[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

21$-55P-32O-65R-53P-54D-21 – No Jacoby, No Beavers – 3-ply XG2 Rollout

Posted By: Taper_Mike
Date: Thursday, 12 June 2014, at 10:32 p.m.

In Response To: Who would see this play over the board? (phil simborg)





White is Player 2

score: 0
pip: 127
Unlimited Game
pip: 151
score: 0

Blue is Player 1
XGID=-a-B-BC-A--BbC--ad-e--b-B-:0:0:1:21:0:0:0:0:10
Blue to play 21

1.Rollout16/4 5/4eq: -0.0009
Player:
Opponent:		51.43% (G:14.30% B:0.58%)
48.57% (G:16.80% B:1.71%)		Conf.: ±0.0063 (-0.0072...+0.0053) - [100.0%]
Duration: 41 minutes 36 seconds
2.Rollout113/10eq: -0.0402 (-0.0393)
Player:
Opponent:		49.59% (G:13.55% B:0.50%)
50.41% (G:15.79% B:1.44%)		Conf.: ±0.0060 (-0.0462...-0.0343) - [0.0%]
Duration: 36 minutes 22 seconds
3.Rollout211/8eq: -0.0553 (-0.0544)
Player:
Opponent:		48.92% (G:13.44% B:0.50%)
51.08% (G:14.88% B:1.18%)		Conf.: ±0.0068 (-0.0621...-0.0485) - [0.0%]
Duration: 23 minutes 35 seconds
 
1 7776 Games rolled with Variance Reduction.
Dice Seed: 55966683
Moves: 3-ply, cube decisions: XG Roller

2 5184 Games rolled with Variance Reduction.
Dice Seed: 55966683
Moves: 3-ply, cube decisions: XG Roller

Rollout by Taper_Mike
2014-Jun-12
eXtreme Gammon Version: 2.10.199.2658
21$-55P-32O-65R-53P-54D-21-n-tm.xgp
n[4 039 854] "<=7*5
n[4 032 842] "&e

eXtreme Gammon Version: 2.10

21$-55P-32O-65R-53P-54D-21 — Unlimited Games
Bot Seed Equity Nacbracs Link
GnuBg 1.02 2-ply Trunc 716797341 –55.177 n[4 027 828] '/0.360t Taper_Mike
GnuBg 1.02 2-ply Trunc 716599599 –55.101 n[4 021 828] '/0.360t Taper_Mike
XG 2.10 3-ply, Cubes: XGR 55966683 –0.9 n[4 039 854] "<=7*5 Taper_Mike (above)
XG 2.10 XGR++ Eval [XGR++ seed] +1.5 n[4 032 842] "&e Taper_Mike

21$-55P-32O-65R-53P-54D-21 — According to Score
Bot Seed Equity Nacbracs Link
XG 2.19.206.pre-release 3-ply, Cubes: XGR, MET: Kazaross XG2 1 +2 –5–5 [4 031] "<=1 Phil Simborg

Key: 0 = Zero [13/10], 4 = Four [6/4, 5/4], 8 = Eight [11/8]

Nactation Tutorial • Explanation of nacbracs • Nacbracs for cube actions

What a great position! Thanks, Phil. I hope Tim will consider this one for inclusion in his best-of-the-year list.

Neil said it all, “This is the right play due to tremendous dupe of 4s and 31 and most importantly, the made 11 pt.” The only 4 that could not be used by the opponent to make either his 4pt or 5pt is 64. But that roll would allow him to make his 2pt, while nicely unstacking his 6pt and 8pt in the process. A roll of 44, of course, makes two points.

This position can be reached by either of the well-played sequences below:

21$-55P-32O-65R-53P-54D-21
21$-65R-32O-55P-53P-54D-21

Although XGR++ and my GnuBg trunc agree 100% on the best play and the ranking of lesser plays, there is a significant difference in the absolute equities they assign. The 0.05 disparity in equity is one of the larger values I have seen in an early game position since the new neural nets were installed in GnuBg version 1.xx. As a double check, I reran the trunc using a different seed, but the result was just about the same.

Nactation: One could use upper case P to describe the switching play, 6/4 5/4, but many players would automatically think you mean the play that is properly nactated using lower case p, 11/8. One solution is to use numerals for both. While you are at it, you may as well use 0 = 13/10 for the second-best play instead of lower case d.

Mike

Messages In This Thread

 

Post Response
Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:
 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.