Fritz5 ECM test suite results (was: Re: Fritz5 is out (to all non-Kornered users in Germany))
Moritz Berger
1. 1s / 6
2. 1s / 6
3. 3s / 8
4. 3s / 8
5. 2s / 8
6. 1s / 6
7. 3s / 7
8. 3s / 7
9. 2s / 7
10. 2s / 8
11. 212s / 11
12. 2s / 8
13. 2s / 7
14. >300s
15. 1s / 5
16. 2s / 7
17. 1s / 6
18. 1s / 5
19. >300s
20. 1s / 5
21. 1s / 6
22. 1s / 7
23. 4s / 8
24. 1s / 8
25. 2s / 7
26. 1s / 5
27. 3s / 9
28. >300s
29. >300s
30. 1s / 4
31. 2s / 8
32. 1s / 6
33. 2s / 7
34. 1s / 6
35. 13s / 9
36. 2s / 7
37. 1s / 5
38. 1s / 7
39. 17s / 7
40. 6s / 8
41. 1s / 6
42. 1s / 5
43. 2s / 6
44. 1s / 5
45. 1s / 7
46. 1s / 6
47. 1s / 7
48. 6s / 9
49. 42s / 10
50. 1s / 5
51. 25s / 10
52. 1s / 6
53. 1s / 6
54. >300s
55. 63s / 9
56. 1s / 5
57. 54s / 9
58. 1s / 6
59. 2s / 8
60. >300s
61. 41s / 10
62. 1s / 5
63. 1s / 5
64. 1s / 5
65. 1s / 5
66. 16s / 9
67. 1s / 5
68. >300s
69. 4s / 7
70. 1s / 6
71. 4s / 8
72. 1s / 5
73. 1s / 5
74. 1s / 6
75. 2s / 7
76. >300s
77. 1s / 6
78. 3s / 7
79. 2s / 9
80. 37s / 8
81. >300s
82. 75s / 11
83. 1s / 4
84. 2s / 6
85. 1s / 6
86. 1s / 6
87. 6s / 8
88. 2s / 5
89. 1s / 5
90. 2s / 8
91. >300s
92. 2s / 8
93. 1s / 6
94. 2s / 7
95. 1s / 6
96. 3s / 7
97. 2s / 4
98. 1s / 7
99. 2s / 7
100. 3s / 8
101. 1s / 4
102. 1s / 5
103. 5s / 7
104. 1s / 6
105. 1s / 6
106. 1s / 6
107. 4s / 8
108. 1s / 4
109. 2s / 7
110. 2s / 6
111. 7s / 9
112. 8s / 10
113. 1s / 8
114. 2s / 7
115. >300s
116. 1s / 5
117. >300s
118. 2s / 7
119. 95s / 10
120. 1s / 6
121. 1s / 5
122. 1s / 6
123. >300s
124. 22s / 9
125. 1s / 3
126. 2s / 8
127. 6s / 8
128. 64s / 9
129. 1s / 6
130. 25s / 9
131. >300s
132. 2s / 8
133. 37s / 10
134. 1s / 4
135. 8s / 8
136. 19s / 10
137. >300s
138. 4s / 7
139. 1s / 6
140. 2s / 8
141. 1s / 4
142. >300s
143. 1s / 6
144. 6s / 8
145. >300s
146. >300s
147. 65s / 10
148. >300s
149. 2s / 7
150. 1s / 5
151. 1s / 5
152. 1s / 4
153. >300s
154. 1s / 5
155. 57s / 10
156. 1s / 5
157. 4s / 9
158. 4s / 8
159. 6s / 8
160. 93s / 10
161. 2s / 6
162. 79s / 10
163. 2s / 5
164. 4s / 9
165. 1s / 6
166. 1s / 5
167. 3s / 8
168. 1s / 5
169. 1s / 6
170. 1s / 5
171. 1s / 5
172. 1s / 6
173. 3s / 7
174. 1s / 4
175. 1s / 3
176. 4s / 8
177. 7s / 8
178. 1s / 4
179. 2s / 8
180. >300s
181. 2s / 6
182. 3s / 7
183. 1s / 5
184. 5s / 7
185. 1s / 6
