I cannot answer everything right now, it's quite late here.
I used Rybka 1.0 Beta 32bit (2.39MB), at 100 ms. I used Strelka 1.8, in fact all engines at 100 ms. Got a result similar to yours among those two, around 74% similarity. I see nothing wrong with your results. I used a variety of many unrelated or wekaly related engines, therefore this 74% was very high. I am astonished to see Strelka 2.0 B having 92% similarity with Rybka 1.0 Beta 2, it is _extremely_ high. Try tu run a renamed Rybka 1.0 Beta 2, to see the noise of this engine. You got me curious to test those too.
My matrix is here
Code: Select all
Key:
1) Critter0.90 (time: 100 ms scale: 1)
2) DRybka3 (time: 100 ms scale: 1.0)
3) DRybka4 (time: 100 ms scale: 1.0)
4) DShredder10 (time: 100 ms scale: 1.0)
5) DShredder12 (time: 100 ms scale: 1.0)
6) Fruit 2.1 (time: 100 ms scale: 1.0)
7) Glaurung 2 (time: 100 ms scale: 1.0)
8) Houdini15aa (time: 100 ms scale: 1.0)
9) Houdini15bb (time: 100 ms scale: 1.0)
10) IvanHoe_B49jA (time: 100 ms scale: 1.0)
11) Komodo13 (time: 100 ms scale: 1)
12) Naum42 (time: 100 ms scale: 1.0)
13) Robbo009a (time: 100 ms scale: 1.0)
14) Ruffian (time: 100 ms scale: 1.0)
15) Rybka 1,0 beta (time: 100 ms scale: 1.0)
16) Spark (time: 100 ms scale: 1.0)
17) Stockfish2.0ja (time: 100 ms scale: 1.0)
18) Strelka18 (time: 100 ms scale: 1.0)
19) Tiger2007 (time: 100 ms scale: 1.0)
1 2 3 4 5 6 7 8 9 10 11 12
13 14 15 16 17 18 19
1. ----- 53.40 51.68 44.93 46.21 49.32 47.72 50.57 51.43 55.29 48.64 51.82 56
.40 44.93 49.34 46.86 51.32 48.48 42.87
2. 53.40 ----- 58.25 45.75 48.43 52.05 48.05 53.57 53.88 57.87 51.53 56.55 59
.65 46.56 52.51 48.36 51.37 52.99 44.80
3. 51.68 58.25 ----- 44.48 47.35 50.51 46.66 53.16 53.65 55.06 49.90 54.53 56
.07 46.22 52.23 47.86 50.30 52.04 44.89
4. 44.93 45.75 44.48 ----- 51.93 45.93 44.87 41.60 40.85 43.53 44.40 44.66 44
.60 46.02 44.61 43.80 44.40 44.40 42.18
5. 46.21 48.43 47.35 51.93 ----- 46.73 43.23 44.46 44.87 46.70 45.84 46.73 47
.78 46.37 45.45 46.18 45.68 46.02 44.68
6. 49.32 52.05 50.51 45.93 46.73 ----- 53.68 44.82 45.36 48.92 48.64 54.96 49
.77 48.05 56.41 47.73 50.45 56.23 43.02
7. 47.72 48.05 46.66 44.87 43.23 53.68 ----- 41.21 41.67 45.05 45.08 49.75 45
.92 46.98 50.00 44.74 49.85 50.02 41.71
8. 50.57 53.57 53.16 41.60 44.46 44.82 41.21 ----- 70.37 58.22 46.15 48.47 58
.64 42.36 44.95 45.99 50.11 44.80 43.86
9. 51.43 53.88 53.65 40.85 44.87 45.36 41.67 70.37 ----- 58.40 46.38 48.41 59
.21 42.73 45.90 46.35 50.06 44.94 43.74
10. 55.29 57.87 55.06 43.53 46.70 48.92 45.05 58.22 58.40 ----- 49.67 52.44 66
.90 44.53 49.53 48.35 50.42 48.66 44.08
11. 48.64 51.53 49.90 44.40 45.84 48.64 45.08 46.15 46.38 49.67 ----- 49.21 49
.84 43.93 48.00 46.77 46.95 47.66 42.18
12. 51.82 56.55 54.53 44.66 46.73 54.96 49.75 48.47 48.41 52.44 49.21 ----- 54
.15 45.82 60.25 47.71 50.70 59.14 44.28
13. 56.40 59.65 56.07 44.60 47.78 49.77 45.92 58.64 59.21 66.90 49.84 54.15 --
--- 45.21 50.97 48.34 51.32 50.64 44.74
14. 44.93 46.56 46.22 46.02 46.37 48.05 46.98 42.36 42.73 44.53 43.93 45.82 45
.21 ----- 45.42 43.48 45.10 45.59 44.08
15. 49.34 52.51 52.23 44.61 45.45 56.41 50.00 44.95 45.90 49.53 48.00 60.25 50
.97 45.42 ----- 46.36 48.65 73.65 41.95
16. 46.86 48.36 47.86 43.80 46.18 47.73 44.74 45.99 46.35 48.35 46.77 47.71 48
.34 43.48 46.36 ----- 48.43 46.86 43.35
17. 51.32 51.37 50.30 44.40 45.68 50.45 49.85 50.11 50.06 50.42 46.95 50.70 51
.32 45.10 48.65 48.43 ----- 48.20 44.04
18. 48.48 52.99 52.04 44.40 46.02 56.23 50.02 44.80 44.94 48.66 47.66 59.14 50
.64 45.59 73.65 46.86 48.20 ----- 41.94
19. 42.87 44.80 44.89 42.18 44.68 43.02 41.71 43.86 43.74 44.08 42.18 44.28 44
.74 44.08 41.95 43.35 44.04 41.94 -----
I am using the latest sim03.exe, it takes longer, but the previous one was noisy and unreliable (too few positions, many of them bad).
For hierarchical cluster analysis I chose the Complete Linkage algorythm: update s_pq by min(s_pr, s_qr) if S is a similarity matrix and s_ij is the similarity measure between cluster i and cluster j. Then perform that itteratively. For measure I have to dig deeper, I did that in SPSS, no clear idea.
Ok, hope that helps.
Kai