| 18-08-12 |
SC |
simpleVoting |
diff
|
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 39006 W: 8281 L: 8520 D: 22205
|
sprt
@ 5+0.05 th 4 |
Tweak weights. |
| 18-08-12 |
Roc |
PawnSupportB |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 21890 W: 4736 L: 4760 D: 12394
|
sprt
@ 10+0.1 th 1 |
PSB_1
Take 2 |
| 18-08-12 |
sni |
pr_1725~4 |
diff
|
LLR: 0.14 (-2.94,2.94) [0.00,5.00]
Total: 77768 W: 17418 L: 17026 D: 43324
|
sprt
@ 10+0.1 th 1 |
Non-linear pawn count bonus in imbalance |
| 18-08-12 |
sni |
pr_1725~1 |
diff
|
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 19683 W: 3409 L: 3206 D: 13068
|
sprt
@ 60+0.6 th 1 |
LTC: Take 4, tweak bishop values like in PR #1733 |
| 18-08-12 |
Roc |
PawnSupport |
diff
|
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 5438 W: 1139 L: 1244 D: 3055
|
sprt
@ 10+0.1 th 1 |
PS_1
Some bonus for squares which would be left over by potential storming pawns |
| 18-08-12 |
Roc |
MinorSafeCheck |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16472 W: 3594 L: 3644 D: 9234
|
sprt
@ 10+0.1 th 1 |
MSC
Consider a Minor check safe if supported by another piece but only defended by a major (the queen case was already in master this test adds the rook case) |
| 18-08-12 |
SC |
simpleVoting |
diff
|
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 13657 W: 2860 L: 3050 D: 7747
|
sprt
@ 5+0.05 th 4 |
Tweak weights, take 2. |
| 18-08-12 |
SC |
simpleVoting |
diff
|
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 10054 W: 2072 L: 2255 D: 5727
|
sprt
@ 5+0.05 th 4 |
Take negative scores into account. |
| 18-08-12 |
sni |
pr_1725~1 |
diff
|
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 42948 W: 9662 L: 9317 D: 23969
|
sprt
@ 10+0.1 th 1 |
Take 4: tweak bishop values like in PR #1733 |
| 18-08-12 |
SC |
PVextension |
diff
|
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 12391 W: 2667 L: 2738 D: 6986
|
sprt
@ 10+0.1 th 1 |
3000000, 2. |
| 18-08-12 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12961 W: 2888 L: 2955 D: 7118
|
sprt
@ 10+0.1 th 1 |
3000000, 4. |
| 18-08-12 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5938 W: 1257 L: 1359 D: 3322
|
sprt
@ 10+0.1 th 1 |
2500000, 3. |
| 18-08-12 |
sni |
pr_1725 |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 17599 W: 3819 L: 3914 D: 9866
|
sprt
@ 10+0.1 th 1 |
Take 5: without the pawn non-linearity |
| 18-08-12 |
sni |
pr_1725~3 |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 16986 W: 3753 L: 3850 D: 9383
|
sprt
@ 10+0.1 th 1 |
Take 2: only the pieces values adjustments |
| 18-08-12 |
sni |
pr_1725~2 |
diff
|
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 13142 W: 2930 L: 2997 D: 7215
|
sprt
@ 10+0.1 th 1 |
Take 3: combine pieces and pawns |
| 18-08-12 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 27142 W: 6009 L: 6006 D: 15127
|
sprt
@ 10+0.1 th 1 |
2000000. |
| 18-08-12 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 2215 W: 456 L: 578 D: 1181
|
sprt
@ 10+0.1 th 1 |
Different logic. |
| 18-08-12 |
SC |
PVextension |
diff
|
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16859 W: 2795 L: 2851 D: 11213
|
sprt
@ 60+0.6 th 1 |
Only extend: PV above beta / 40000. Speculative LTC |
| 18-08-12 |
Roc |
SimpleImbalance1 |
diff
|
ELO: 1.20 +-2.0 (95%) LOS: 87.7%
Total: 50000 W: 11221 L: 11048 D: 27731
|
50000
@ 10+0.1 th 1 |
Measuring a simplification of GuardianRM imbalance formula. |
| 18-08-12 |
Roc |
SimpleImbalance2 |
diff
|
ELO: -0.40 +-2.0 (95%) LOS: 35.1%
Total: 50000 W: 11080 L: 11137 D: 27783
|
50000
@ 10+0.1 th 1 |
SI2
Same as previous test, but without the pawn imbalances |
| 18-08-11 |
SC |
simpleVoting |
diff
|
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 32194 W: 6830 L: 7056 D: 18308
|
sprt
@ 5+0.05 th 4 |
Use 0 as reference. |
| 18-08-11 |
Gua |
ClosedRank |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 25746 W: 5780 L: 5783 D: 14183
|
sprt
@ 10+0.1 th 1 |
ClosedRank penalty. For cases, if there is an advantage in pawn endings, but it can not be realize it due to the fact that the contact line is closed by own pawns and is attacked by enemy pawns. Example
3k4/8/8/p1p2p1p/P1P2P1P/8/P4P1P/3K4 w - - 0 1 |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 45821 W: 10312 L: 10215 D: 25294
|
sprt
@ 10+0.1 th 1 |
3000000. |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 19545 W: 4360 L: 4395 D: 10790
|
sprt
@ 10+0.1 th 1 |
5000000. |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 64263 W: 14360 L: 14173 D: 35730
|
sprt
@ 10+0.1 th 1 |
Only extend: PV above beta / 40000. |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 36445 W: 8129 L: 8080 D: 20236
|
sprt
@ 10+0.1 th 1 |
Extend PV over beta, only extend. |