186. 149s / 10
187. >300s
188. 1s / 6
189. 59s / 10
190. >300s
191. >300s
192. 1s / 5
193. 50s / 10
194. 84s / 12
195. 5s / 8
196. 8s / 8
197. 2s / 6
198. >300s
199. 2s / 8
200. 10s / 10
201. 2s / 8
202. 1s / 4
203. 1s / 5
204. 1s / 5
205. 2s / 7
206. >300s
207. 2s / 9
208. 2s / 7
209. 11s / 8
210. 30s / 10
211. >300s
212. 22s / 9
213. 62s / 10
214. 2s / 6
215. >300s
216. 1s / 6
217. 1s / 6
218. 2s / 9
219. 1s / 5
220. 6s / 8
221. 1s / 7
222. 4s / 7
223. 3s / 7
224. 1s / 5
225. 2s / 6
226. 11s / 10
227. 4s / 9
228. >300s
229. 3s / 7
230. 2s / 5
231. >300s
232. 3s / 7
233. 1s / 5
234. 2s / 8
235. 1s / 6
236. 2s / 6
237. >300s
238. 17s / 9
239. 11s / 9
240. 17s / 10
241. 60s / 9
242. 1s / 5
243. 2s / 7
244. 1s / 5
245. 1s / 6
246. 2s / 6
247. 75s / 10
248. 1s / 6
249. 1s / 5
250. 78s / 10
251. 3s / 8
252. 1s / 6
253. 1s / 4
254. 2s / 5
255. 4s / 8
256. 1s / 6
257. 1s / 5
258. 1s / 7
259. 15s / 10
260. 1s / 4
261. 14s / 8
262. 8s / 9
263. 2s / 6
264. 5s / 8
265. 1s / 4
266. 55s / 11
267. 15s / 9
268. 3s / 7
269. 1s / 4
270. 98s / 11
271. >300s
272. 8s / 7
273. 1s / 5
274. >300s
275. 2s / 6
276. 6s / 8
277. >300s
278. 1s / 5
279. 9s / 8
280. 55s / 10
281. 1s / 4
282. 2s / 6
283. 2s / 8
284. 1s / 6
285. 119s / 10
286. >300s
287. >300s
288. 2s / 6
289. >300s
290. 1s / 5
291. >300s
292. 10s / 8
293. 3s / 7
294. 3s / 6
295. 4s / 7
296. 8s / 8
297. 234s / 11
298. >300s
299. 1s / 4
300. >300s
301. 2s / 7
302. 2s / 9
303. 1s / 5
304. 3s / 6
305. 1s / 5
306. 1s / 5
307. 2s / 7
308. 3s / 8
309. 2s / 7
310. >300s
311. 4s / 10
312. 42s / 9
313. 3s / 7
314. 1s / 5
315. >300s
316. 8s / 9
317. 2s / 9
318. >300s
319. 4s / 8
320. 4s / 8
321. >300s
322. 1s / 5
323. 1s / 5
324. 6s / 8
325. 2s / 8
326. 1s / 5
327. >300s
328. 2s / 9
329. 3s / 8
330. >300s
331. 6s / 9
332. 210s / 11
333. 4s / 7
334. 2s / 7
335. >300s
336. 85s / 10
337. 29s / 10
338. 1s / 6
339. 7s / 8
340. 10s / 8
341. 1s / 5
342. >300s
343. 30s / 10
344. 1s / 5
345. 4s / 8
346. 2s / 7
347. 2s / 7
348. 2s / 8
349. >300s
350. 1s / 5
351. 1s / 6
352. 1s / 5
353. >300s
354. 1s / 6
355. 14s / 10
356. >300s
357. >300s
358. 8s / 9
359. 1s / 5
360. 1s / 5
361. 1s / 6
362. 3s / 9
363. 2s / 7
364. 1s / 4
365. 2s / 7
366. >300s
367. >300s
368. 60s / 12
