Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Texel eval tuning and Tapered eval #61

Merged
merged 15 commits into from May 13, 2021
Merged

Texel eval tuning and Tapered eval #61

merged 15 commits into from May 13, 2021

Conversation

amanjpro
Copy link
Owner

@amanjpro amanjpro commented May 7, 2021

No description provided.

@amanjpro
Copy link
Owner Author

amanjpro commented May 8, 2021

So far, as commit 8c30f4c

Untuned Tapered eval:

vs Zahak 2.0.0,

Finished game 200 (zahak-darwin-amd64-latest vs zahak_next): 1/2-1/2 {Draw by 3-fold repetition}
Score of zahak_next vs zahak-darwin-amd64-latest: 81 - 54 - 65  [0.568] 200
...      zahak_next playing White: 40 - 28 - 32  [0.560] 100
...      zahak_next playing Black: 41 - 26 - 33  [0.575] 100
...      White vs Black: 66 - 69 - 65  [0.492] 200
Elo difference: 47.2 +/- 39.9, LOS: 99.0 %, DrawRatio: 32.5 %
Finished match

vs
Achilliees, Gopher-Check, Baisclicka:

Finished game 599 (zahak_next vs Achillees): 1-0 {White wins by adjudication}
Rank Name                          Elo     +/-   Games    Wins  Losses   Draws   Points   WWins  WLoss.  WDraws   BWins  BLoss.  BDraws
   0 zahak_next                    -22      25     600     224     262     114    281.0     128     112      60      96     150      54
   1 baislicka                      65      44     200     100      63      37    118.5      55      28      17      45      35      20
   2 Achillees                      21      44     200      88      76      36    106.0      56      32      12      32      44      24
   3 gopher_check                  -19      43     200      74      85      41     94.5      39      36      25      35      49      16

Finished match

vs Clever Girl (clever girl crashed):

Score of zahak_next vs clever-girl: 82 - 63 - 55  [0.547] 200
...      zahak_next playing White: 42 - 31 - 27  [0.555] 100
...      zahak_next playing Black: 40 - 32 - 28  [0.540] 100
...      White vs Black: 74 - 71 - 55  [0.507] 200
Elo difference: 33.1 +/- 41.3, LOS: 94.3 %, DrawRatio: 27.5 %
Finished match

@amanjpro
Copy link
Owner Author

With the latest commit 2f69a25, this is my current tests:

Tuned Tapered eval:

vs Zahak 2.0.0,

Score of zahak_next vs zahak-darwin-amd64-latest: 112 - 38 - 50  [0.685] 200
...      zahak_next playing White: 64 - 15 - 21  [0.745] 100
...      zahak_next playing Black: 48 - 23 - 29  [0.625] 100
...      White vs Black: 87 - 63 - 50  [0.560] 200
Elo difference: 135.0 +/- 43.9, LOS: 100.0 %, DrawRatio: 25.0 %
Finished match

vs
Achilliees, Gopher-Check, Baisclicka:

Rank Name                          Elo     +/-   Games    Wins  Losses   Draws   Points   WWins  WLoss.  WDraws   BWins  BLoss.  BDraws
   0 zahak_next                    137      26     600     347     122     131    412.5     180      64      56     167      58      75
   1 gopher_check                 -111      44     200      46     108      46     69.0      23      54      23      23      54      23
   2 Achillees                    -145      46     200      39     118      43     60.5      19      58      23      20      60      20
   3 baislicka                    -156      46     200      37     121      42     58.0      16      55      29      21      66      13

Finished match

vs Clever Girl (clever girl crashed):

Score of zahak_next vs clever-girl: 137 - 26 - 37  [0.777] 200
...      zahak_next playing White: 71 - 12 - 17  [0.795] 100
...      zahak_next playing Black: 66 - 14 - 20  [0.760] 100
...      White vs Black: 85 - 78 - 37  [0.517] 200
Elo difference: 217.3 +/- 50.1, LOS: 100.0 %, DrawRatio: 18.5 %
Finished match

Vice, Rustic

With Bayeselo

ResultSet>readpgn /Users/amanj/Documents/MyDocuments/Programming/zahak/src/github.com/zahak/zahak_games/tuned-evals-3.pgn
1000 game(s) loaded
ResultSet>elo
ResultSet-EloRating>mm
00:00:00,00
ResultSet-EloRating>exactdist
00:00:00,00
ResultSet-EloRating>ratings
Rank Name                       Rating   Δ     +    -     #     Σ    Σ%     W    L    D   W%    =%   OppR
---------------------------------------------------------------------------------------------------------
   1 zahak_next                  3227   0.0   17   17  1000  705.0  70.5  596  186  218  59.6  21.8  3075
   2 gopher_check                3115 111.9   35   36   200   69.0  34.5   46  108   46  23.0  23.0  3227
   3 zahak-darwin-amd64-latest   3095  20.4   35   36   200   63.0  31.5   38  112   50  19.0  25.0  3227
   4 Achillees                   3081  13.3   36   37   200   60.5  30.2   39  118   43  19.5  21.5  3227
   5 baislicka                   3070  11.6   36   38   200   58.0  29.0   37  121   42  18.5  21.0  3227
   6 clever-girl                 3013  56.8   38   41   200   44.5  22.2   26  137   37  13.0  18.5  3227
---------------------------------------------------------------------------------------------------------
  Δ = delta from the next higher rated opponent
  # = number of games played
  Σ = total score, 1 point for win, 1/2 point for draw

That shows an elo for Zahak around 2270-2340, which is at least 130 elo points ahead of whatever Zahak has now

@amanjpro amanjpro merged commit 8d17f9b into master May 13, 2021
@amanjpro amanjpro deleted the texel-eval-tuning branch September 21, 2021 19:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

1 participant