True or False Chess is a Draw with Best Play from Both Sides
Sort:
MorganJackson
Mar 2, 2020
0
#2661
With so many possible games “perfect play” is impossible to determine, at least it is using today’s best engines. That said to eliminate the possibility that the outcome of perfect play from both sides is in fact a win for black is incorrect. Who knows maybe as a hypothetical engine’s rating approaches infinity it becomes able to win every time as black and can at best draw with white when playing itself. Maybe moving first is actually a disadvantage at such a high rating because it creates vulnerabilities only a god-like engine can properly exploit yet to us mere mortals it is clear that moving first increases winning chances.
Mar 2, 2020
0
#2662
kindaspongey escribió:
Haven’t there been certain endgames that were believed (by the authorities) to be drawn, but were discovered to be wins after sufficient computer investigation?
Yes, KBB vs KN is one of them.
Mar 2, 2020
0
#2663
Laskersnephew escribió:
"These days, a properly installed Stockfish can hit tablebases as early as move 12 or move 15 of opening phase"
If there are seven or fewer pieces on the board.
Nope, they hit TBs even with all 32 pieces although you need to leave them thinking for hours at a time, still very impressive.
"Nope, they hit TBs even with all 32 pieces although you need to leave them thinking for hours at a time, still very impressive."
Why do they do that when there are no 32-piece--or even 8-piece tablebases?
Mar 2, 2020
0
#2665
Ziryab escribió:
@KibiDangoman The word you are looking for is blunder.
Don't rely too heavily on engine analysis. There are still positions that Stockfish gets wrong, or does not get right fast enough.
For instance, I played Black against Stockfish from this position last week. It took some time before the initial evaluation of +2.00 gave way to +0.50, and several moves before it recognized the position as 0.00.
What computer are you using? Even the slow chess.com Stockfish 10 shows ~1.10 immediately and drops to 0.8 in about 3 seconds, less than a minute later and it evaluates 0.36. This is all without TBs.
Mar 2, 2020
0
#2666
Laskersnephew escribió:
"Nope, they hit TBs even with all 32 pieces although you need to leave them thinking for hours at a time, still very impressive."
Why do they do that when there are no 32-piece--or even 8-piece tablebases?
Hitting a tablebase just means the engine has accessed the TB, the search tree the engine generates can go deep enough to reach positions with 7 pieces and that's when it'll ask the TB for an evaluation, i.e., it hits the TB.
Mar 2, 2020
0
#2667
Also, 8-piece tablebases or more can be generated with FinalGen for a single position albeit with some limitations.
Mar 2, 2020
0
#2668
drmrboss wrote:
Ziryab wrote:
@KibiDangoman The word you are looking for is blunder.
Don't rely too heavily on engine analysis. There are still positions that Stockfish gets wrong, or does not get right fast enough.
For instance, I played Black against Stockfish from this position last week. It took some time before the initial evaluation of +2.00 gave way to +0.50, and several moves before it recognized the position as 0.00.
@Ziryab,
You just need 6 men syzgy tablebase to solve your issue.
http://oics.olympuschess.com/tracker/index.php
The torrent is free and legal to download as the owner publicy share those database.
You dont need to spend a few minutes to check the position. Stockfish can tell you the outcome 100% correctly within 1 second.
P.S. I wouldnt trust Stockfish evaluation without Tablebase. These days, a properly installed Stockfish can hit tablebases as early as move 12 or move 15 of opening phase ( Theoretically engines in these days can see door to door, from opening to ending in her Principal Variations or her predicted lines).
You missed the point. I could use an example with more pieces. Positions still exist that Stockfish, even with tablebases, will evaluate incorrectly. Of course I use engines. For some people—maybe you’re one of them—engines are a crutch, and people put too much faith in their evaluations.
Stockfish played this position perfectly without tablebases, but it did not evaluate the position correctly. Of course with six-piece tablebases, it would need no calculation, and would know in a nanosecond that it is drawn.
Mar 2, 2020
0
#2669
Prometheus_Fuschs wrote:
Ziryab escribió:
@KibiDangoman The word you are looking for is blunder.
Don't rely too heavily on engine analysis. There are still positions that Stockfish gets wrong, or does not get right fast enough.
For instance, I played Black against Stockfish from this position last week. It took some time before the initial evaluation of +2.00 gave way to +0.50, and several moves before it recognized the position as 0.00.
What computer are you using? Even the slow chess.com Stockfish 10 shows ~1.10 immediately and drops to 0.8 in about 3 seconds, less than a minute later and it evaluates 0.36. This is all without TBs.
I don’t recall. It may have been my iPad. The point, of course, is that 0.36 is inaccurate. The position is a dead draw with correct play. 0.00 would be accurate.
Mar 3, 2020
0
#2670
Ziryab escribió:
Prometheus_Fuschs wrote:
Ziryab escribió:
@KibiDangoman The word you are looking for is blunder.
Don't rely too heavily on engine analysis. There are still positions that Stockfish gets wrong, or does not get right fast enough.
For instance, I played Black against Stockfish from this position last week. It took some time before the initial evaluation of +2.00 gave way to +0.50, and several moves before it recognized the position as 0.00.
What computer are you using? Even the slow chess.com Stockfish 10 shows ~1.10 immediately and drops to 0.8 in about 3 seconds, less than a minute later and it evaluates 0.36. This is all without TBs.
