Stockfish Testing Queue

Pending - 0 tests 0.0 hrs

None

Active - 1 tests

20-09-17 An avoidrepsB diff
LLR: -1.34 (-2.94,2.94) [0.00,5.00]
Total: 35324 W: 6448 L: 6354 D: 22522
sprt @ 10+0.1 th 1 Take 2.

Finished - 109 tests

19-09-17 An avoidrepsA diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 22245 W: 4041 L: 4071 D: 14133
sprt @ 10+0.1 th 1 Take on ElbertoOne's avoidrep. Add the condition of alpha >= 0, ie, a draw would not make us happy.
19-09-17 An deceptiveQuiets3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 17251 W: 3081 L: 3133 D: 11037
sprt @ 10+0.1 th 1 Try with 2PLY bonus. Add in the stats cleanup (Correctly this time). Local results on commit comment.
17-09-17 An cleanupStatBonuses diff
LLR: 0.21 (-2.94,2.94) [-3.00,1.00]
Total: 556 W: 107 L: 97 D: 352
sprt @ 10+0.1 th 1 Resolve weird behavior when dealing with stat_bonus at depth 17. First time submitting a patch on [3, 1]. Is this an acceptable one?
17-09-17 An improvingByPlies diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 14149 W: 2558 L: 2623 D: 8968
sprt @ 10+0.1 th 1 No longer using improving as a bool, but as a 3 state variable. For moveCountPruning, index by no improvement, improvement from (ss-2), improvement from (ss-4).
17-09-17 An tacticalResolution diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9311 W: 1669 L: 1755 D: 5887
sprt @ 10+0.1 th 1 Increase reduction in LMR if a tactical move resolves the check, and holds as the best thus far.
17-09-17 An deceptiveQuiets2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 7109 W: 1251 L: 1346 D: 4512
sprt @ 10+0.1 th 1 Take 2 (fixed). Scale with reductions, not just +ONE_PLY for researches.
16-09-17 An deceptiveQuiets diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 53682 W: 9802 L: 9698 D: 34182
sprt @ 10+0.1 th 1 Quiets that beat alpha when reduced, but not at full depth, are deceptive. Apply a bigger malus to push them later in the move list.
16-09-17 An futileReductions diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 26193 W: 4672 L: 4686 D: 16835
sprt @ 10+0.1 th 1 Add a futility pruning like component for reduction in LMR
15-09-17 An tune_volatile diff
12001/12500 iterations
24586/25000 games played
25000 @ 10+0.1 th 1 Tune with better params
15-09-17 An volatileIsolated2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 31087 W: 5540 L: 5534 D: 20013
sprt @ 10+0.1 th 1 Last take, using values from a good local test.
14-09-17 An tune_volatile diff
11544/25000 iterations
23570/50000 games played
50000 @ 10+0.1 th 1 Tune isolated pawns that are contested (fix fix)
14-09-17 An volatileIsolated diff
LLR: -1.38 (-2.94,2.94) [0.00,5.00]
Total: 43060 W: 7854 L: 7729 D: 27477
sprt @ 10+0.1 th 1 If an isolated pawn may be a point of contention, (can attack or be attackedby another pawn), reduce the penalty. If this fails spectacularly, try the opposite.
12-09-17 An ttMoveFromRazor diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 17539 W: 3182 L: 3232 D: 11125
sprt @ 10+0.1 th 1 If we don't have a ttMove, get one when we call qsearch from razoring
10-09-17 An Doubled2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 25158 W: 4445 L: 4464 D: 16249
sprt @ 10+0.1 th 1 Fixed (Was doing the opposite of what I said). Using adjacent_files_bb instead of stoppers, as that makes more sense.
10-09-17 An Doubled diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11694 W: 2092 L: 2167 D: 7435
sprt @ 10+0.1 th 1 Heavier penalty for doubled pawns if their are no enemy pawns that would allow the pawns to 'undouble'
09-09-17 An faltering4 diff
LLR: -2.03 (-2.94,2.94) [0.00,5.00]
Total: 14031 W: 2552 L: 2578 D: 8901
sprt @ 10+0.1 th 1 Last try; Apply to Captures instead of Quiets
07-09-17 An faltering diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 55690 W: 10257 L: 10143 D: 35290
sprt @ 10+0.1 th 1 Increase reduction if we have, from our perspective and our opponents, been losing control of the game over the last 4 half-moves.
08-09-17 An faltering3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 23861 W: 4290 L: 4314 D: 15257
sprt @ 10+0.1 th 1 Same as Take 1, but do it for either player.
08-09-17 An faltering2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 25226 W: 4696 L: 4712 D: 15818
sprt @ 10+0.1 th 1 Take 2.
01-09-17 An ttCaptureAndIID2 diff
LLR: -3.69 (-2.94,2.94) [0.00,5.00]
Total: 30974 W: 5615 L: 5639 D: 19720
sprt @ 10+0.1 th 1 Take 2: Try the opposite.
31-08-17 An ttCaptureAndIID diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12446 W: 2214 L: 2286 D: 7946
sprt @ 10+0.1 th 1 Don't apply ttCapture to cutnodes when the ttMove came from IID. (Fixed Master)
30-08-17 An futileFromIID3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 30102 W: 5680 L: 5674 D: 18748
sprt @ 10+0.1 th 1 Take 3. Final go, barring a very good yellow.
29-08-17 An futileFromIID2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 9922 W: 1729 L: 1812 D: 6381
sprt @ 10+0.1 th 1 Take 2. Reduce range of depths IID-Futile is applied. Increase margin by 150 (ONE_PLY)
29-08-17 An futileFromIID diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 20758 W: 3759 L: 3796 D: 13203
sprt @ 10+0.1 th 1 Futile prune in IID, treating the value from the reduced search as an eval, to reduce the gap between depth and reduced depth.
28-08-17 An qsAMCP diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9314 W: 1641 L: 1727 D: 5946
