Re: [Gnushogi-devel] Comparing GNUShogi 1.2 and later versions

[h . g . muller]

Hi Yann,

I hacked UCI2WB to make an xShogi-CECP adapter. I tried it on GNU Shogi
1.4.1, which I happened to have installed, and it worked! It played a game
with sente against CrazyWa. And it did not play bad at all: it was
actually winning, and CrazyWa conceeded that GNU Shogi could checmate it
in 3 moves for some time. But instead of doing the required moves GNU
Shogi at some point did something else, after which CrazyWa mated it in 1
through a Silver drop. See the game below.

I attached the XS2WB source. Link with lpthread. The XBoard command I used
was:

xboard -variant shogi -fcp "./xs2wb debug ./gnushogi" -scp ./crazywa

(all binaries were in the current directory). A match can then be started
though the Mode -> Machine Match menu.

Regards, H.G.

[Event "Computer Chess Game"]
[Site "hgm-xboard"]
[Date "2018.02.19"]
[Round "-"]
[White "xs2wb"]
[Black "CrazyWa 1.0.5"]
[Result "0-1"]
[TimeControl "40/60"]
[Variant "shogi"]
[Annotator "1... -0.15"]

1. c4 Gf8 {-0.15/7 1.7} 2. h4 Kd8 {-0.40/6 4} 3. Ke2 g6 {-0.55/6 3} 4. Sf2
g5 {+0.11/8 2.0} 5. Bxh8+ Sxh8 {+0.01/6 2.7} 6. address@hidden Gg7 {+0.84/6 
1.4} 7.
h5 a6 {+0.75/6 1.5} 8. Kd2 a5 {+1.16/5 2.1} 9. Gfe2 Sg9 {+1.16/5 0.9} 10.
Sc2 b6 {+1.87/6 1.0} 11. i4 b5 {+2.01/6 1.0} 12. Rh4 f6 {+2.10/5 1.0} 13.
i5 f5 {+1.87/6 1.1} 14. f4 fxf4 {+1.92/6 1.4} 15. Rxf4 Ge9 {+1.33/5 1.3}
16. Rh4 Sg8 {+1.68/5 1.4} 17. h6 hxh6 {+2.36/5 1.0} 18. address@hidden h5
{+0.31/6 1.8} 19. Ri4 Gf8 {-0.35/3 1.6} 20. hxh9+ Sxh9 {-0.91/6 1.0} 21.
address@hidden Sh8 {-0.58/6 0.9} 22. Nxg7+ Sxg7 {-0.80/5 1.0} 23. Rf4 
address@hidden
{-1.14/5 1.8} 24. Rf5 Sf6 {-0.01/5 0.9} 25. Rf4 Sg7 {+0.00/6 1.1} 26. Gee1
Sf6 {-0.58/4 1.2} 27. address@hidden address@hidden {+2.05/5 2.2} 28. Ni3 Bh1+ 
{+2.68/6 1.4}
29. Nxh5 +Bxi1 {+2.69/4 3} 30. address@hidden address@hidden {+2.68/2 1.0} 31. 
Ri4 +Bh2
{-2.36/5 0.8} 32. fxf6 fxf6 {-2.35/4 0.9} 33. Bxf6 +Bh3 {-3.58/5 1.6} 34.
Re4 Li8 {-3.15/5 1.2} 35. Bh8+ +Bf5 {-2.97/5 1.3} 36. Nxi7+ Lxi7
{-2.20/5 0.7} 37. +Bxi7 Nf4 {-2.20/5 1.6} 38. Gf3 address@hidden {-1.96/3 0.6} 
39.
+Bh6 address@hidden {-7.47/3 2.8} 40. +Bxf8 +Bxe4 {-8.93/3 0.6} 41. 
address@hidden Kc8
{-14.49/7 2.5} 42. exe4 Nxc4 {-14.37/5 1.4} 43. Kc3 c6 {-12.01/4 1.1} 44.
address@hidden Kc7 {-11.94/4 1.1} 45. Gxf4 d6 {-13.42/3 1.3} 46. +Be9 Sd8
{-15.22/4 1.7} 47. Gxd8 c5 {-22.28/3 1.1} 48. Gc8 Kc6 {-1000.04/4 1.3} 49.
address@hidden Kb7 {-1000.03/8 1.0} 50. address@hidden Ka7 {-1000.03/8 1.3} 51. 
address@hidden Kxa6
{-1000.03/13 1.1} 52. Lxb8+ address@hidden {+1000.01/3 0.3}
{Xboard adjudication: Checkmate} 0-1




