The FFO endgame test suite is a common benchmark for endgame searching. Computer completely solves each testcase, and find the best move. This benchmark evaluates the exact time for searching and the speed (NPS: Nodes Per Second).
| No. | Depth | Best Move | Score | Time (sec) | Nodes | NPS |
|---|---|---|---|---|---|---|
| #40 | 20 | a2 | +38 | 0.037 | 18551129 | 501381864 |
| #41 | 22 | h4 | +0 | 0.063 | 26869773 | 426504333 |
| #42 | 22 | g2 | +6 | 0.123 | 71405077 | 580529081 |
| #43 | 23 | g3 | -12 | 0.183 | 102405636 | 559593639 |
| #44 | 23 | b8 | -14 | 0.071 | 21069844 | 296758366 |
| #45 | 24 | b2 | +6 | 0.466 | 416937288 | 894715210 |
| #46 | 24 | b3 | -8 | 0.177 | 92389718 | 521975807 |
| #47 | 25 | g2 | +4 | 0.091 | 23819448 | 261752175 |
| #48 | 25 | f6 | +28 | 0.353 | 168442255 | 477173526 |
| #49 | 26 | e1 | +16 | 0.46 | 319391867 | 694330145 |
| #50 | 26 | d8 | +10 | 1.629 | 1145908606 | 703442974 |
| #51 | 27 | e2 | +6 | 0.855 | 587578234 | 687226004 |
| #52 | 27 | a3 | +0 | 0.688 | 440750924 | 640626343 |
| #53 | 28 | d8 | -2 | 3.701 | 2689891689 | 726801320 |
| #54 | 28 | c7 | -2 | 5.462 | 5001985645 | 915779136 |
| #55 | 29 | g6 | +0 | 14.02 | 10545655917 | 752186584 |
| #56 | 29 | h5 | +2 | 1.855 | 918717949 | 495265740 |
| #57 | 30 | a6 | -10 | 2.95 | 1918737622 | 650419532 |
| #58 | 30 | g1 | +4 | 2.433 | 1364592150 | 560868125 |
| #59 | 34 | e8 | +64 | 0.469 | 6199035 | 13217558 |
| All | - | - | - | 36.086 | 25881299806 | 717211656 |
Without speedup by SIMD
| No. | Depth | Best Move | Score | Time (sec) | Nodes | NPS |
|---|---|---|---|---|---|---|
| #40 | 20 | a2 | +38 | 0.04 | 15910301 | 397757525 |
| #41 | 22 | h4 | +0 | 0.086 | 24980177 | 290467174 |
| #42 | 22 | g2 | +6 | 0.187 | 78603581 | 420340005 |
| #43 | 23 | c7 | -12 | 0.167 | 66554688 | 398531065 |
| #44 | 23 | d2 | -14 | 0.075 | 13925836 | 185677813 |
| #45 | 24 | b2 | +6 | 0.73 | 404624370 | 554279958 |
| #46 | 24 | b3 | -8 | 0.238 | 85301685 | 358410441 |
| #47 | 25 | g2 | +4 | 0.091 | 22426387 | 246443813 |
| #48 | 25 | f6 | +28 | 0.508 | 174331347 | 343171942 |
| #49 | 26 | e1 | +16 | 0.756 | 389448889 | 515144033 |
| #50 | 26 | d8 | +10 | 2.549 | 1151727651 | 451835092 |
| #51 | 27 | e2 | +6 | 1.231 | 560750484 | 455524357 |
| #52 | 27 | a3 | +0 | 1.165 | 516146220 | 443043965 |
| #53 | 28 | d8 | -2 | 6.587 | 3314489603 | 503186519 |
| #54 | 28 | c7 | -2 | 8.898 | 5070789929 | 569879740 |
| #55 | 29 | g6 | +0 | 22.012 | 10713139055 | 486695395 |
| #56 | 29 | h5 | +2 | 2.81 | 1037039643 | 369053253 |
| #57 | 30 | a6 | -10 | 5.86 | 2882233211 | 491848670 |
| #58 | 30 | g1 | +4 | 3.672 | 1454112944 | 396000257 |
| #59 | 34 | e8 | +64 | 0.522 | 7485471 | 14339982 |
| All | - | - | - | 58.184 | 27984021472 | 480957333 |
Edax 4.4 is one of the best Othello AI in the world.
If I set the game from the very beginning, same line appears a lot. To avoid this, I set the game from many different near-draw lines.
I used XOT for its testcases.
No opening books used.
If Egaroucid Win Ratio is over 0.5, then Egaroucid wins more than Edax do. "Black" and "White" means Egaroucid played black/white. In all conditions, Egaroucid is stronger than Edax.
| Level | Egaroucid win | Draw | Edax Win | Egaroucid Win Ratio |
|---|---|---|---|---|
| 1 | 1246(Black: 602 White: 644) | 59(Black: 29 White: 30) | 695(Black: 369 White: 326) | 0.638 |
| 5 | 1152(Black: 586 White: 566) | 101(Black: 50 White: 51) | 747(Black: 364 White: 383) | 0.601 |
| 10 | 1062(Black: 623 White: 439) | 223(Black: 97 White: 126) | 715(Black: 280 White: 435) | 0.587 |
| 15 | 480(Black: 234 White: 246) | 159(Black: 74 White: 85) | 361(Black: 192 White: 169) | 0.559 |
| 21 | 86(Black: 56 White: 30) | 52(Black: 25 White: 27) | 62(Black: 19 White: 43) | 0.56 |