369. 2s / 7
370. 29s / 9
371. >300s
372. 2s / 6
373. >300s
374. 1s / 4
375. 6s / 9
376. 5s / 7
377. >300s
378. 1s / 5
379. >300s
380. 5s / 8
381. 1s / 4
382. 9s / 8
383. 4s / 8
384. 1s / 5
385. 2s / 8
386. 2s / 6
387. 39s / 9
388. 2s / 7
389. 5s / 10
390. 2s / 7
391. 6s / 9
392. 3s / 8
393. 2s / 8
394. >300s
395. >300s
396. >300s
397. 121s / 11
398. 99s / 11
399. 5s / 8
400. 4s / 8
401. 50s / 9
402. >300s
403. 3s / 7
404. 1s / 4
405. 1s / 5
406. 17s / 9
407. 2s / 7
408. 7s / 9
409. 6s / 8
410. 9s / 9
411. 3s / 7
412. 69s / 10
413. 1s / 5
414. 7s / 9
415. 2s / 6
416. 22s / 9
417. 7s / 8
418. 24s / 10
419. 15s / 10
420. >300s
421. 1s / 4
422. >300s
423. 1s / 5
424. 1s / 4
425. 7s / 8
426. 1s / 6
427. 2s / 8
428. 2s / 7
429. 1s / 5
430. 2s / 7
431. 2s / 8
432. 2s / 6
433. 3s / 8
434. >300s
435. 2s / 4
436. 3s / 7
437. 2s / 7
438. 82s / 10
439. >300s
440. >300s
441. 2s / 8
442. 13s / 9
443. >300s,
444. 1s / 6
445. 2s / 9
446. 3s / 8
447. >300s
448. 1s / 6
449. 5s / 8
450. 3s / 9
451. 2s / 7
452. 2s / 7
453. 1s / 7
454. 1s / 7
455. 1s / 6
456. >300s
457. 1s / 5
458. 2s / 7
459. 8s / 8
460. 42s / 10
461. 7s / 9
462. 1s / 5
463. 230s / 10
464. >300s
465. >300s
466. >300s
467. 10s / 9
468. >300s
469. 165s / 11
470. >300s
471. 9s / 8
472. 146s / 11
473. 3s / 9
474. 12s / 9
475. >300s
476. >300s
477. 9s / 8
478. 1s / 7
479. 2s / 7
480. 1s / 7
481. 1s / 5
482. >300s
483. 5s / 8
484. 1s / 6
485. 13s / 9
486. 1s / 7
487. 40s / 10
488. 164s / 10
489. >300s
490. 3s / 7
491. >300s
492. >300s
493. 1s / 5
494. 14s / 11
495. 3s / 8
496. 6s / 7
497. >300s
498. >300s
499. 1s / 6
500. >300s
501. 45s / 10
502. 1s / 7
503. >300s
504. 196s / 10
505. >300s
506. 6s / 8
507. 24s / 8
508. 1s / 6
509. 1s / 6
510. 22s / 10
511. 1s / 6
512. 8s / 8
513. 1s / 6
514. 10s / 8
515. 2s / 7
516. 27s / 9
517. 1s / 6
518. 2s / 8
519. >300s
520. 5s / 8
521. 11s / 8
522. 1s / 6
523. 1s / 6
524. 1s / 6
525. >300s
526. 2s / 8
527. 1s / 5
528. 1s / 5
529. 1s / 5
530. 1s / 5
531. 2s / 7
532. 1s / 3
533. 1s / 4
534. 2s / 8
535. 1s / 5
536. 1s / 4
537. 2s / 8
538. 2s / 8
539. 19s / 10
540. 3s / 6
541. 1s / 7
542. 1s / 4
543. 1s / 6
544. 6s / 9
545. 6s / 10
546. 1s / 7
547. 1s / 7
548. 1s / 7
549. 3s / 11
550. 2s / 9
551. 3s / 11
552. 2s / 8
