Author Topic: Gothic Chess Correspondence Game  (Read 182 times)

GothicChessInventor

  • Full Member
  • ***
  • Posts: 175
    • View Profile
Gothic Chess Correspondence Game
« on: January 29, 2018, 04:21:23 pm »
I hereby declare that Panzerschiff can beat anyone on this site in a correspondence game of Gothic Chess played at the rate of 1 move per 7 days (faster on occasion and as the players deem fit).

:)

Share on Facebook Share on Twitter


chilipepper

  • Full Member
  • ***
  • Posts: 100
  • chilipepper
  • Location: U.S.
    • View Profile
    • QyiQ23607 Pisano
Re: Gothic Chess Correspondence Game
« Reply #1 on: January 29, 2018, 05:41:11 pm »
Who is Panzerschiff? Is it another screenname for you, a friend, a computer program, or a secret? ???
the "chilipepper"👹

GothicChessInventor

  • Full Member
  • ***
  • Posts: 175
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #2 on: January 29, 2018, 09:37:24 pm »
Who is Panzerschiff? Is it another screenname for you, a friend, a computer program, or a secret? ???

Panzerschiff is my friend from Ohio. I think his correspondence chess rating was 2200+ in the early 1980s.

By the way, programs are horrible at correspondence time controls for Gothic Chess, and they are the easiest opponents to defeat. They just can't plan ahead.
« Last Edit: January 29, 2018, 09:39:51 pm by GothicChessInventor »

chilipepper

  • Full Member
  • ***
  • Posts: 100
  • chilipepper
  • Location: U.S.
    • View Profile
    • QyiQ23607 Pisano
Re: Gothic Chess Correspondence Game
« Reply #3 on: January 30, 2018, 11:14:28 am »
Oh, Thanks. Any idea which of the two programs (Fairy Max, and ChessV) is stronger at Gothic Chess? I've tried both programs (with other games) but have never played them against each other in one particular game. Regardless of which one is stronger - I think they're both well designed, and like them a lot. :)
the "chilipepper"👹

ebinola

  • Jr. Member
  • **
  • Posts: 77
  • Location: nowhere in particular
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #4 on: January 30, 2018, 05:49:29 pm »
I'll take a crack at it. Not to say that I'll win but I'd like to play nonetheless.
Chess player, amateur writer and on the side shitposter extraordinaire.

GothicChessInventor

  • Full Member
  • ***
  • Posts: 175
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #5 on: January 30, 2018, 08:51:57 pm »
I'll take a crack at it. Not to say that I'll win but I'd like to play nonetheless.

OK I'll play 1. d4 if you want to play against me. It was not fair for me to volunteer Panzer :)

GothicChessInventor

  • Full Member
  • ***
  • Posts: 175
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #6 on: January 30, 2018, 08:53:53 pm »
Oh, Thanks. Any idea which of the two programs (Fairy Max, and ChessV) is stronger at Gothic Chess? I've tried both programs (with other games) but have never played them against each other in one particular game. Regardless of which one is stronger - I think they're both well designed, and like them a lot. :)

I made a new topic out of this question, but here is the short answer for the list of programs from that played in the 2007 Gothic Chess Computer World Championship:

PlaceProgramScore
1-2Gothic Vortex13.0-1.0
1-2Variant Pulverizer13.0-1.0
3Tornado9.0-5.0
4SMIRF8.0-6.0
5TSCP Gothic 64-bit6.0-8.0
6fmax44.5-9.5
7ChessV2.0-10.0
8Zillions 20.5-13.5

Gothic Vortex and Variant Pulverizer each won their games as white against one another. They were the dominant programs. Sadly, Vortex does not run under Windows 10, so I need to rewrite it.
« Last Edit: January 30, 2018, 08:56:43 pm by GothicChessInventor »

Greg Strong

  • Newbie
  • *
  • Posts: 21
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #7 on: January 30, 2018, 11:25:01 pm »
Oh, Thanks. Any idea which of the two programs (Fairy Max, and ChessV) is stronger at Gothic Chess? I've tried both programs (with other games) but have never played them against each other in one particular game. Regardless of which one is stronger - I think they're both well designed, and like them a lot. :)

The newest ChessV beats Fairy-Max pretty regularly.  If you download the Windows installer, it includes Fairy-Max, so you can use ChessV to pay them against each other with no configuration needed.

GothicChessInventor

  • Full Member
  • ***
  • Posts: 175
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #8 on: January 31, 2018, 04:24:30 am »
The newest ChessV beats Fairy-Max pretty regularly.  If you download the Windows installer, it includes Fairy-Max, so you can use ChessV to pay them against each other with no configuration needed.

Chess V is very easy to crush. I gave it 3 minutes per move tonight on my Intel i7-5960X overclocked to 4.6 GHz and I was able to beat it by giving it tons of material in positions it obviously did not understand.