Op Zo, 18 februari, 2018 4:04 pm schreef address@hidden:
>

>
> ----- Mail original -----
>
>> It would probably be more revealing to try GNU Shogi 1.2 against an
>> engine known to be reasonably bug-free. This might be difficult, however,
>> as all available engines use XBoard protocol, and I suppose GNU Shogi
>> 1.2
>> is not yet patched for that, and is still stuck with xShogi protocol, for
>> which there are no opponents at all.
>
> For this my Omaha UI could be of use, as it's able to talk both
> protocols, but i've not worked on it for some time now, and it's still
> unable to handle clocks, which makes it of limited use at best.  Maybe I
> should have a look at this, it was what I last worked on on this software,
> I just
> remember it was not finished...
>
> The obvious other idea would be to backport your XBoard patches, but then
>  this work would be of limited use aside from checking the AI sanity, so
> I'm less inclined to go this path :}
>
>
>
>> Otherwise it would be a good idea to try it against CrazyWa (which is
>> virtually knowledgeless; it has for instance no idea what generals are,
>> and that it is good strategy to keep some of those close to your King). Or
>> even one of the configurable chess-variant engines, such as Sjaak II,
>> configured for Shogi (which also have zero specific Shogi knowledge). If
>> GNU Shogi 1.2. is totally
>> crushed by those (which would not surprise me), finding out exactly where
>> it regressed seems a waste of time.
>
> Sounds like a good plan.
>
>
>> @Yann:
>> So if you are interested and have time, you
>> could package the latest snapshot (version 1.0.5) from the repository at
>> http://hgm.nubati.net . I
>> added a Makefile with 'make install'
>> and 'make dist' targets. I did not add any Wa- or Tori-Shogi piece
>> graphics.
>
> Great news, I guess I can find a slot for such a small task :)
> Thanks!
>
>
>> Regards,
>> H.G
>>
>>
>> Op 2/11/2018 om 10:29 PM schreef address@hidden:
>>
>>> I resumed a bit of work on my 1.2-revival branch (now pushed to
>>> Savannah), in order
>>> to make it buildable with "gcc-6 -m32" (64bit was probably not the
>>> target at that time, and in fact the tests segfault on x86_64,
>>> although there are build options for alpha...).
>>>
>>> I launched a pair of games between that 1.2p03 (even without
>>> opening moves) and 1.5 snapshot, and the results are as interesting as
>>> I imagined they could be:
>>>
>>>
>>> * 1.2p03 wins all the games
>>> * 1.5 even lets its clock expire quite often
>>> * xshogi UI blocks when it gets "White mates" from engine so it
>>> takes manual work to extract the games
>>>
>>> Games recorded with:
>>> xshogi -debug True -fsp gnushogi -ssp ./gnushogix-1.2pl03 Output then
>>> processed with to get what looks like a decent PGN: grep "Received
>>> from first:" |grep -v time|cut -d: -f2|tail -n+3
>>>
>>> Sample games attached.  Sente is 1.5 (despite filename, sorrry),
>>> Gote is 1.2p03
>>>
>>>
>>> Game 1: sente gets invaded at move 10 already, and could give up
>>> around move 20 Game 2: sente gets seriously invaded only at move 25,
>>> and what follows does look like crap (did not take the time to analyse
>>> early moves anyway, my own level is so weak :) Game 3: sente invaded at
>>> move 15
>>>
>>> Looks like we have a pattern here :)
>>>
>>>
>>>
>>> In further tests:
>>>
>>>
>>> 1.3.2 vs 1.2p03 (no book on either side):
>>> 1. stength difference is not as drastic, but 1.3.2 still looses
>>> on time (game goes on, it would look like 1.3.2 gets slowly
>>> overwhelmed, see game) 2. 1.3.2 still looses on time, "but" gets ahead
>>> and mates - strange things happen, though, like the 37-38 moves
>>>
>>> 1.5pre vs 1.3.2 (no book for 1.3.2):
>>> 1. 1.3.2 wins after 75 moves (no clock expiration on either side)
>>> 2. 1.5 wins after 88 moves (xshogi says "black's flag has
>>> fallen", but I'm not sure what it's supposed to be, no clock is
>>> negative here)
>>>
>>>
>>> My conclusions so far:
>>> * there may still be some remaining hope in this engine
>>> * real benchmarking has to be done to identify when the various
>>> regressions occurred (controlled conditions, 32 vs 64bits, book or not)
>>>  * then we'll know where it stands more precisely
>>>
>>>
>>>
>>>
>>>
>>>
>>> ---
>>> This email has been checked for viruses by AVG.
>>> http://www.avg.com
>>>
>>
>>
>>
>

XS2WB.c
Description: Text Data