553. >300s
554. 2s / 5
555. 1s / 6
556. 1s / 5
557. 17s / 9
558. 2s / 8
559. >300s
560. >300s
561. 1s / 5
562. 1s / 5
563. 1s / 7
564. 1s / 5
565. 14s / 7
566. >300s
567. 2s / 8
568. 2s / 6
569. 18s / 10
570. 1s / 6
571. 1s / 6
572. 1s / 5
573. 1s / 7
574. 1s / 6
575. 1s / 6
576. 1s / 6
577. 1s / 6
578. 9s / 9
579. 8s / 10
580. >300s
581. 1s / 6
582. 1s / 6
583. 1s / 6
584. 1s / 6
585. 2s / 6
586. 3s / 7
587. 45s / 9
588. 1s / 7
589. 20s / 8
590. 1s / 7
591. 20s / 9
592. >300s
593. 1s / 6
594. 7s / 8
595. 55s / 10
596. 1s / 6
597. 2s / 7
598. 10s / 11
599. 2s / 12
600. >300s
601. >300s
602. 1s / 7
603. 1s / 8
604. 3s / 10
605. >300s
606. 1s / 8
607. >300s
608. >300s
609. 40s / 9
610. 53s / 10
611. 1s / 6
612. 2s / 10
613. 11s / 9
614. >300s
615. >300s
616. 187s / 11
617. >300s
618. 1s / 6
619. 1s / 7
620. 2s / 8
621. >300s
622. 1s / 5
623. 227s / 14
624. >300s
625. 7s / 9
626. 5s / 8
627. >300s
628. 1s / 6
629. >300s
630. 1s / 6
631. 3s / 9
632. 1s / 6
633. 23s / 10
634. 3s / 8
635. 156s / 11
636. >300s
637. >300s
638. 1s / 8
639. 2s / 10
640. >300s
641. 9s / 10
642. 1s / 7
643. 12s / 13
644. 89s / 13
645. 1s / 7
646. >300s
647. 1s / 6
648. >300s
649. 1s / 7
650. 3s / 8
651. 1s / 6
652. 53s / 9
653. >300s
654. >300s
655. 3s / 6
656. 2s / 6
657. 2s / 4
658. 10s / 11
659. 67s / 11
660. 1s / 6
661. 2s / 6
662. 6s / 8
663. 11s / 10
664. 4s / 8
665. >300s
666. >300s
667. 28s / 8
668. 197s / 10
669. 42s / 8
670. 2s / 8
671. 1s / 6
672. 6s / 7
673. >300s
674. >300s
675. 24s / 9
676. 10s / 9
677. >300s
678. 1s / 5
679. >300s
680. 109s / 11
681. >300s
682. 23s / 10
683. >300s
684. >300s
685. 55s / 10
686. >300s
687. >300s
688. >300s
689. 1s / 7
690. 93s / 11
691. 2s / 7
692. 6s / 9
693. >300s
694. 6s / 9
695. 4s / 8
696. >300s
697. 184s / 11
698. >300s
699. >300s
700. 2s / 5
701. 1s / 6
702. 2s / 7
703. 1s / 6
704. 1s / 6
705. 1s / 7
706. 25s / 10
707. 2s / 6
708. 8s / 8
709. 7s / 9
710. 1s / 5
711. 1s / 5
712. >300s
713. >300s
714. >300s
715. 19s / 9
716. 73s / 10
717. 4s / 8
718. 194s / 11
719. >300s
720. 2s / 7
721. >300s
722. >300s
723. 1s / 6
724. 2s / 7
725. >300s
726. 2s / 8
727. 2s / 7
728. 42s / 9
729. 275s / 11
730. >300s
731. 2s / 7
732. 1s / 6
733. 38s / 9
734. 30s / 10
735. >300s
