I wonder if chess is a draw with less ranks

[jkominek]

A couple adjunct points.

1) I re-partitioned my Chess324 games under Armageddon scoring, grouping together the castling rights of KQ/q, KQ/-, K/q, and K/-. Thus a single match between players consists of 4 x 648 = 2592 games. The total stats came out to 5067320 / 10238400 = 49.5% White wins, with ordo reporting a White advantage = -11.15 +/- 0.21 Elo. This makes for a strikingly balanced competition in computer play, with the benefit that the engines are entirely responsible for all piece development. By construction, no more draw death!

2) I did not elaborate on my spicy comment that Komodo Dragon's Monte Carlo evaluations are "suspicious crap that I can't take seriously." I'll illustrate what caught me by surprise. From a 1 million node search:

Code: Select all

position fen rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1
setoption name Use MCTS value true
setoption name MCTS Hash value 32768
info string MCTS Hash table size is now 32768 meg
setoption name MultiPV value 6
...
info multipv 1 depth 29 seldepth 69 time 2797416 nodes 1039051 nps 371 score cp 108 tbhits 0 pv d2d4 d7d5 c2c4 c7c6 g1f3 e7e6 b1c3 g8f6 e2e3 f6e4 f3d2 f7f5 d2e4 f5e4 f2f3 e4f3 d1f3 d8f6 e3e4 f6f3 g2f3 f8e7 h2h4 h8f8 c4d5 e6d5 e4d5 b8d7 c1g5 d7f6 e1d2 e8d8 a1e1 a7a6 d5c6 b7c6 d2c1 h7h6 g5f4 f6d5 c3d5 c6d5 f4c7 d8d7 c7e5 g7g5 f1g2 d7d8 e5c7 d8d7 h4g5 e7g5 f3f4 d7c7
info multipv 2 depth 29 seldepth 69 time 2797416 nodes 1039051 nps 371 score cp 81 tbhits 0 pv g1f3 d7d5 d2d4 c8f5 c2c4 e7e6 d1b3 b8c6 c1d2 a8b8 e2e3 a7a6 b1c3 g8f6 a1c1 f8b4 f1e2 h7h5 e1g1 e8f8 a2a3 b4c3 d2c3 f5g4 h2h3 f6e4 g1h1 h8h6 c3e1
info multipv 3 depth 29 seldepth 69 time 2797416 nodes 1039051 nps 371 score cp 63 tbhits 0 pv c2c4 e7e5 g2g3 g8f6 f1g2 f8c5
info multipv 4 depth 29 seldepth 69 time 2797416 nodes 1039051 nps 371 score cp 57 tbhits 0 pv e2e4 c7c5 g1f3 b8c6 f1b5 e7e6
info multipv 5 depth 29 seldepth 69 time 2797416 nodes 1039051 nps 371 score cp 58 tbhits 0 pv e2e3 c7c6 g1f3 d7d5 c2c4 c8g4
info multipv 6 depth 29 seldepth 69 time 2797416 nodes 1039051 nps 371 score cp 46 tbhits 0 pv g2g3 c7c5 g1f3 b8c6
stop
info nodes 1039545
info depth 29 seldepth 69 time 2799028 nodes 1039545 nps 371 score cp 181 tbhits 0 hashfull 863 pv d2d4 d7d5 c2c4 c7c6 g1f3 e7e6 b1c3 g8f6 e2e3 f6e4 f3d2 f7f5 d2e4 f5e4 f2f3 e4f3 d1f3 d8f6 e3e4 f6f3 g2f3 f8e7 h2h4 h8f8 c4d5 e6d5 e4d5 b8d7 c1g5 d7f6 e1d2 e8d8 a1e1 a7a6 d5c6 b7c6 d2c1 h7h6 g5f4 f6d5 c3d5 c6d5 f4c7 d8d7 c7e5 g7g5 f1g2 d7d8 e5c7 d8d7 h4g5 e7g5 f3f4 d7c7
bestmove d2d4 ponder d7d5
quit

At this point in building its search tree Komodo Dragon 3.3 reported 1.d4 as having an evaluation of 108, which is consistent with an alpha-beta search run for the same amount of time. Yet when I stopped the search Komodo's final evaluation is given as 181. Why the jump? To me this does not make mathematical sense. Note that the PV is unchanged.

