BGonline.org Forums

Kleinman count investigation

Posted By: Stein Kulseth
Date: Thursday, 7 April 2011, at 8:11 p.m.

I regularily use the Kleinman count to evaluate races (Those of you not familiar with it, or who wants a refresher can see Chuck's presentation of it here: http://www.bkgm.com/rgb/rgb.cgi?view+500), but I cannot recall having seen any modern day investigations of it. It is fairly old, pre-bot, and it would not be too much work to try to verify it with a set of rollouts. So that's what I did, Rollouts were with GNU, initially with my normal 1296 settings, but I shortened it to 432 and 216, seeing that those were sufficient for convergence. I used normal-looking and efficient positions, where there would be no need to modify the pip count for wastage. There still may be distribution issues that makes a position do better or worse than the mere pip count would indicate - but I haven't gone into details on that.

I was pleased to see that the KC did quite well - better than I had thought - for a wide range of pip counts. Without further ado here are my results:

Pip A	Pip B	KC coeff	KC wins	Rollout	Diff %-pt	Pos
40	40	0,21	62,3%	64,9%	-2,6%	Position ID: tm0AALBtAwAAAA
40	41	0,32	65,2%	67,9%	-2,7%	Position ID: um0AALBtAwAAAA
40	43	0,62	71,0%	71,5%	-0,5%	Position ID: 2m4AALBtAwAAAA
40	45	1,00	76,0%	76,0%	0,0%	Position ID: 6nYAALBtAwAAAA
40	47	1,46	80,4%	80,0%	0,4%	Position ID: bHcAALBtAwAAAA
40	50	2,28	85,4%	86,1%	-0,7%	Position ID: 6u4AAGDbBgAAAA
40	55	3,97	91,9%	92,8%	-0,9%	Position ID: 2u4BAMC2DQAAAA
60	61	0,21	62,3%	65,0%	-2,7%	Position ID: 7W4XAADt7hYAAA
60	63	0,41	67,1%	68,0%	-0,9%	Position ID: bXcXAADt7hYAAA
60	65	0,67	71,8%	71,8%	0,0%	Position ID: bXdHAADt7hYAAA
60	67	0,98	76,5%	77,9%	-1,4%	Position ID: bbdLAADt7hYAAA
60	70	1,56	81,0%	82,2%	-1,2%	Position ID: bbuTAADt7hYAAA
60	75	2,76	87,8%	88,5%	-0,7%	Position ID: td2VAADt7hYAAA
60	80	4,24	92,7%	92,6%	0,1%	Position ID: td2VAADt7hYAAA
60	85	5,96	96,0%	95,9%	0,1%	Position ID: 2u6VAADt7hYAAA
80	81	0,16	60,7%	61,4%	-0,7%	Position ID: 2u6ZAADa7lkAAA
80	83	0,31	64,8%	65,7%	-0,9%	Position ID: 2u4ZAgDa7lkAAA
80	85	0,50	68,9%	70,3%	-1,4%	Position ID: 2u4cAgDa7lkAAA
80	87	0,74	73,0%	73,7%	-0,7%	Position ID: 2u4sBADa7lkAAA
80	90	1,18	77,8%	78,6%	-0,8%	Position ID: 2u4MBQDa7lkAAA
80	95	2,11	84,6%	85,8%	-1,2%	Position ID: 2nYWBgDa7lkAAA
80	100	3,27	89,8%	90,2%	-0,4%	Position ID: bDubBADa7lkAAA
80	105	4,65	93,9%	93,7%	0,2%	Position ID: tJ2bBADa7lkAAA
100	101	0,13	59,5%	60,0%	-0,5%	Position ID: 2D03AgDYPTcBAA
100	103	0,25	63,2%	63,9%	-0,7%	Position ID: 2J07AQDYPTcBAA
100	105	0,40	66,9%	67,5%	-0,6%	Position ID: 2J07BADYPTcBAA
100	107	0,60	70,6%	70,8%	-0,2%	Position ID: 2J2bBADYPTcBAA
100	110	0,95	76,0%	75,5%	0,5%	Position ID: 2J0rBgDYPTcBAA
100	115	1,71	82,1%	82,2%	-0,1%	Position ID: 2M5NBgDYPTcBAA
100	120	2,67	87,3%	87,4%	-0,1%	Position ID: 2M4mBwDYPTcBAA
100	125	3,81	91,4%	91,4%	0,0%	Position ID: aOdGBwDYPTcBAA
100	130	5,12	95,1%	94,3%	0,8%	Position ID: sHNHBwDYPTcBAA
120	123	0,21	62,1%	62,4%	-0,3%	Position ID: cOfsBABw59wBAA
120	125	0,34	65,5%	65,6%	-0,1%	Position ID: cOdsBgBw59wBAA
120	130	0,80	73,8%	73,2%	0,6%	Position ID: sHNuBgBw59wBAA
120	135	1,44	80,3%	79,6%	0,7%	Position ID: sDM3BwBw59wBAA
120	140	2,25	85,3%	84,9%	0,3%	Position ID: sDmbBwBw59wBAA
120	150	4,35	93,0%	92,3%	0,7%	Position ID: 0MyuBwBw59wBAA
140	145	0,29	64,3%	65,8%	-1,5%	Position ID: 4OycBwDQ3DkHAA
140	150	0,69	72,1%	71,4%	0,7%	Position ID: 4GzOBwDQ3DkHAA
140	155	1,24	78,4%	77,9%	0,5%	Position ID: YDbPBwDQ3DkHAA
140	160	1,95	83,6%	83,2%	0,4%	Position ID: YGrnBwDQ3DkHAA
140	170	3,78	91,3%	91,0%	0,3%	Position ID: INrzBwDQ3DkHAA

Messages In This Thread

• Kleinman count investigation
  Stein Kulseth -- Thursday, 7 April 2011, at 8:11 p.m.
  • Kleinman count investigation
    Bob Koca -- Friday, 8 April 2011, at 12:42 a.m.
  • Kleinman count investigation
    Phil Simborg -- Friday, 8 April 2011, at 4:06 a.m.
    • +1
      Casper van der Tak -- Friday, 8 April 2011, at 7:22 a.m.
  • Some checker distribution issues
    Stein Kulseth -- Friday, 8 April 2011, at 7:05 a.m.
  • Kleinman count investigation
    Timothy Chow -- Saturday, 9 April 2011, at 5:38 p.m.
    • Kleinman count investigation
      Stein Kulseth -- Wednesday, 20 April 2011, at 7:56 a.m.
  • Kleinman count investigation
    rambiz -- Tuesday, 19 April 2011, at 10:01 p.m.
    • Kleinman count investigation
      Bob Koca -- Wednesday, 20 April 2011, at 6:15 a.m.
      • Kleinman count investigation
        rambiz -- Wednesday, 20 April 2011, at 9:51 a.m.
        • TMG is online ^^
          rambiz -- Thursday, 21 April 2011, at 12:33 a.m.
        • Kleinman count investigation
          Timothy Chow -- Thursday, 21 April 2011, at 1:13 a.m.
          • Kleinman count investigation
            rambiz -- Thursday, 21 April 2011, at 5:19 a.m.