736. 64s / 10
737. 1s / 5
738. 1s / 5
739. 1s / 5
740. >300s
741. 3s / 8
742. 5s / 8
743. 4s / 8
744. 2s / 7
745. 6s / 8
746. 1s / 6
747. 1s / 6
748. 2s / 7
749. 8s / 8
750. 19s / 8
751. 3s / 7
752. 24s / 9
753. 6s / 8
754. >300s
755. 4s / 8
756. 1s / 7
757. 2s / 7
758. 1s / 4
759. 1s / 5
760. 2s / 7
761. 3s / 6
762. >300s
763. 6s / 8
764. >300s
765. 5s / 8
766. >300s
767. >300s
768. 2s / 7
769. >300s
770. 1s / 6
771. 68s / 10
772. 3s / 7
773. 6s / 8
774. >300s
775. 1s / 5
776. 143s / 10
777. 2s / 9
778. 2s / 7
779. 259s / 11
780. 100s / 12
781. 3s / 8
782. 2s / 8
783. 4s / 7
784. 1s / 5
785. 36s / 10
786. 2s / 7
787. 3s / 7
788. >300s
789. >300s
790. 3s / 7
791. 2s / 8
792. >300s
793. 3s / 8
794. 7s / 9
795. 34s / 11
796. 2s / 7
797. 1s / 5
798. 1s / 6
799. >300s
800. 1s / 5
801. 58s / 10
802. 1s / 6
803. 6s / 10
804. >300s
805. 205s / 12
806. 1s / 6
807. 2s / 7
808. 1s / 5
809. 7s / 7
810. >300s
811. 2s / 6
812. 7s / 9
813. 25s / 10
814. 24s / 8
815. 2s / 6
816. 17s / 9
817. 2s / 8
818. >300s
819. 8s / 7
820. 2s / 8
821. 1s / 7
822. 4s / 8
823. 1s / 7
824. 8s / 10
825. >300s
826. 179s / 11
827. 33s / 9
828. 1s / 7
829. >300s
830. 2s / 7
831. >300s
832. 49s / 10
833. >300s
834. >300s
835. 1s / 6
836. >300s
837. 40s / 11
838. 116s / 10
839. 2s / 8
840. 1s / 5
841. >300s
842. >300s
843. 254s / 11
844. 31s / 10
845. 39s / 11
846. 47s / 10
847. 1s / 6
848. 11s / 9
849. 11s / 9
850. 8s / 8
851. 1s / 5
852. 7s / 8
853. 86s / 10
854. 6s / 9
855. 1s / 5
856. >300s
857. 85s / 11
858. >300s
859. 2s / 6
860. >300s
861. 5s / 9
862. 1s / 4
863. 7s / 8
864. >300s
865. 3s / 9
866. 46s / 10
867. >300s
868. 3s / 8
869. 1s / 5
870. 2s / 7
871. 24s / 9
872. 24s / 9
873. 1s / 5
874. 5s / 8
875. 118s / 11
876. 1s / 8
877. 41s / 10
878. >300s
879. >300s
brucemo
Moritz Berger wrote:
>
> 1. 1s / 6
> 2. 1s / 6
What hardware? Is this the P5/166?
I checked this out. Here's a table:
Time Solved
---- ------
<=10 564
<=20 598
<=30 622
<=40 634
<=50 648
<=60 660
<=70 669
<=80 674
<=90 679
<=100 685
<=110 686
<=120 689
<=180 697
<=240 708
< 300 711
Which seems pretty impressive to me, it is a hard suite.
Did you really run this for 879*5min = 73 hours, or does it quit the ply
after it finds the solution or something like that?