[lkaufman]

If Black cannot castle gives evals around 1.2 it is clearly unsuitable for Armageddon, a 75 to 25 expected score is totally unacceptable. That's why I proposed Black cannot castle short, that should be much closer to 1.0. Even if your fast stats suggest it's not as good a fit as no Black castling, I would trust the eval more.

The WDL model of Stockfish, Leela, and others including I presume Komodo Dragon, are based on computer-computer self-play measurements at bullet time controls. It is a guide but not one we should over-reach.

The question in my mind, and one I am interested in your opinion on, is How do you know? By that I mean how do you know that computer results translate directly to equivalent human results without further empirical studies? Absent abundant evidence from human play I treat the measurements I've made (and yours) as suggestive but not definitive.

What stands out to me as the confounding variable is the ability of human grandmasters to semi-reliably secure a draw when needed, in this case by Black. We also know that computer ratings lists are dilated with respect to FIDE ratings, e.g. as expressed in Komodo Dragon's UCI Elo settings, a fact I attribute to human's greater tendency to play for draws, especially when the players have a large rating separation. (Versus engines fighting it out for the win for many more moves.)

Breaking it down,
i) Under Armageddon rules the ideal opening position score 50% between equally rated players.
ii) Modern engines are calibrated such that an evaluation of 100 centipawns equates to a 50% win/draw ratio - in computer-computer play. This puts us in the ballpark but should not be considered definitive.
iii) Observation: human grandmasters have the ability to successfully play for a draw when needed. Also, placing White in a must-win situation adds clutch-play pressure.
iv) Together this tilts the advantage further in Black's favor.
v) To compensate the opening position needs to more strongly favor White that what computer WDL models consider balanced.
vi) This argues for a value somewhat above 100 centipawns being the target point. A value of 120 may seem too high but should not be ruled out as suitable in human play.
vii) Likewise I would not rule out a position with an evaluation of 80, either. Because maybe I'm wrong about statement iv).

In short I advocate for a broader range of consideration (appreciation of uncertainty) until thoroughly put to the test.

I'm limited by not having a database of human Armageddon games at my disposal. Do you have a collection of games to study from your connections to either chess.com or Lichess? Maybe you have analysis that you could share. Even if only for the standard opening position it would provide a useful point of comparison.

More ideal of course a large experiment on either of those playing sites to study variant castling rights.

I wouldn't put any weight on the idea that human grandmasters have the ability to play for a draw when needed. Sure, they have memorized many defenses for Black like the Berlin with the goal to reach a drawable ending, but this is irrelevant for play with alternate rules such as limited castling rights, which throws all opening theory out the window. I don't know whether Armageddon play with such rules is more favorable for one color or the other with human players than with equal strength bots, but I suspect there is very little difference. Also, I am interested in Armageddon rules that are as fair as possible both for human grandmasters and for engines (which means effectively for human correspondence play); if there is some difference in the eval margin needed I would advocate an intermediate value. But as I recall the difference between the eval needed for 2800 engines and for 3600 engines is very small. It is probably roughly as difficult to find "only" winning moves and "only" drawing moves, highly position-dependent.

I don't have an Armageddon database. For computer games with various castlin rights and Armageddon rules, Stephan Pohl's website should be the place to look, he did a lot of work on that some years ago.

[tmokonen]

It's pretty easy to set up these sorts of chess variants in Fairy Stockfish. For example, I added this to my variants.ini file to implement 5 and 6 rank chess:

Code: Select all

[5rankchess]
pieceToCharTable = PNBRQ................Kpnbrq................k
pawn = p
knight = n
bishop = b
rook = r
queen = q
king = k
maxRank = 5
maxFile = 8
startFen = rnbqkbnr/pppppppp/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
promotionRank = 5
promotionPieceTypes = qrbn
doubleStep = false
castling = true

[6rankchess]
pieceToCharTable = PNBRQ................Kpnbrq................k
pawn = p
knight = n
bishop = b
rook = r
queen = q
king = k
maxRank = 6
maxFile = 8
startFen = rnbqkbnr/pppppppp/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
promotionRank = 6
promotionPieceTypes = qrbn
doubleStep = true
castling = true

You can then upload your modified variants.ini to https://fairyground.vercel.app/ and use that interface to test your ideas. Of course, running Fairy Stockfish locally is faster than through a web interface, but it's nice to have that web GUI available.