1. d4 d5
2. Nh3 Nc6
3. i3

The first step to defeating a program: play positionally and avoid tactics. This is a slow, developing move that programs can't decipher. It looks to "weaken" the kingside from the perspective of most evaluation functions, but it is simply preparing to castle, which is always a good idea in Gothic Chess.

3...g6
4. c3 h5?

Black is self-inflicting wounds already.

5. Bi2 e5
6. dxe5 Nxe5
7. f4 Ng4

A move with the aim of forcing the white archbishop to "babysit" the h2 square, otherwise the fork ...Nxh2+ picks up the rook on j1. This motif is ubiquitous in Gothic Chess, and many games feature such play. However, here, the program is missing the most obvious way to dodge the veiled threat made by this knight.

8. Ac5+ Ae7

Of course, trading archbishops by using the tempo of check to castle out of danger.

9. Axe7+ Cxe7
10. O-O Qd6?

The game is functionally over already from a strategic point of view. All this move does is allow a trivial skewer threat along the a3-f8 diagonal, which is the diagonal on which the game ends very shortly.

11. b3 Bf5

Ignoring the threat of Ba3 with the mistaken belief that blocking with the pawn move to c5 is "safe enough." Even though the program is completing 12- and 13-ply searches, the danger associated with the bishop on a3 is actually permanent, and not temporary. Programs have an impossible task if they wish to try and label such positional stratagems.

12. e4 Bxe4

I decide to donate a pawn to start "tilting" the alpha-beta window. As the program is now a "pawn ahead," literally every move it searches will be good for it and any lurking danger will be beyond its search horizon. This is the perfect square to lure the bishop onto, since there is a semi-pin on it considering my chancellor is gunning down the e-file and with Ng5 I compound the attack on it on e4.

13. Ba3

Making this move now since BxN on b1 would deny the safety of deploying it.

13...c5
14. Ng5 Nf2

Allowing black to fork my queen and rook with its forlorn knight now on f2, because my rook is actually quite safe. I will further illustrate this in upcoming moves by planting the rook directly in harm's way where the knight can capture it.

15. Qe2!

The game is definitely over now. Get out your scorecards and start counting how many hanging pieces white allows. Currently I am down one pawn and allowing the knight to take my rook.

 15...Nd3

And the knight declines winning The Exchange, wisely. I thought with the extra pawn it was ahead, it would grab it and allow Nh7+ inflicting more than collateral damage. This is actually a good move for black.

16. Cd1

White really doesn't have a choice here. One thing I have noticed in all of my years of play, if one side plays Cc2 (or Cc7 as black) that almost never works out well. Chancellors need to be treated as rooks until the late middlegame, where they become more deadly.

16...f5

This move basically turns the black bishop on e4 into a pawn. It has no safe moves anymore. It does stop Nh7+ by virtue of allowing the chancellor to cover that square, free from any pawn obstructions on the 7th rank now.

17. Nd2 Nh6
18. Re1!!

A move which wins in all variations, as is often said. Black has closed the position and white needs to pry it open in order to exploit the defects in the enemy camp.

 18...Nxe1
19. Ndxe4!!

And white does the unthinkable, capturing the "bad bishop" while letting the knight that took the rook remain unscathed. The knight on e1 is dead meat anyway, and by keeping it around the program is burdened with move generation for it to no avail. The white knight on e4 is impervious to the pawn threats to capture it. If 19...fxe4? 20. Ne6+!! and if 19...dxe4?? 20. Cxd6 takes the queen. In the game, white is now down a Rook and a pawn for a mere bishop, and one knight is en prise. These are the types of positions that confuse programs.


19...Qc7
20. Nh7+!!

And white sacrifices another knight!

20... Cxh7


21. Cxd5!!

Leaving the black knight on e1, and only capturing the "mission critical" d5 pawn for it while leaving the knight on e4 hang, still.

21...Kg8??

Going from the frying pan to the fire.

22. Qc4!

Again, setting up play with the concept of a deadly pin rather than try and recover material or save the hanging knight. None of these moves would ever make it "within the bounds" of an alpha-beta search window. Yet they are deadly.

22...Nxg2+

A spite check if there ever was one. The materialistic computer snags another pawn. The move is about as worthless as they come.

23. Kh1 Qf7
24. Ng5 Nxf4

Look at all of the threatened pieces in this position.

25. Ce7+!!

I decide to let the black king take my chancellor. Not now, but in 2 moves. The program still doesn't understand. How can it be so far ahead in material and still be losing?
 

25...Kf8
26. Nxh7+ Kxe7

27. Qxc5+

And now that a3-f8 diagonal is fatal.

27... Ke8
28. Re1+ Ne6
29. Qd6 Be5

The program must start tossing pieces to stop the checkmate. It's a cake walk from here to the end.