My own results are 595 in <= 20.00 seconds on a P6/200. I always do
"time until found and held until end of test".
bruce
brucemo
Moritz Berger wrote:
> This might be one thing I could do as well. 595 doesn't sound bad,
> either. Did you ever try it at longer time controls? I found the
> positions solved in <300 seconds but not in <20 seconds to be more
> interesting. Maybe you could run the suite for Ferret with all the
> positions it doesn't get in <= 20 seconds? 595 for Ferret (albeit on a
> P6 200) vs. 598 for Fritz (P166) sounds pretty close, but what about
> number of positions solved at longer time controls? E.g. how about
> #877?
I'm going to respond to this to let you know I got it, and so it won't
disappear from my server.
I'll run some more ECM stuff. #877 had a fail-high fail-low at 8
seconds, maybe it would have found it in the next ply.
bruce
brucemo
Moritz Berger wrote:
> This might be one thing I could do as well. 595 doesn't sound bad,
> either. Did you ever try it at longer time controls? I found the
> positions solved in <300 seconds but not in <20 seconds to be more
> interesting. Maybe you could run the suite for Ferret with all the
> positions it doesn't get in <= 20 seconds? 595 for Ferret (albeit on a
> P6 200) vs. 598 for Fritz (P166) sounds pretty close, but what about
> number of positions solved at longer time controls? E.g. how about
> #877?
I would like to go into some detail with my answer.
I just got done running my thing for one day approximately. I ran each
position for 2 minutes (120 seconds).
I would search less than 2 minutes in exactly one case: if a position
was determined a mate in N, and I had searched to depth (N*2)-1, I would
quit. This isn't strictly correct, since a null move program doesn't
necessarily find the fastest mate this way, it can take you longer than
(N*2)-1 plies to find a mate in N, so perhaps I could miss some shorter
mates this way. I didn't see any examples of this here.
I would search more than 2 minutes in some special cases, which I will
not detail. If you cut the search time off at exactly 2 minutes, my
program finds 667 correct answers. If you tell it to search for 2
minutes, throw out any cases where it finds the correct answer in more
than 2 minutes, yet make the program responsible for instances where it
has the correct answer at 2 minutes, but switches away from the correct
answer after 2 minutes is over, it finds 663.
Fritz finds 689 in 2 minutes, on wimpier hardware.
But I got to wondering what the effect of this, "go for one ply beyond
where you get a solution", mechanism would be.
I searched for cases where the following occurred:
1) It had had a correct answer after *finishing* depth D (for some D).
2) It had also had a correct answer after *finishing* depth D+1.
3) At some later depth it switched to a wrong move, and would still have
played a wrong move at the end of 2 minutes.
I was expecting to find only a couple of these cases, but I found 37 of
them.
I did not attempt to break down these cases, but I noticed some stuff in
passing. In some, it was obviouly not understanding the position, but
had the correct answer for a while, before switching to some other move,
which either won some material or didn't. In other cases, it had +3, +5,
or whatever, and switched away to something else where the score was
higher.
But anyway, I think that this shows that it is important to be clear
about how you are scoring correct answers, since how you do this
seemingly trivial piece of bookkeeping had, in my case, a big affect upon
the score. 667 is a lot different than 704.
By the way, #877 Qe3 is played after 28.3 seconds, ply 9, evaluated after
29.4 seconds, also ply 9, as a draw, so obviously this was a case where
it's all happy because it thinks it's found a perpetual check. It still
thinks it's a perpetual check after 2 minutes.
I may try your idea of making a sub-suite containing all the positions
mine missed, but tonight I'm going to run that Corvax suite.
bruce
Moritz Berger
On Sat, 30 Aug 1997 19:36:35 -0700, brucemo <bru...@seanet.com>
wrote:
< snip >
>I may try your idea of making a sub-suite containing all the positions
>mine missed, but tonight I'm going to run that Corvax suite.