[tmokonen]

The two variants I specified above can be defined even more simply in the .ini file:

Code: Select all

[5rankchess:chess]
maxRank = 5
maxFile = 8
startFen = rnbqkbnr/pppppppp/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
promotionRank = 5
doubleStep = false

[6rankchess:chess]
maxRank = 6
maxFile = 8
startFen = rnbqkbnr/pppppppp/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
promotionRank = 6

[Ajedrecista]

Hello Tony:

Great contribution! I wrote about Los Alamos chess earlier in this thread because is historically significant for computer chess. The .ini file for that variant should be:

Code: Select all

[LosAlamoschess:chess]
pieceToCharTable = PNRQ................Kpnrq................k
pawn = p
knight = n
rook = r
queen = q
king = k
maxRank = 6
maxFile = 6
startFen = rnqknr/pppppp/6/6/PPPPPP/RNQKNR w - - 0 1
promotionRank = 6
promotionPieceTypes = qrn
doubleStep = false
castling = false

However, I am not able to run the variant at https://fairyground.vercel.app/ site once I load the .ini file. I did not found any help here. I do not know if 6 files requires less dots at 'pieceToCharTable' string: there are 16 and 16 in both 5×8 and 6×8 variants, so maybe 12 and 12 in 6×6? I edited the .ini and reloaded, without success again.

I also do not know how to run locally rather than through the web interface. Should I download Fairy Stockfish from GitHub, install the engine and Arena GUI would change the typical 8×8 chessboard to 6×6 with the correct .ini file (and get back to 8×8 when needed)? variants.ini at GitHub does not provide settings for Los Alamos variant.

There is other link from the help that I managed to run an analysis of a preloaded Los Alamos .ini file, not my own:

https://fairy-stockfish-nnue-wasm.verce ... lysis.html

Threads = 1 (for being deterministic), hash = 1024 MB and MultiPV = 10 (all possible first moves):

