) easier to win as white - as someone else already pointed out black needs to make fewer mistakes in order to end up in a losing position due to being a tempo down to start with. And (as fmgaijn already implied) the notion that computers playing against either themselves or other computers always draw is nonsense, they just don't blunder heavily (tactically) most of the time, but they do win and lose (just let your own computer program play a dozen games against itself and you are bound to see a couple of decisive games). |
| i_play_slowly 10/21/2004 17:30:15 [ report this post ] | It's still all in the mind Message: True, Fischer did win more and lose less with White. It is precisely because of his profound understanding of White that his observation carries weight. Kant's observations re: the limitations of Reason are especially meaningful because few people have ever demonstrated the capabilities of Reason better than Kant. He, if anyone, was in a position to know its limitations. Similarly, few people have ever been as qualified as Fischer in assessing the limitations of White. * True, a computer playing against itself can produce a decisive game, nor did I claim that a computer will ALWAYS draw against itself. Still, it will tend to draw, and that tendency is significant. Given that a computer playing against itself would tend to err equally as White and Black, the tendency should be towards victories for White, if there was indeed a first-move advantage. * One more quote to support my claim that there is no first-move advantage: "The better player will win with either color, but it takes longer with Black" (Larry Evans). * And one last quote to support the alternative: "The advantage of first move is increased rather than diminished in correspondence chess" (Adrian Hollis). |
| fmgaijin 10/21/2004 20:23:02 [ report this post ] | "Theoretical Draw" Does Not Mean Equal Message: What Fischer meant was that IF Black plays well, there is no particular reason that Black should lose. No one here argues otherwise. However, many positions evaluated as "slightly worse" are in fact drawn but easier for one side to play. That is, as several people have pointed out, in these positions Black needs FEWER mistakes in order to lose than White, which accounts for White's small edge in winning %. In GM play as in computer self-play (and YES, the major computer self-play studies show more White victories than Black--just run a Google search and you will find a number of these), that translates into many draws but a few more wins for White--nothing decisive, just a consistent statistical edge. As for myself, having grown up like Fischer playing Swiss-system tournaments where one must try to win with Black as well as White, I have always tried to win with both colors, but I find that I must take more risks and hence incur more losses with Black in attempting this--for example, all of my GK losses thus far have come with Black in sharp openings where the consequences of my mistakes were more disastrous than when I slipped with White. |
| bonsai 10/22/2004 01:41:06 [ report this post ] | Message: Actually the Larry Evans quote "The better player will win with either color, but it takes longer with Black" surely does support the theory that there's an advantage to playing with white, why else would it take longer to win as black (for the better player)? And I very much agree with fmgaijn's post above. |
| leo_london 10/22/2004 05:46:38 [ report this post ] | Having Message: started the thread, thought I should do a little research. Found a couple of observations with a simple Google search. On one site the writer claims that "white wins some two thirds of the time at the highest level ". He then speculates that " Some entertain that black would be superior at still higher levels "..presumably levels as yet un-attainable, going on to say " if both sides could play perfectly black would gain an advantage because the game would tend towards a lock-down, where being forced to move ( white's tempo advantage ) would be a disadvatage " Another site, more concerned with statistics, maybe betting and odds (is there much betting on chess ? ) gives the following..." When the top 20 humans play white is considered worth about 50 ratings points advantage, against computers only about half that. In 12 games Kasparov v Deep blue white won six times, black merely once. Interestingly in the decade leading up to the Deep blue matches Kasparov showed an advantage of 50 points rating playing white, since then only 35 points " I dont know if any of that is of interest..probablly not. I would have liked to find the win/loss/draw percentages of GM's playing white/black over the last century. |