>
>bruce
Fritz 5 (P166) / Covax results:
No. 10 (Nxf7): 155s (11 ply), doesn't change in 10'
No. 22 (..f4): 358s (11 ply), doesn't change in 10'
These are the only 2 positions Fritz gets right within 10 minutes.
Moritz
Ren Wu
On Sat, 30 Aug 1997 19:36:35 -0700, brucemo <bru...@seanet.com>
wrote:
>Moritz Berger wrote:
Mine only found 624 in search exact 2 minutes, on a P6-200.
>But I got to wondering what the effect of this, "go for one ply beyond
>where you get a solution", mechanism would be.
>
>I searched for cases where the following occurred:
>
>1) It had had a correct answer after *finishing* depth D (for some D).
>2) It had also had a correct answer after *finishing* depth D+1.
>3) At some later depth it switched to a wrong move, and would still have
>played a wrong move at the end of 2 minutes.
>
>I was expecting to find only a couple of these cases, but I found 37 of
>them.
>
>I did not attempt to break down these cases, but I noticed some stuff in
>passing. In some, it was obviouly not understanding the position, but
>had the correct answer for a while, before switching to some other move,
>which either won some material or didn't. In other cases, it had +3, +5,
>or whatever, and switched away to something else where the score was
>higher.
I also noticed this and it make me think if the key move is really the
best move. There are many positions that key move is the best move
with the score 3+ or something like that, but then the program found a
move with even higher score.
I wonder if we should compare our results on this suite, not just the
best moves but also the scores etc. I post this a while ago, but I
guess you didn't see it, or maybe you didn't like the idea. I really
like to have a test suite which is big (300+), hard (> n seconds, or >
d ply) and correct.
>But anyway, I think that this shows that it is important to be clear
>about how you are scoring correct answers, since how you do this
>seemingly trivial piece of bookkeeping had, in my case, a big affect upon
>the score. 667 is a lot different than 704.
>
>By the way, #877 Qe3 is played after 28.3 seconds, ply 9, evaluated after
>29.4 seconds, also ply 9, as a draw, so obviously this was a case where
>it's all happy because it thinks it's found a perpetual check. It still
>thinks it's a perpetual check after 2 minutes.
Mine needs 130 seconds to realize it can draw by play the key move.
Too bad!
Ren Wu
On Thu, 04 Sep 1997 22:53:40 -0700, brucemo <bru...@seanet.com>
wrote:
>Ren Wu wrote:
>
>> I wonder if we should compare our results on this suite, not just the
>> best moves but also the scores etc. I post this a while ago, but I
>> guess you didn't see it, or maybe you didn't like the idea. I really
>> like to have a test suite which is big (300+), hard (> n seconds, or >
>> d ply) and correct.
>
>I don't remember this. I miss things now and again, they come in days
>late and look like they are very old, and I don't read them before they
>are recycled, etc.
>
>It's hard to compare something so monstrous as output from from ECM, even
>if it is summarized.
That's true. I may try to come up a way to deal with this. Or I may
post a just few interesting positions so that everyone can try them.
>It might be very useful to have a hard tactical suite made up of very
>hard problems from all the suites we have, but I haven't tried to
>construct such a thing.
I am not sure. I like to have a big and hard tactic suite so that I
can try to minimize the solotion time with confidence. I mainly use
LCT2, Bt2630, and now BS2830, etc as benchmark. These are too small
and I don't really want optimze my program against them. however,
everytime I change my program, I always run them to see if anything I
did wrong.
>Is your program going to be in Paris? Will you be there with it?
>
>Sorry if I already asked, I have been distracted.
Unfortunetely, I will not. I am waiting my PR here and it is not good
time for me to leave US. Also, currently I spent 90% of my time doing
my thesis, there is very little time left for me on my chess program.
This is also the same reason I missed the Jakata event. I hope I will
participate next WMCCC, or even WMCC if possible.
i really wish I can go there through. It must be a great event.
>bruce
Ren
-delete one loop for the proper email address