BGonline.org Forums

Hit or not

Posted By: Daniel Murphy
Date: Wednesday, 2 December 2009, at 6:41 a.m.

In Response To: Hit or not (Timothy Chow)

but this example illustrates that they can be very confusing and misleading sometimes.

Well, (again) I'll say that if one remembers to understand what the bot is telling you, one shouldn't be "very" confused about the presentation.

Xavier is right -- a detailed description of all the probable outcomes in a position where doubles (and perhaps redoubles) and gammons (perhaps for both sides, and on two or more different cube values) are possible, is probably more data than you'd find useful to absorb, while the wrong kind of simplified summary (for instance a presentation of a raw win percentage, where what really matters is how many wins occur on which cube levels) would mislead you in a different way.

Still, in a position like the one below ...

Money game. Blue on roll.

White	5
Blue	7
Position ID: EQAAEAEAAAAAAA Match ID: UQkAAAAAAAAA

Alert: wrong double ( -0.1281)!

Cube decision
Rollout cubeless equity		+0.2380
Cubeful equities:
1.	No double	+0.2392
2.	Double, pass	+1.0000	+0.7608
3.	Double, take	+0.1111	-0.1281
Proper cube action:		No redouble, take (14.4%)
Rollout details
Win W g W bg Lose L g L bg Cubeless Cubeful Player Blue owns 2-cube 0.6190 0.0000 0.0000 - 0.3810 0.0000 0.0000 +0.2380 +0.2392 Standard error 0.0000 0.0000 0.0000 - 0.0000 0.0000 0.0000 0.0000 0.0000 Player White owns 4-cube 0.6190 0.0000 0.0000 - 0.3810 0.0000 0.0000 +0.4760 +0.1111 Standard error 0.0000 0.0000 0.0000 - 0.0000 0.0000 0.0000 0.0000 0.0000
Full cubeful rollout with var.redn.
46656 games, Mersenne Twister dice gen. with seed 842838737 and quasi-random dice
Play: 0-ply cubeful [expert]
Cube: 2-ply cubeful prune [world class]

... althought the equities are obviously correct, it might be nice if a bot were smart enough not to seem to be telling you that Blue wins 61.96% of the time whether or not Blue redoubles.

Messages In This Thread

• Hit or not
  eXtreme Gammon -- Monday, 30 November 2009, at 1:53 a.m.
  • Hit or not
    Stick -- Monday, 30 November 2009, at 2:02 a.m.
    • Hit or not
      neilkaz -- Monday, 30 November 2009, at 3:51 a.m.
      • Hit or not
        Phil Simborg -- Monday, 30 November 2009, at 5:13 a.m.
        • Hit or not
          Matt Cohn-Geier -- Monday, 30 November 2009, at 5:51 a.m.
        • Hit or not
          Daniel Murphy -- Monday, 30 November 2009, at 7:19 a.m.
  • Hit or not
    Matt Cohn-Geier -- Monday, 30 November 2009, at 2:03 a.m.
    • Hit or not
      eXtreme Gammon -- Monday, 30 November 2009, at 2:33 a.m.
      • Hit or not
        lenny -- Monday, 30 November 2009, at 2:40 a.m.
        • Hit or not
          Stick -- Monday, 30 November 2009, at 3:41 a.m.
        • Hit or not
          Stick -- Monday, 30 November 2009, at 4:39 a.m.
  • Hit or not
    Bill Riles -- Monday, 30 November 2009, at 3:10 a.m.
  • Hit or not
    John O'Hagan -- Monday, 30 November 2009, at 3:47 a.m.
  • Hit or not
    Rich Munitz -- Monday, 30 November 2009, at 5:09 a.m.
  • Hit or not
    garyo -- Monday, 30 November 2009, at 4:45 p.m.
  • Rollout and Comments
    eXtreme Gammon -- Tuesday, 1 December 2009, at 12:09 a.m.
    • Rollout and Comments (money)
      eXtreme Gammon -- Tuesday, 1 December 2009, at 12:12 a.m.
    • Rollout and Comments
      Phil Simborg -- Tuesday, 1 December 2009, at 7:28 a.m.
  • Hit or not
    Timothy Chow -- Tuesday, 1 December 2009, at 9:25 p.m.
    • Hit or not
      eXtreme Gammon -- Wednesday, 2 December 2009, at 3:13 a.m.
    • Hit or not
      Daniel Murphy -- Wednesday, 2 December 2009, at 6:41 a.m.