I don’t recall. It may have been my iPad. The point, of course, is that 0.36 is inaccurate. The position is a dead draw with correct play. 0.00 would be accurate.
Anyone who knows how engines work would know such low evaluations only indicate advantage and not a winning position, this is especially true for an endgame. Whether it should evaluate with 0.00 a drawn position with "advantage" is a subjective decision.
Mar 3, 2020
0
#2671
There is no advantage in a simple textbook draw. It is your chess knowledge that is missing.
My limited knowledge of chess engines is built upon the thin data compiled through only a few years of near daily use since 1989.
Mar 3, 2020
0
#2672
Ziryab escribió:
There is no advantage in a simple textbook draw. It is your chess knowledge that is missing.
My limited knowledge of chess engines is built upon the thin data compiled through only a few years of near daily use since 1989.
Advantage is always the product of holes in anybody's chess understanding, the only true objective evaluations are winning, equal or losing positions.
Why did the chicken cross the road?
Mar 3, 2020
0
#2674
Tepeyac escribió:
Why did the chicken cross the road?
Que la virgencita lo protega 
Mar 14, 2020
0
#2675
Ziryab escribió:
Prometheus_Fuschs wrote:
Ziryab escribió:
@KibiDangoman The word you are looking for is blunder.
Don't rely too heavily on engine analysis. There are still positions that Stockfish gets wrong, or does not get right fast enough.
For instance, I played Black against Stockfish from this position last week. It took some time before the initial evaluation of +2.00 gave way to +0.50, and several moves before it recognized the position as 0.00.
What computer are you using? Even the slow chess.com Stockfish 10 shows ~1.10 immediately and drops to 0.8 in about 3 seconds, less than a minute later and it evaluates 0.36. This is all without TBs.
I don’t recall. It may have been my iPad. The point, of course, is that 0.36 is inaccurate. The position is a dead draw with correct play. 0.00 would be accurate.
Technically, 0.36 is about a third of a pawn of an advantage, such advantage in virtually any endgame is meaningless.
Mar 14, 2020
0
#2676
Prometheus_Fuschs wrote:
Ziryab escribió:
Prometheus_Fuschs wrote:
Ziryab escribió:
@KibiDangoman The word you are looking for is blunder.
Don't rely too heavily on engine analysis. There are still positions that Stockfish gets wrong, or does not get right fast enough.
For instance, I played Black against Stockfish from this position last week. It took some time before the initial evaluation of +2.00 gave way to +0.50, and several moves before it recognized the position as 0.00.
What computer are you using? Even the slow chess.com Stockfish 10 shows ~1.10 immediately and drops to 0.8 in about 3 seconds, less than a minute later and it evaluates 0.36. This is all without TBs.
I don’t recall. It may have been my iPad. The point, of course, is that 0.36 is inaccurate. The position is a dead draw with correct play. 0.00 would be accurate.
Technically, 0.36 is about a third of a pawn of an advantage, such advantage in virtually any endgame is meaningless.
When the computer sees a clear draw, the numbers are 0.00. That 0.36 is equal is another matter. In the diagram that I posted, the draw is clear, simple, and undeniable. Nonetheless, the engine running sans tablebases takes more than five seconds to reach that conclusion.
My point—simple and clear—don’t put all your stock in engine analysis. It will retard your chess development.
I’m done arguing this point with people who understand neither chess nor chess engines.
0.36 advantage in the opening is not meaningless to a very strong chess player. It improves practical chances.
But of course don't put all your stock in an evaluation of 0.36 advantage by a chess engine.
About 4 years ago a grand master gave two positions [here on chess com] which the chess engines got wrong --because I am a human and can think out side the box--I solved both positions.
Mar 15, 2020
0
#2679
ponz111 wrote:
0.36 advantage in the opening is not meaningless to a very strong chess player. It improves practical chances.
But of course don't put all your stock in an evaluation of 0.36 advantage by a chess engine.
About 4 years ago a grand master gave two positions [here on chess com] which the chess engines got wrong --because I am a human and can think out side the box--I solved both positions.
The discussion the past two pages emanates from the position here, where Boguljubow missed a draw with 70...Kg4.
Ziryab wrote:
@KibiDangoman The word you are looking for is blunder.
Don't rely too heavily on engine analysis. There are still positions that Stockfish gets wrong, or does not get right fast enough.
For instance, I played Black against Stockfish from this position last week. It took some time before the initial evaluation of +2.00 gave way to +0.50, and several moves before it recognized the position as 0.00.
Here's a screenshot of Stockfish 10 not finding the true 0.00, and also showing evaluations while playing out the position. It was six moves before Stockfish showed 0.01 and ten until 0.00.
Of course, it was pointed out quite accurately, that with tablebases, Stockfish sees this position correctly in an instant.
Mar 15, 2020
0
#2680
ponz111 wrote:
0.36 advantage in the opening is not meaningless to a very strong chess player.
Actually it is.
Even professionals frequently choose to "spend" (so to speak) that small amount of "advantage" to get positions where they have better practical chances.
Only weak players think of 0.3 out of the opening as meaningful.
Of course, 0.3 could be good for white, but without seeing the position for context, 0.3 all by itself is meaningless. It could even offer black better chances in a real game.