BGonline.org Forums

32Z-32 vs 52S-32

Posted By: Nack Ballard
Date: Wednesday, 2 December 2009, at 3:05 a.m.

In Response To: 3-2 rev split - 3-2 vs 5-2 split -3-2 (Thomas Myhr)

First, some data:

For 32Z-32, Snowie is [K X11 H21 S49] 31k 5kS, Gnu is [K X3 H6] 46k, and XG is [K H8 X9] 36k, for a straight (i.e., non-weighted) bot average of [K X7 H12].

For 52S-32, Snowie is [H S3 K10 X19] 20k, Gnu is [H S30 K34 X38] 5k, and XG is [H K11 S13 X19] 36k, for a straight average of [H S15 K18 X25]. (I predict the margins of Gnu’s rollout, and therefore the bot-average margins, will narrow a bit after it has been extended past its current 5k.)

[Standard nacbrac abbreviations: Plays within brackets are listed in order of strength, where K = Kill (hit two) 6/3*/1*, X = Hit-and-split 24/22 6/3*, H = Hit (and down) 13/11 6/3*, and S = Split (and down) 24/21 13/11. Error sizes are listed in thousandths after each letter. Outside the brackets, the number of trials rolled out are given in rounded-down thousands; e.g., 5k = 5184.]

It is widely understood that by hitting one checker instead of none, you deny your opponent most of her immediate point-making numbers. For example, for 52S-32, (non-hitting) S is a reasonable play, only .015 worse than H, whereas for 32Z-32, S is worse than H by .028 (and worse than K by .049). Hitting is a lot more worthwhile when the opponent has a 10pt blot/builder (which she can cover or use to make a point or better point with 21 51 53 62 63 64) than when she merely has another spare stacked on her 8pt.

The incentive for hitting two checkers instead of one is similar though not as widely understood. K does relatively better with 32Z-32 than with 52S-32 because there are more cumulative chances to (a) hit the outfield blot, and (b) to take away point-making numbers on subsequent rolls. For both reasons, there is an increased value in keeping the opponent deflected or off balance (“take away” half or full rolls) as long as possible.

The duplication of 3s that you noted is relevant, but it is overshadowed by the effects described above. If there were no duplication of 3s, K’s relative performance versus H in comparing 32Z-32 with 52S-32 would be even greater.

Hope that helps.

Nack

Messages In This Thread

• 3-2 rev split - 3-2 vs 5-2 split -3-2
  Thomas Myhr -- Tuesday, 1 December 2009, at 10:03 p.m.
  • 32Z-32 vs 52S-32
    Nack Ballard -- Wednesday, 2 December 2009, at 3:05 a.m.
    • 32Z-32 vs 52S-32
      rew -- Wednesday, 2 December 2009, at 8:41 a.m.
    • 32Z-32 vs 52S-32
      Thomas Myhr -- Wednesday, 2 December 2009, at 9:19 a.m.
  • 3-2 rev split - 3-2 vs 5-2 split -3-2
    Timothy Chow -- Wednesday, 2 December 2009, at 4:30 a.m.