Code: Select all

info depth 20 seldepth 31 multipv 1 score cp 43 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv c2c3 b5b4 f2f3 e6f4 e2e3 c6b5 c3c4 b6c4 e3f4 e5f4 c1c2 a5a4 f1f2 a4a3 b2a3 c4a3 b1a3 a6a3 a1b1 f6e6 b1b3 a3a6 e1d3
info depth 20 seldepth 27 multipv 2 score cp 15 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv f2f3 d5d4 c2c3 d4c3 d2c3 c6d5 b1d2 b6c4 e1c2 e6f4 b2b3 d5d2 c1d2 c4d2 d1d2 c5c4 e2e3 f4d5 e3e4 f5e4 f3e4 d5f4 b3b4 d6c6 a1d1 a5b4 c2b4 c6c5 b4a6 f6a6
info depth 20 seldepth 29 multipv 3 score cp 0 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv b2b3 a5a4 c1b2 b5b4 c2c3 c6b5 e1c2 a4a3 b2c1 e5e4 c3b4 c5b4 f2f3 e6f4 c2d4 b5c5 e2e3 f4d3 c1c5 d6c5 d1e2
info depth 20 seldepth 29 multipv 4 score cp 0 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv d2d3 c5c4 c2c3 c4d3 e1d3 e5e4 d3e1 d5d4 c3d4 c6c1 d1c1 e6d4 e2e3 d4e6 f2f3 b6d5 f3e4 f5e4 f1f6 d5f6 e1c2 a6c6 b1c3 b5b4 c3b5 d6c5 c2d4 e6d4 b5d4 c6d6
info depth 20 seldepth 27 multipv 5 score cp -76 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv e1f3 e5e4 f3e1 d5d4 b2b3 b5b4 a2a3 b6d5 f2f3 e6f4 f3e4 f5e4 e2e3 f4e6 f1f6 d5f6 a3b4 a5b4 a1a6 c6a6 c2c3 b4c3 b1a3 a6f1
info depth 20 seldepth 31 multipv 6 score cp -123 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv e2e3 f5f4 f2f3 b5b4 a2a3 c6b5 f1f2 c5c4 b2b3 e6c5 a3a4 b5c6 c1b2 d6e6 f2e2 f4e3 d2e3 e6f5 f3f4 c4c3
info depth 20 seldepth 29 multipv 7 score cp -123 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv b1a3 f5f4 a3b1 c5c4 d2d3 b5b4 e1f3 e5e4 d3e4 d5e4 f3d2 d6e5 a2a3 c4c3 a3b4 c3d2 b1d2 c6b5 c2c3 a5b4 a1a6 b5a6 c3b4 a6a4 b2b3
info depth 20 seldepth 32 multipv 8 score cp -129 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv a2a3 c5c4 d2d3 c4d3 e1d3 e5e4 d3f4 c6c5 c2c3 a5a4 f4e6 d6e6 c1d2 b6c4 d2d4 a6c6 d4c5 c6c5 d1c1 e6e5 b1d2 d5d4 c3d4 e5d4 d2c4 d4c4 c1c2 c4d4 c2b1 f6c6 b1a2 c5c2
info depth 20 seldepth 32 multipv 9 score cp -188 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv b1c3 d5d4 c3b1 c5c4 c2c3 d4d3 e2d3 c4d3 f2f3 e6f4 b2b3 b6c4 b1a3 d6d5 a1b1 c4a3 c1a3 f6d6 c3c4 d5e6 b1c1 b5b4 a3b2 c6c5 a2a3 a6c6
info depth 20 seldepth 31 multipv 10 score cp -197 nodes 27882225 nps 140326 hashfull 245 tbhits 0 time 198695 pv e1d3 c5c4 d3e1 d5d4 c2c3 d4d3 e2d3 c4d3 f2f3 e6f4 b1a3 b5b4 b2b3 c6c5 c3b4 a5b4 c1c5 d6c5 a1c1 c5d5 e1c2 d3c2 a3c2 f4d3

With less verbosity:

Code: Select all

Depth 20:

+0.43  c3 (d20/31)  P Q  - 3
+0.15  f3 (d20/27)  P KR - 3
 0.00  b3 (d20/29)  P QN - 3
 0.00  d3 (d20/29)  P K  - 3
-0.76 Nf3 (d20/27) Kt KR - 3
-1.23  e3 (d20/31)  P KN - 3
-1.23 Na3 (d20/29) Kt QR - 3
-1.29  a3 (d20/32)  P QR - 3
-1.88 Nc3 (d20/32) Kt Q  - 3
-1.97 Nd3 (d20/31) Kt K  - 3

Moving a knight first seems bad for white. More graphically, regarding pawn moves:

Code: Select all

-1.29    0.00   +0.43    0.00   -1.23   +0.15
  P       P       P       P       P       P
  R       N       Q       K       N       R
=============================================
  a       b       c       d       e       f

Regarding knight moves:

Code: Select all

-1.23           -1.88   -1.97           -0.76
  P       P       P       P       P       P
  R       N       Q       K       N       R
=============================================
  a       b       c       d       e       f

------------

Meanwhile, I found that the variant was preloaded together with many other variants at the web interface you pointed to. I was able to win against a random mover: 1.- d3, Nd4; 2.- c3, Nb3?; 3.- axb3, d4; 4.- cxd4, b4; 5.- Qxc5+, Ke6??; 6.- Qxc8# (1-0). I can win against a random mover!

Maggie's IQ Test | The Simpsons

Well, congratulations. You are now as intelligent as a pig. Let's see if you can move up to a dolphin.

(From The Simpsons S15E13 'Smart & Smarter', first aired on February 22, 2004).

Regards from Spain.

Ajedrecista.