Test Position 1.Nxb4! followed by Rc7!!

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

acase
Posts: 994
Joined: Wed Mar 15, 2006 1:14 am
Location: Columbus, Ohio USA
Full name: Andrew R Case

Re: Test Position 1.Nxb4! followed by Rc7!!

Post by acase »

GARGAMEL wrote:Hi Andrew

Here are my results of analysis with SPARC 0.3a!
Hello Anton,

I've found that SPARC 0.3a is very good at tactics and it is rapidly becoming one of my favorite engines for tactical positions, thanks for pointing this great engine out to me (by solving my test positions with it).

It also looks like you have a very fast computer getting around 16 million nodes per second with SPARC 0.3a. I only get around 8-9 million nodes per second on my Phenom II 955 running at stock speed (3.2GHZ).

Andrew
acase
Posts: 994
Joined: Wed Mar 15, 2006 1:14 am
Location: Columbus, Ohio USA
Full name: Andrew R Case

Re: Test Position 1.Nxb4! followed by Rc7!!

Post by acase »

Albert Silver wrote:Great position. A rare bird to find a tactical win that is so hard to spot for today's engines and hardware.

Thanks Albert,

I have a few more difficult to very difficult ones up my sleeve still :twisted: . I'll probably post one once a week.

Andrew
acase
Posts: 994
Joined: Wed Mar 15, 2006 1:14 am
Location: Columbus, Ohio USA
Full name: Andrew R Case

Re: Test Position 1.Nxb4! followed by Rc7!!

Post by acase »

I'd like to thank Dann, Louis, and John for trying out this tough position.

Andrew
BBauer
Posts: 658
Joined: Wed Mar 08, 2006 8:58 pm

Re: Test Position 1.Nxb4! followed by Rc7!!

Post by BBauer »

Changing the following line in search.cpp from
Value v = search(pos, ss, beta, depth - 5 * OnePly, ply, false, threadID);
to
Value v = search(pos, ss, beta, depth - 6 * OnePly, ply, false, threadID);
in the Dann Corbit sources stockfish1.6.3dc
stockfish finds Nxb4 in 1h 30 min 32 sec on a 4-cpu machine running at 2GHz

20 +331 17409 486742415 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
21 +343 36497 1052951712 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
22 +331 68652 2011853360 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
23 +315 154550 4558110030 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
24 +315 330544 9741064901 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
24 +359 543187 15896069942 Nxb4 Bxb4 Rc7 Nf6 Rxb7

Btw "- 6 * OnePly" looks sometimes better than "- 5 * OnePly".

kind regards
Bernhard
Dann Corbit
Posts: 12778
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Test Position 1.Nxb4! followed by Rc7!!

Post by Dann Corbit »

BBauer wrote:Changing the following line in search.cpp from
Value v = search(pos, ss, beta, depth - 5 * OnePly, ply, false, threadID);
to
Value v = search(pos, ss, beta, depth - 6 * OnePly, ply, false, threadID);
in the Dann Corbit sources stockfish1.6.3dc
stockfish finds Nxb4 in 1h 30 min 32 sec on a 4-cpu machine running at 2GHz

20 +331 17409 486742415 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
21 +343 36497 1052951712 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
22 +331 68652 2011853360 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
23 +315 154550 4558110030 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
24 +315 330544 9741064901 Nf5 Bf6 Nxb4 Bxb2 Rxb2 Re8
24 +359 543187 15896069942 Nxb4 Bxb4 Rc7 Nf6 Rxb7

Btw "- 6 * OnePly" looks sometimes better than "- 5 * OnePly".

kind regards
Bernhard
You can change the equation via UCI parameters instead of recompiling.