Cinnamon 2.1 sacs it’s queen for nothing on Picochess.

[ScallyBag]

Hi Giuseppe,

I’ve been trying out your new version of Cinnamon 2.1 in Picochess. It was paired agains Tasc 2.2, Cinnamon has the Black pieces in a 90 seconds per move game.

This is the move sequence:

1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.c3 dxc3 5.Nxc3 Nc6 6.Bc4 Qc7 7.Qe2 Nf6 8.h3 Bd6 9.Nb5 Bb4+ 10.Bd2 Bxd2+ 11.Qxd2 Qb6 12.Nd6+ Ke7 13.Rd1 Qxf2+ ?? 14.Kxf2 Nxe4+ 15 Nxe4

I have replayed the game several times at different time intervals, but Cinnamon gives it’s queen away every time at move 13?

Here’s the FEN before Blacks bad move:

r1b4r/pp1pkppp/1qnNpn2/8/2B1P3/5N1P/PP1Q1PP1/3RK2R b K - 4 13

I was using your Raspberry Pi Binary, I also compiled it myself with the same results.

Thanks,

Alan Cooper.

[gekomad]

try this version https://drive.google.com/file/d/1n0Glkq_XE3U6XaGnkhHxY4xZhY-z6mGC/view?usp=sharing

On Thu, Apr 4, 2019 at 11:56 AM Al ‘Scally’ Cooper notifications@github.com wrote:

Hi Giuseppe,

I’ve been trying out your new version of Cinnamon 2.1 in Picochess. It was paired agains Tasc 2.2, Cinnamon has the Black pieces in a 90 seconds per move game.

This is the move sequence:

1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.c3 dxc3 5.Nxc3 Nc6 6.Bc4 Qc7 7.Qe2 Nf6 8.h3 Bd6 9.Nb5 Bb4+ 10.Bd2 Bxd2+ 11.Qxd2 Qb6 12.Nd6+ Ke7 13.Rd1 Qxf2+ ?? 14.Kxf2 Nxe4+ 15 Nxe4

I have replayed the game several times at different time intervals, but Cinnamon gives it’s queen away every time at move 13?

Here’s the FEN before Blacks bad move:

r1b4r/pp1pkppp/1qnNpn2/8/2B1P3/5N1P/PP1Q1PP1/3RK2R b K - 4 13

I was using your Raspberry Pi Binary, I also compiled it myself with the same results.

Thanks,

Alan Cooper.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/gekomad/Cinnamon/issues/47, or mute the thread https://github.com/notifications/unsubscribe-auth/AL1hGYygdRdZYTvQy701A40hT9y46oZMks5vdcxSgaJpZM4ccZLg .

[ScallyBag]

Hi Giuseppe,

Thanks for the quick response, it’s still playing the same Queen sac. After 9.Nb5 it seems Cinnamon believes it’s mate in 1 against it, hence the checks and Queen sac.

Thanks,

Al.

[gekomad]

I can't replicate the issue, please check this https://drive.google.com/file/d/1Dqndr56Kq7mxydt2dauqRlQE8-wYmLp7/view?usp=sharing thanks

On Thu, Apr 4, 2019 at 7:49 PM Al ‘Scally’ Cooper notifications@github.com wrote:

Hi Giuseppe,

Thanks for the quick response, it’s still playing the same Queen sac. After 9.Nb5 it seems Cinnamon believes it’s mate in 1 against it, hence the checks and Queen sac.

Thanks,

Al.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/gekomad/Cinnamon/issues/47#issuecomment-479997763, or mute the thread https://github.com/notifications/unsubscribe-auth/AL1hGXDqMC02nBTbNXQ1bSg7BmNCJY_4ks5vdjsjgaJpZM4ccZLg .

[ScallyBag]

Hi Giuseppe,

It’s still playing the same Queen sacrifice in all 3 time controls within Picochess, that’s Fischer, game in x or move in x. However outside of Picochess running your engine natively via ./cinnamon, it plays a different move when I give it the FEN of the position.

I’m at a loss to think what it might be?

Thanks,

Al.

[gekomad]

I think the issue is on trasposition table

Il ven 5 apr 2019, 16:01 Al ‘Scally’ Cooper notifications@github.com ha scritto:

Hi Giuseppe,

It’s still playing the same Queen sacrifice in all 3 time controls within Picochess, that’s Fischer, game in x or move in x. However outside of Picochess running your engine natively via ./cinnamon, it plays a different move when I give it the FEN of the position.

I’m at a loss to think what it might be?

Thanks,

Al.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/gekomad/Cinnamon/issues/47#issuecomment-480286459, or mute the thread https://github.com/notifications/unsubscribe-auth/AL1hGRqGdNEMBBFRXOGjNFjaHShndDRPks5vd1c_gaJpZM4ccZLg .

[ScallyBag]

Ok, if it’s any help I just compiled Cinnamon v2.0 and it doesn’t happen on that version, all seems well.

Al.

[gekomad]

How can I replicate the moves?

Il ven 5 apr 2019, 19:04 Al ‘Scally’ Cooper notifications@github.com ha scritto:

Ok, if it’s any help I just compiled Cinnamon v2.0 and it doesn’t happen on that version, all seems well.

Al.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/gekomad/Cinnamon/issues/47#issuecomment-480350249, or mute the thread https://github.com/notifications/unsubscribe-auth/AL1hGZLQxjuBRVPDwoGmE2UK3HAq_4KBks5vd4IigaJpZM4ccZLg .

[ScallyBag]

I can’t think how. Cinnamon v2.1 & v2.2 produce the Queen problem only on Picochess not runnng it stand alone via ./cinnamon.

Your old version 2.0 works perfectly fine as tried again this afternoon.

Al.

[gekomad]

@ScallyBag I think I found the problem can you re-check ? here new exe http://cinnamonchess.altervista.org/tmp/cinnamon22a thanks

[ScallyBag]

Hi Giuseppe,

Thanks for the new binary, unfortunately it’s still happening ....

https://i.postimg.cc/6ykGRY5P/5-EC0-BE24-2493-4-F2-C-B304-0-BA81-F16-CBA4.jpg

Al.

[tpoppins]

Cannot reproduce with the Windows x64 v2.2a binary under Cute Chess GUI using the TC of 90s/move. Even with hash as low as 1MB Cinnamon finds 13...Qa5 at d=8 and sticks with it.

Must be a bug with the Raspberry P binary.

[ScallyBag]

Hi Giuseppe,

I compiled v2.2a myself and the Queen sac no longer happens within Picochess.

Thanks very much for your patience and help.

Al.

[gekomad]

Great. thanks

[LocutusOfPenguin]

so, is this solved now? What was the (real) cause of these problems? Shouldnt be related on compiling. So, i worry abit (for picochess).

Jürgen

[gekomad]

yes, It seems a compilation problem

[LocutusOfPenguin]

i meant:

• if solved, should be closed.
• i would feel better, if the cause is clear! Sadly, the old bins (mentioned above) no longer available. So, if y take an old version => error, and with 2.2A no longer (13.Apr). For me - when i read the above text, it seems to be a mixture of compile, some source code bugs (say: transposition table) or no problem on windows.

Thats strange. But OK, i dont need to care too much.

Yours, Jürgen