| 18-08-11 |
SC |
permatune5 |
diff
|
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 23761 W: 4155 L: 3923 D: 15683
|
sprt
@ 60+0.6 th 1 |
Several promising parameter tweaks at once. LTC |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 27102 W: 6049 L: 6046 D: 15007
|
sprt
@ 10+0.1 th 1 |
Only extend: PV above beta / 30000. |
| 18-08-11 |
jos |
no_value_eg |
diff
|
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17469 W: 3820 L: 4020 D: 9629
|
sprt
@ 10+0.1 th 1 |
Run a first simplification test. The basic question is do we really need to assign big scores to (some) won endgames? This not only creates eval 'holes', but also leads to more complicated code. (Value and ScaleFactor based endgames.) |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13373 W: 2992 L: 3057 D: 7324
|
sprt
@ 10+0.1 th 1 |
Only extend: PV above beta margin. |
| 18-08-11 |
jdo |
simple_razor |
diff
|
ELO: 0.23 +-2.1 (95%) LOS: 58.6%
Total: 36412 W: 6168 L: 6144 D: 24100
|
40000
@ 60+0.6 th 1 |
Both versions of razoring simplification passed [-3, 1] and framework is empty, so at low prio -5 let's collect more data - which is better? |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8000 W: 1730 L: 1822 D: 4448
|
sprt
@ 10+0.1 th 1 |
Only extend: PV above beta / 20000. |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 2243 W: 432 L: 552 D: 1259
|
sprt
@ 10+0.1 th 1 |
Only extend: PV below alpha. |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7687 W: 1663 L: 1757 D: 4267
|
sprt
@ 10+0.1 th 1 |
Extend PV over beta. |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5769 W: 1229 L: 1332 D: 3208
|
sprt
@ 10+0.1 th 1 |
Extend PV over alpha, only extend. |
| 18-08-11 |
SC |
PVextension |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 4886 W: 1015 L: 1122 D: 2749
|
sprt
@ 10+0.1 th 1 |
Extend PV over alpha. |
| 18-08-11 |
SC |
permatune5 |
diff
|
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 10668 W: 2445 L: 2239 D: 5984
|
sprt
@ 10+0.1 th 1 |
Several promising parameter tweaks at once. |
| 18-08-11 |
fau |
FauTune |
diff
|
14152/200000 iterations
29319/400000 games played
|
400000
@ 30+0.3 th 1 |
Long tune for mobility |
| 18-08-11 |
xor |
minNullD |
diff
|
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 4166 W: 898 L: 1010 D: 2258
|
sprt
@ 10+0.1 th 1 |
always search at least one. |
| 18-08-11 |
And |
razor_ether |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 16740 W: 2837 L: 2937 D: 10966
|
sprt
@ 60+0.6 th 1 |
LTC: Simplify to just depth 1 razoring, maintain the PvNode condition. Test with [0, 4], despite being a simplification, to see if this is better than the [-3, 1] running now. |
| 18-08-11 |
And |
razor_ether |
diff
|
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 60550 W: 13543 L: 13139 D: 33868
|
sprt
@ 10+0.1 th 1 |
Simplify to just depth 1 razoring, maintain the PvNode condition. Test with [0, 4], despite being a simplification, to see if this is better than the [-3, 1] running now. |
| 18-08-10 |
jdo |
simple_razor |
diff
|
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 15572 W: 2678 L: 2549 D: 10345
|
sprt
@ 60+0.6 th 1 |
Fix of green STC simplification - possibility 1 |
| 18-08-11 |
Voy |
se-imp |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8585 W: 1855 L: 1944 D: 4786
|
sprt
@ 10+0.1 th 1 |
stc |
| 18-08-10 |
jdo |
simple_razor |
diff
|
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 62670 W: 10749 L: 10697 D: 41224
|
sprt
@ 60+0.6 th 1 |
Fix, possibility 2 |
| 18-08-10 |
Gua |
FieldsControl2c |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 7793 W: 1691 L: 1784 D: 4318
|
sprt
@ 10+0.1 th 1 |
v2c |
| 18-08-10 |
pro |
ps_passed19 |
diff
|
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 42958 W: 9373 L: 9623 D: 23962
|
sprt
@ 10+0.1 th 1 |
try another simplified passed pawn determination. |
| 18-08-10 |
Gua |
FieldsControl2b |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13004 W: 2823 L: 2890 D: 7291
|
sprt
@ 10+0.1 th 1 |
Some better to control fields on queen flank some worse on king (v2b some more simple) |
| 18-08-10 |
Gua |
FieldsControl2 |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 19288 W: 4257 L: 4293 D: 10738
|
sprt
@ 10+0.1 th 1 |
Some better to control fields on queen flank some worse on king |
| 18-08-10 |
Gua |
FieldsControl |
diff
|
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6588 W: 1410 L: 1509 D: 3669
|
sprt
@ 10+0.1 th 1 |
Some better to control fields on king flank some worse on queen |
| 18-08-10 |
jdo |
simple_razor |
diff
|
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 66803 W: 14937 L: 14905 D: 36961
|
sprt
@ 10+0.1 th 1 |
Fix of green STC simplification - possibility 1 |