MatPlus.Net

 Website founded by
Milan Velimirović
in 2006

23:50 UTC
ISC 2024
 
  Forum*
 
 
 
 

Username:

Password:

Remember me

 
Forgot your
password?
Click here!
SIGN IN
to create your account if you don't already have one.
CHESS
SOLVING

Tournaments
Rating lists
1-Jan-2024

B P C F





 
 
MatPlus.Net Forum General Help with a proof game
 
You can only view this page!
(1) Posted by Bob Baker [Friday, Jun 9, 2023 00:12]

Help with a proof game


Leither123 at chess.com asked me to check the position below for a possible unique proof game in 31.0 moves. I gave the problem to Stelvio 1.3 which ran for 28 hours without finding any solution. He asked if another program might solve it and I said I would try to find out. Interesting to me is that the position was obtained by adding one move to a 30.0 move game which Stelvio finds in less than a minute is unique.

B4b2/r2ppp2/n1prRrR1/kb5p/P1RQ2n1/N1R2N1r/KBPPPP2/7q
62
 
(Read Only)pid=24818
(2) Posted by Joost de Heer [Friday, Jun 9, 2023 07:44]

(= 15+15 )

 
 
(Read Only)pid=24820
(3) Posted by Andrew Buchanan [Friday, Jun 9, 2023 14:03]

How about the position at 30.5?
 
 
(Read Only)pid=24824
(4) Posted by Bob Baker [Friday, Jun 9, 2023 18:16]

Retracting Nh6-g4 and Qg4-d4 gives a position that Stelvio 1.31 can find a unique 30.0 move proof game for in under a minute. Euclide 1.11 agrees after analyzing the position for more than five hours.
 
   
(Read Only)pid=24829
(5) Posted by Michel Caillaud [Saturday, Jun 10, 2023 21:50]

Testing the 30.5 first moves with Stelvio 1.31 provides cooks in about 4 hours. For example :
1.g4 a5 2.h4 a4 3.h5 a3 4.h6 a×b2 5.h×g7 h5 6.a4 Nh6 7.Na3 b1=R 8.g8=R Rb6 9.Rg6 Rd6 10.Ré6 b5 11.g5 b4 12.g6 b3 13.g7 b2 14.g8=R b1=R 15.Rh4 Rb3 16.Bb2 Ra7 17.Qb1 Ba6 18.Qa2 Bb5 19.0-0-0 Na6 20.Kb1 Rh3 21.Rg3 Qa8 22.Rç3 Qh1 23.Bg2 Kd8 24.Ba8 Rg8 25.Qd5 Rg6 26.Ka2 Rf6 27.Nf3 ç6 28.Rhç4 Kç7 29.Rg1 Kb6 30.Qd4+ Ka5 31.Rg6 Ng4
 
   
(Read Only)pid=24832
(6) Posted by Bob Baker [Sunday, Jun 11, 2023 11:10]

Michel, thank you for cooking the 31.0 move game. For some reason, it didn't occur to me to ask Stelvio to solve the position a half-move rather than a full move past the unique position. It's a lesson I won't forget.
 
   
(Read Only)pid=24836
(7) Posted by Bob Baker [Sunday, Jul 9, 2023 06:46]

I had asked for help in determining the uniqueness of a position in 31.0 moves that had been obtained by adding one full move to a unique 30-move proof game. Michel Caillaud kindly sent me a cook found by testing the first 30.5 moves using Stelvio 1.31, with a solving time of about 4 hours.

Subsequently I decided to see if I could shorten the solving time for the 30.5 move position by adjusting Stelvio's configuration values. My best result to date is a cook found by Stelvio 1.4 with a solving time of 56:12 minutes. Other Stelvio users may like to know the setup that gave this result. Below I have listed the configuration values I used that would be expected to have an effect on the solving time.

java17 startup flag: -Xmx30g

Stelvio 1.4 configuration values:

parallelize = true
numStrategySeekers = 6
numStrategyPlayers = 6
strategyQueueSize = 50000
positionCacheMaxOverallExponent = -1
positionCacheSplitExponent = 2
positionCacheMaxChainLength = 100
positionCacheFullEvictionThreshold = 0.85
positionCacheSizeAdjustment = 0
 
 
(Read Only)pid=24896

No more posts


MatPlus.Net Forum General Help with a proof game