30. Rxe5 a5
31. Rxe6+ Qxe6
32. Qxe6+ Kd8
33. Qe7+ Kc8
34. Bd6 Ng8
35. Qc7#

So don't put too much faith in programs, no matter what statistics their programmers throw at you. They are easy to defeat with creative play that takes them out of what they consider "the principal variation," if you can identify the key segments of the board that need occupation.
« Last Edit: January 31, 2018, 04:46:36 am by GothicChessInventor »

HGMuller

  • Global Moderator
  • Full Member
  • *****
  • Posts: 120
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #9 on: January 31, 2018, 09:07:35 am »
I am not sure what exactly Gregory means by 'beats it regularly'. Does that imply more than half of the time?

Note that Fairy-Max is what I consider a weak program. It is derived from micro-Max by just changing the possible board sizes and piece moves, and micro-Max was optimized for being small (maximum Elo per character) rather than for being strong. As a result virtually all Chess knowledge was culled out of it, and the search algorithms are also pretty basic.

For orthodox Chess, where people keep rating lists, mico-Max' rating is aroud 2000. For comparison, current top programs are over 3000. People that build a new egine from scratch, and put a by-now standard amount of minimal knowledge in it (like about Pawn structure, i.e. passers, doubled and isolated Pawns, the King's Pawn shield) typically are 2200 Elo as soon as it is ready to play, and rise to 2400 when the worst bugs are ironed out. Joker, the program from which Joker80 for 10x8 Chess is derived, is rated around 2400. And Bihasa is much stronger in Gothic Chess than Joker80.

Furthermore, there seems to be a strategic aspect to 10x8 Chess that is not very important in orthodox Chess, and makes it relatively easy for Humas that know the trick to beat most programs. So using the same algoritm as an orthodox engine, just enlarging the board size, does not automatically preserve the strength.

chilipepper

  • Full Member
  • ***
  • Posts: 100
  • chilipepper
  • Location: U.S.
    • View Profile
    • QyiQ23607 Pisano
Re: Gothic Chess Correspondence Game
« Reply #10 on: January 31, 2018, 11:09:05 am »
So don't put too much faith in programs...

I still think competitions between engines is pretty cool. Even if a human can beat a software program, a program that can win over all other programs is a pretty cool invention! 8)
the "chilipepper"👹

Greg Strong

  • Newbie
  • *
  • Posts: 21
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #11 on: January 31, 2018, 11:48:05 am »
Chess V is very easy to crush. I gave it 3 minutes per move tonight on my Intel i7-5960X overclocked to 4.6 GHz and I was able to beat it by giving it tons of material in positions it obviously did not understand.

Thanks, Ed.  When I get a little time I'll go over this and see what I can learn ...

Greg Strong

  • Newbie
  • *
  • Posts: 21
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #12 on: January 31, 2018, 12:11:56 pm »
I am not sure what exactly Gregory means by 'beats it regularly'. Does that imply more than half of the time?

Yes, sorry, that's what I meant.  I was surprised to be able to reach this point since Fairy-Max sees about three times as many nodes per second and ChessV has a pretty basic evaluation also.  The difference must be primarily in the search algorithm.  Chessv has most of the advanced stuff: PVS, IID, TT, Null move, LMR, razoring, futility, and delta pruning.  It is missing ProbCut and singular extension - not sure how much those are worth.  And, of course, it's not multi-threaded, although I do plan to do that.  The Stockfish 'Lazy SMP' approach looks pretty easy.

Although, chilipepper, if you are looking for strong 10x8 engines, also check out Joker80, another of H.G.'s engines.  Although I haven't tested it recently, it's probably still stronger than ChessV.  To use it with chessv is easy.  Just make a new "Joker80" sub-folder under chessv's Engines\XBoard folder, and throw the joker80.exe in it (download here.) Next time you fire up chessv, it will notice the new engine, fire it up, interrogate it about what variants and features it supports, and will make it available as an option for any supported games.  Or if you prefer to store the exe somewhere else, you can use Engines > New > Discover Engine.


HGMuller

  • Global Moderator
  • Full Member
  • *****
  • Posts: 120
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #13 on: January 31, 2018, 12:45:12 pm »
Ah, so you made good progress. Fairy-Max of course has stayed the same forall these years (in the variants it already supported), so it is a good opponent to guage that. BTW, Fairy-Max also has IID, TT, null move, LMR, delta pruning. I guess the most limiting factor for it is poor move ordering (hash move, and then the rest of the move in move-genration order).

ebinola

  • Jr. Member
  • **
  • Posts: 77
  • Location: nowhere in particular
    • View Profile
Re: Gothic Chess Correspondence Game
« Reply #14 on: January 31, 2018, 02:40:40 pm »
I'll take a crack at it. Not to say that I'll win but I'd like to play nonetheless.

OK I'll play 1. d4 if you want to play against me. It was not fair for me to volunteer Panzer :)
Are you wanting to do it in this thread?
If so, 1... h6
Chess player, amateur writer and on the side shitposter extraordinaire.