sprt @ 10+0.1 th 1 STC: MCP in QS, based off of actual number of moves played
21-08-17 An comboStats diff
LLR: -3.36 (-2.94,2.94) [0.00,5.00]
Total: 36576 W: 4588 L: 4606 D: 27382
sprt @ 60+0.6 th 1 LTC
21-08-17 An firstIsTactical diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13909 W: 2482 L: 2548 D: 8879
sprt @ 10+0.1 th 1 Increase reduction if first move is tactical and is to be the best move thus far. Don't double apply with ttCapture
20-08-17 An comboStats diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 73913 W: 13625 L: 13184 D: 47104
sprt @ 10+0.1 th 1 Combo two patches consistently yellow/green
20-08-17 An statsTTWasBad diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 44315 W: 8311 L: 8244 D: 27760
sprt @ 10+0.1 th 1 Increase malus for ttMove if it was beaten by a quiet move. Opposite has been done, but has this?
19-08-17 An researchStatBonus5 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 25762 W: 4667 L: 4682 D: 16413
sprt @ 10+0.1 th 1 STC. Use the adjustment from statscore for scaling the bonus. Make this into more of a corrective bonus for when moves have reductions increased via statscore incorrectly.
19-08-17 An lateMoveReductionPrunin diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13637 W: 2560 L: 2626 D: 8451
sprt @ 10+0.1 th 1 Prune moves facing massive reductions in LMR
19-08-17 An myRazorPV diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 25080 W: 4637 L: 4654 D: 15789
sprt @ 10+0.1 th 1 STC
19-08-17 An statMalusByOrder diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9037 W: 1581 L: 1668 D: 5788
sprt @ 10+0.1 th 1 Increase malus for quiet moves tried first.
18-08-17 An statBonusByQuietCount2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10857 W: 1697 L: 1777 D: 7383
sprt @ 10+0.1 th 1 Take 2. No cap on bonus. Add a malus component.
18-08-17 An statBonusByQuietCount diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 43433 W: 7138 L: 7090 D: 29205
sprt @ 10+0.1 th 1 Stats bonus for best quiet move, based on how many bad quiets were tried first. Capped at 2*depth.
16-08-17 An researchStatBonus4A diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 75210 W: 12416 L: 12242 D: 50552
sprt @ 10+0.1 th 1 Add depth to the equation. (Park with prio -1 until researchStatBonus4 fails)
17-08-17 An researchStatBonus4B diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 82295 W: 13532 L: 13331 D: 55432
sprt @ 10+0.1 th 1 STC : Same as researchStatBonus4, but double the bonus and limit it to depths at or below twelve, similar to stat_bonus(). May I stop researchStatBonus4A?
17-08-17 An researchStatBonus4 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 15961 W: 1763 L: 1833 D: 12365
sprt @ 60+0.6 th 1 LTC
15-08-17 An researchStatBonus4 diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 151824 W: 25179 L: 24459 D: 102186
sprt @ 10+0.1 th 1 STC Take 4. Scale with R value from LMR instead of node depth.
15-08-17 An researchStatBonus3 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 9186 W: 1503 L: 1590 D: 6093
sprt @ 10+0.1 th 1 STC Take 3. Last try using this idea of stat_bonus.
14-08-17 An researchStatBonus2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 45399 W: 7318 L: 7264 D: 30817
sprt @ 10+0.1 th 1 Take 2. Increase bonus to the best move, but do not increase malus for the remainder.
14-08-17 An researchStatBonus diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 11335 W: 1838 L: 1916 D: 7581
sprt @ 10+0.1 th 1 Stat bonus if the best move had to be researched
14-08-17 An lmrIntoQS2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7192 W: 1252 L: 1347 D: 4593
sprt @ 10+0.1 th 1 Take 2, STC.
14-08-17 An lmrIntoQS diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 17337 W: 3097 L: 3149 D: 11091
sprt @ 10+0.1 th 1 Let LMR descend into QS if we are moving a non-pawn to a square attacked by our opponent. Hope is that the best move for our opponent is to capture the piece we just moved, removing the loss of accuracy from not doing a regular search.
13-08-17 An lmrDepthMin diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 18322 W: 3291 L: 3338 D: 11693
sprt @ 10+0.1 th 1 LMR depth (in actual LMR) won't ever be DEPTH_ZERO, so don't let lmrDepth for pruning by DEPTH_ZERO either.
13-08-17 An statsForBadCaps diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 25768 W: 4545 L: 4562 D: 16661
sprt @ 10+0.1 th 1 Extend LMR inc/dec based on stats to bad captures that were not preceded by an appropriately scoring quiet move
13-08-17 An ensureStatScore diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8193 W: 1426 L: 1516 D: 5251
sprt @ 10+0.1 th 1 If the previous move was tactical, statscore is not set. Do not perform LMR inc/dec based on an non set statscore.
12-08-17 An cutNodesImproving2 diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 10798 W: 1879 L: 1958 D: 6961
sprt @ 10+0.1 th 1 Take 2. Passed locally at 2s+.02s...
12-08-17 An cutNodesImproving diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 27844 W: 4893 L: 4901 D: 18050
sprt @ 10+0.1 th 1 Req no researches and a non improving position for +2PLY for cutNode LMR
12-08-17 An poorImprovementB diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 42488 W: 7484 L: 7432 D: 27572
sprt @ 10+0.1 th 1 Take 2, try a different usage