Why don't you post a few of the positions 22 moves down the road do we can compare them?
Just based on a quick look, you are obviously believing that the increased power of the white bishop should make up for the loss of the exchange.
StockFish Help
Sort:
just to be clear, the reason I believe in Rxc5 is because Alphazero played it and won.
in my limited understanding it's a combination of the bishop pin like you said, eliminating the powerful pony, but maybe the most important is the super fast tempo to eventually imprison the Q on h8 ???
Here is what happened in the game...
Feb 17, 2018
0
#4
I'm analyzing this on a crappy PC (weak cpu, 1 thread). Stockfish 8 still can't find Rxc5 after 49 ply. (It prefers 1. Qf4 eval approx +0.25). However, Stockfish 9 finds 1. Rxc5 at ply 39. It actually finds the same line via two different move orders. The other one is 1. Qh4 Rde8 2. Rxc5 bxc5 3. Rf6. The eval for the two move orders is about +0.9 to +1.1.
Next, black responded with Rf8
but what puzzle me is that if I change the number of lines to 9, it will show totally different evaluation,
for example, with 9 lines... at dept>28 it start to drop Rf8 from (0.00) all the way down to (5.51) at dept=31
but with only 1 line open at dept=31 it evaluate (0.00)
do you know what am doing wrong?
(other then failing to communicate a simple question)
in other words... what is the function of the "lines" in the Chess.com fish engine tool?
@Esch, Interesting! how do you switch to fish 9
I only see stockfish.js 8
Feb 17, 2018
0
#7
funindsun wrote:
@Esch, Interesting! how do you switch to fish 9
I only see stockfish.js 8
I'm analyzing the position on my desktop PC, not in Chess.com.
Feb 17, 2018
0
#8
funindsun wrote:
Next, black responded with Rf8
but what puzzle me is that if I change the number of lines to 9, it will show totally different evaluation,
for example, with 9 lines... at dept>28 it start to drop Rf8 from (0.00) all the way down to (5.51) at dept=31
but with only 1 line open at dept=31 it evaluate (0.00)
do you know what am doing wrong?
(other then failing to communicate a simple question)
in other words... what is the function of the "lines" in the Chess.com fish engine tool?
I don't know, this multi-PV stuff is like black magic to me. :)
For my analysis, I just arbitrarily chose multi-PV = 6.
Feb 17, 2018
0
#9
EscherehcsE wrote:
funindsun wrote:
@Esch, Interesting! how do you switch to fish 9
I only see stockfish.js 8
I'm analyzing the position on my desktop PC, not in Chess.com.
Install your favorite GUI, then download Stockfish 9 from here:
Thank you
I had the engine Stockfish 6 play against stockfish 9, I accidentally only gave an opening book to Stockfish 6, perfect 2017. Stockfish 6 played with the white pieces in the game.
Anyway, on move 9, Stockfish 9 only searched to a depth of 13, and it has low search depths throughout this game, did I do something wrong?
I gave both engines 2048 MB of hash, my laptop has 6GB of RAM, so did I give them too much? I have read that for the best performance, use half of your computers RAM, but not more than that. So do I need to use 3GB or less between the two engines?
interesting
Mar 17, 2018
0
#13
y-e-l-l-o-w wrote:
I had the engine Stockfish 6 play against stockfish 9, I accidentally only gave an opening book to Stockfish 6, perfect 2017. Stockfish 6 played with the white pieces in the game.
Anyway, on move 9, Stockfish 9 only searched to a depth of 13, and it has low search depths throughout this game, did I do something wrong?
I gave both engines 2048 MB of hash, my laptop has 6GB of RAM, so did I give them too much? I have read that for the best performance, use half of your computers RAM, but not more than that. So do I need to use 3GB or less between the two engines?
Stockfish 9 should be searching deeper than Stockfish 6. So, yeah, something's wrong.
Mar 17, 2018
0
#14
Stockfish 9 at ply 38 opts for Rxc5.
Evaluation goes up and down, but it's +1.00 and up.
Mar 17, 2018
0
#15
EscherehcsE wrote:
y-e-l-l-o-w wrote:
I had the engine Stockfish 6 play against stockfish 9, I accidentally only gave an opening book to Stockfish 6, perfect 2017. Stockfish 6 played with the white pieces in the game.
Anyway, on move 9, Stockfish 9 only searched to a depth of 13, and it has low search depths throughout this game, did I do something wrong?
I gave both engines 2048 MB of hash, my laptop has 6GB of RAM, so did I give them too much? I have read that for the best performance, use half of your computers RAM, but not more than that. So do I need to use 3GB or less between the two engines?
Stockfish 9 should be searching deeper than Stockfish 6. So, yeah, something's wrong.
13...b6 is a big blunder. I didn't check the moves after that. Something is definitely hosed.
Mar 17, 2018
0
#16
pfren wrote:
Yes, Stockfish is late to get there, dur to its sharp pruning. Actually I use ASMFish (Stockfish code complied in asembly), which obviously prunes Rxc5 initially and analyses it as good for white after manually accepting the move (1.Rxc5 and 1.Qh4 lead to the same position).
A good player (OTB or correspondence) would naturally consider sacrificing the exchange, as the Bb3 becomes a monster.
Here Asmfish finds Rxc5 at 14 ply depth and holds it. In this particular position, Asmfish is uber fast, gets to the 35th ply in 5.7 seconds.
From my previous tests, it turned out that Asmfish is faster to reach shallow depths (up to 22 - 23) and later it gets behind with SF8.
I set this position here on the chess.com SF tool for unlimited time and a single line.
at depth=44 the engine recommend next move to be Qf4 (+0.06)
I believe Rxc5 is a better move yet my computer seem to freeze at depth#44.
Perhaps you have better hardware and engine knowledge to run this position and see if SF ever change its mind from Qf4 to Rxc5?
~I'm totally new to this tool and might just doing it all wrong