For engine authors

Graham Banks · Post by **Graham Banks** » Fri Mar 10, 2023 11:30 am

Since I started using 6 piece Syzygy tablebases a couple of days ago, I've come across two engines that lose with an illegal move message if the opponent takes a piece to leave a drawn 6 piece syzygy tablebase position.

Could you please check that your engine doesn't have this issue?

Graham Banks · Post by **Graham Banks** » Sat Mar 11, 2023 4:36 am

Clover 3.3.1 is now fixed.

Uralochka 3.39d - still waiting for a fix.

Velvet 5.1.0 now loses on time without even reaching the tablebases stage. Author notified.

Chessqueen · Post by **Chessqueen** » Sat Mar 11, 2023 4:50 am

Graham Banks wrote: ↑Sat Mar 11, 2023 4:36 am Clover 3.3.1 is now fixed.

Uralochka 3.39d - still waiting for a fix.

Velvet 5.1.0 now loses on time without even reaching the tablebases stage. Author notified.

My antivirus told me that Clover 3.3.1 contained a Trojan Horse before it downloaded and it prevented me from downloading it

RubiChess · Post by **RubiChess** » Sat Mar 11, 2023 7:08 am

Graham Banks wrote: ↑Sat Mar 11, 2023 4:36 am Velvet 5.1.0 now loses on time without even reaching the tablebases stage. Author notified.

The search reaches tablebase probing using the wdl/.rtbw files long before the root position. On strong hardware like TCEC you see tablebase hits starting from the very first move in the game.
What kind of storage do you use for your tablebase files? Especially the .rtbw files should be on very fast storage like ssd. Using 6-men .rtbw on normal harddisk will slow down the engine and can lead to timeouts and/or worse play.

Regards, Andreas

Modern Times · Post by **Modern Times** » Sat Mar 11, 2023 7:15 am

He has them on a normal hard drive. Not the best idea.

Graham Banks · Post by **Graham Banks** » Sat Mar 11, 2023 7:58 am

Modern Times wrote: ↑Sat Mar 11, 2023 7:15 am He has them on a normal hard drive. Not the best idea.

I'll see if I can find an easy way to switch.

RubiChess · Post by **RubiChess** » Sat Mar 11, 2023 10:20 am

Or just stay at 5-men. Not the worst option in my opinion.

Chessqueen · Post by **Chessqueen** » Sat Mar 11, 2023 11:19 am

RubiChess wrote: ↑Sat Mar 11, 2023 10:20 am Or just stay at 5-men. Not the worst option in my opinion.

I believe that 5 men is more than sufficient, and let the two engines find out for themselves the best continuation of the endgame stage as a matter of fact are we trying to find out which of the two engines play all the stages of the game or simply let the EGTB find out the best and continuation in certain chess endgames

NOTE: As a matter fact, I do NOT even use EGTB at all for the same reason stated above. Do GMs use open chess book for their games such as Silman’s Complete Endgame Course chess knowledge? It should be the same for chess engines

For engine authors

For engine authors

Re: For engine authors

Re: For engine authors

Re: For engine authors

Re: For engine authors

Re: For engine authors

Re: For engine authors

Re: For engine authors