OliverBr wrote:
I remember that I tried one of those position and the engine didn't make any mistakes at that very position. So it's really hard to find and fix something.
Perhaps 5.2.2 is better, but I wouldn't hope too much... If anyone would provide some optimized MSVC binaries, would be nice.
I'll see if I still have any of the debug files from those games. Otherwise I'll make sure I send any new ones to you should similar problems occur.
Fingers crossed because I like the playing style of OliThink.
Cheers,
Graham.
Hi,
Unfortunately I took out Olithink 5.2.1 out of one of my tournaments because of this bug....8 games out of 20 were presented with this bug...
Here is an example:
[d]r2q1rk1/pb1n1ppp/1p2pb2/3n4/3Q4/1PN2NP1/PB2PPBP/R2R2K1 w - - 2 13
Olithink 5.2.1 played here Qf4??
I hope this horrible bug to be fixed soon as I also like the playing style of Olithink
Dr.D
Yes - same experience with 5.2.1. Have you tried 5.2.2 yet Wael?
Hi Graham,
No,but my inner feeling tells me that this awful bug is still alive
Dr.D
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
If I cut and paste this FEN into Arena, then sic OliThink on it, it seems to play normally, without trying to give away the queen. I'm wondering if it's a hash table issue, as it only seems to come up after playing for a little while, and it's tricky to reproduce the strange results that testers are getting.
tmokonen wrote:If I cut and paste this FEN into Arena, then sic OliThink on it, it seems to play normally, without trying to give away the queen. I'm wondering if it's a hash table issue, as it only seems to come up after playing for a little while, and it's tricky to reproduce the strange results that testers are getting.
You are correct,but it's happening in real tournament games with a percantage of 40....and yes,I do think that it's a hash table issue....
Dr.D
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
Dr.Wael Deeb wrote:
Hi,
Unfortunately I took out Olithink 5.2.1 out of one of my tournaments because of this bug....8 games out of 20 were presented with this bug...
Here is an example:
[d]r2q1rk1/pb1n1ppp/1p2pb2/3n4/3Q4/1PN2NP1/PB2PPBP/R2R2K1 w - - 2 13
Olithink 5.2.1 played here Qf4??
I hope this horrible bug to be fixed soon as I also like the playing style of Olithink
Dr.D
When I try the position, I got only the following moves from OliThink:
Qd3
Qd2
but never ever Qf4. What about anybody else?
PS: I would need the log-output of OliThink when such bug occurs: When I tried the position, it's:
Dr.Wael Deeb wrote:
Hi,
Unfortunately I took out Olithink 5.2.1 out of one of my tournaments because of this bug....8 games out of 20 were presented with this bug...
Here is an example:
[d]r2q1rk1/pb1n1ppp/1p2pb2/3n4/3Q4/1PN2NP1/PB2PPBP/R2R2K1 w - - 2 13
Olithink 5.2.1 played here Qf4??
I hope this horrible bug to be fixed soon as I also like the playing style of Olithink
Dr.D
When I try the position, I got only the following moves from OliThink:
Qd3
Qd2
but never ever Qf4. What about anybody else?
Yes,if you import the FEN and analyse the position with Olithink 5.2.1,it will not make that losing move with the queen....I saw this move with my very own eyes when I came back from work today....it was the 19th move of the game,about 6-7 moves after the engine was out of the opening book....so I inturrepted the game and posted this position....I've seen it many times before and I was just hoping that it doesn't happen that often but I found that this bug is present in 4 games out of 10....so I stopped testing the engine....
Dr.D
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
tmokonen wrote:If I cut and paste this FEN into Arena, then sic OliThink on it, it seems to play normally, without trying to give away the queen. I'm wondering if it's a hash table issue, as it only seems to come up after playing for a little while, and it's tricky to reproduce the strange results that testers are getting.
You are correct,but it's happening in real tournament games with a percantage of 40....and yes,I do think that it's a hash table issue....
Dr.D
A hash table problem is possible and could theoretically be solved in the new version as I did touch the hash tables. Let's hope.
OliverBr wrote:I would need the log-output of OliThink when such bug occurs
I'll see if I can find some debug files for you when I get home from work. I'll email you one way or the other.
However, it sounds like the latest version might hopefully be okay if it's a hash issue. Fingers crossed.
OliverBr wrote:I would need the log-output of OliThink when such bug occurs
I'll see if I can find some debug files for you when I get home from work. I'll email you one way or the other.
However, it sounds like the latest version might hopefully be okay if it's a hash issue. Fingers crossed.
Cheers,
Graham.
I am really going to find this bug. It seems to be a hashtable problem after opening book and transition to normal play. But let us focus on 5.2.2:
Please let the engine play and if the bug occurs, post the log and all other parameters like time control at each move etc. This is the only way to reproduce it.
OliverBr wrote:
I am really going to find this bug. It seems to be a hashtable problem after opening book and transition to normal play. But let us focus on 5.2.2:
Please let the engine play and if the bug occurs, post the log and all other parameters like time control at each move etc. This is the only way to reproduce it.
