7000, 7001, 7002 are the same. Devilish Endgame, too. Spoller Alert

#1

Here’s the Kingsrow play. Masterful play.

[FEN “B:WK1,24,25,27,28,30,32:B7,10,12,16,18,20,21”]

  1. 10-14 24-19 2. 16x23 25-22 3. 18x25 27x9 4. 25-29 9-6 5. 29-25 6-2 6. 7-11 1-6 7. 25-22
    6-10 8. 22-17 32-27 9. 11-16 28-24 10. 17-22 10-14 11. 22-17 14-18 12. 17-13 2-6 13. 13-17
    6-10 14. 17-13 18-23 15. 13-17 23-26 16. 17-13 10-14 17. 13-17 14-18 18. 17-13 18-23 19.
    13-17 24-19 20. 17-14 27-24 21. 20x27 23x32 22. 16x23 26x19 23. 14-18 32-27 24. 18-14
    27-23 25. 14-10 23-18 26. 10-6 18-22 27. 6-10 30-26 28. 10-7 26-23 29. 7-11 19-24 30.
    11-16 24-20 31. 16-19 23x16 32. 12x19 22-26 33. 21-25 26-31 34. 25-30 31-27 35. 30-26
    27-23 36. 26-22 23x16 37. 22-18 16-11 38. 18-14 11-7 39. 14-9 7-10 40. 9-5 10-14 41. 5-1
    20-24 42. 1-5 24-27 43. 5-1 27-23 44. 1-5 23-18 45. 5-1 14-9 46. 1-5 18-14 47. 5-1 9-5 48.
    1-6 14-18 49. 6-1 18-15 50. 1-6 5-1 51. 6-9 1-6 52. 9x2 15-10 53. 2-7 10x3 *
#2

Having multiple version of puzzles is intentional, each being a move shorter than the last. The rating system should separate them to easy, medium, and hard versions. It doesn’t work as well here where the primary difficulty is winning the endgame afterwards but there are lots of examples like this. For instance, problems 1, 2, and 3 are the same tactic but shorter each time.

I could prevent this, and chesstempo does prevent it. But I’m not sure why they do. I guess it loses some variety for the same number of puzzles, and more puzzles takes longer for the rating system to catch up. Maybe also if they get rated close together they look like repeats? Hmm, maybe there are good reasons. Any thoughts for the next version of the problem generator?