| ormus 10/22/2004 07:19:04 [ report this post ] | Rating Equality Message: interesting, I wondered that myself. If White has the advantage given two players of average strength, how much do you have to increase the average rating of the players with the Black pieces so that the White advantage is nullified. 50 rating points seems low but that is at top-level GM play. I doubt it's a linear function. If you take the absolute opposite of top-level super GM play and look at games between two beginners, you can see that White has no advantage at all... {beginners may start a game 1.h4 f5 2.b4 c5 - in other words, totally random moves that don't exploit the initial tempo advantage in any way} There is probably a stat average for every rating-class that equalizes the White/Black advantage over a large pool of games... i.e. White(1500) - Black(1575) 50/50 This is all academic though, it shouldn't really interfere with your playing or your attitude entering a game with the Black pieces. They {whoever 'they' is} can only draw out averages for White/Black over large pools of games, no one can predict ahead of time which games will end up losers for Black {aside from possibly GOD, or maybe 33rd Degree Masons} No doubt some rating junkie out there has done an actual study on this... |
| fmgaijin 10/22/2004 10:25:51 [ report this post ] | Hey, Leo! Message: If you have a large database (for example, I use ChessBase with about 3 million games in my bases), you can actually get SOME of the kinds of stats you wanted. Just restrict the players by rating and data (advanced search options) and you can find out % scores after any set of moves (or even no moves--the starting position). These scores are more accurate for the past 30 years because older players have no ratings listed (BE--"before Elo") and do not have all of their games in the database (sometimes just the "best" or most famous games; the others are lost in the mists of time). NOTE: If you turn off the rating restrictions, you wind up with games by 1100 players in the Moldovan Under-10 Championship, which defeats the purpose.) That's how I and others derived the "55%" stat. You can also pick out a particular player and get that player's lifetime stats (again, more complete for big names and more recent players). |
| fmgaijin 10/22/2004 10:28:26 [ report this post ] | Hey, Ormus! Message: As you surmised, some stats junkies have done things like figure out what players (or a particular player) tend towards in certain situations. For example, some of them have looked at whether a player is most likely to win/lose/draw after a previous loss or string of losses . . . |
| i_play_slowly 10/22/2004 11:45:34 [ report this post ] | Re: Larry Evans quote Message: Evans surely implies that having the first move does not lead to an increased percentage of victories for White. White's victories may come sooner, but where is the advantage? An advantage that does not result in an increased percentage of victories is no advantage at all. Black's victories may take more time, but there is no disadvantage to taking longer if the game ends in victory all the same. * I would like to know if White tends to win in machine vs. machine tounaments. Does anyone know how I might find out? |
| leo_london 10/22/2004 11:45:37 [ report this post ] | fmgaijin Message: Thanks for the information. I haven't access to a large database at present. However I might try going through the records of some of the " greats " of the game, that information should be easy enough to find on line with a hour or so to spare. The consensus seems to be that, the higher the rating, therefore the fewer mistakes (unforced errors if you like)..the more advantage to white ? The additional speculation about " perfect play " seems unresolved. I guess two computers with infinite capabilities would solve that, but we would not be around to see it. |
| bonsai 10/23/2004 03:05:14 [ report this post ] | the last time I'm going to comment on that quote Message: i_play_slowly: I think you are misunderstanding Evans. He just says that if a player is sufficiently strong then he will win with either colour against a much weaker player. I.e. if one player is much stronger then even the advantage of having white won't help his opponent. This certainly does not mean that there is no advantage to having white. If even under those circumstances black victories take more time then that shows that black has to overcome a certain advantage that white has before going on to win - i.e. this does point to an advantage in playing white. This advantage may not exist "theoretically"/"objectively", but in practical play it simply does exist and white does score better than black does. |
| fmgaijin 10/24/2004 14:32:12 [ report this post ] | Final Fischer Thought Message: Just remembered a further Fischer comment on this. GM Robert Byrne (a boyhood friend of RJF) tells how back in the 1970's he was showing Bobby some of his games where he played aggressively with Black (trying to channel his inner Fischer) and Bobby criticized him, commenting that "you have to equalize with Black BEFORE you start playing for the win." Interesting observation from a player who, as noted, scored well with Black. As the last few posters have suggested, he probably won with Black because he was stronger than his opponents--a difference of 100 Elo points in playing strength is MUCH more significant statistically than the 5% swing for playing White. |
Post a reply to this message: |
