Race cube - ISight, Trice, Keith, something else?

[ah...Clem]

XGID=-AD-FD--------------cdcbc-:1:-1:-1:00:0:0:0:0:10

X:Player 2 O:Player 1
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| | | O O O O |
| | | O O O |
| | | O O O |
| | | O O O |
| | | 6 |
| |BAR| |
| | | |
| | | X |
| | | X X X X | +---+
| | | X X X X X | | 2 |
| | | X X X X X | +---+
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 47 O: 53 X-O: 0-0
Cube: 2, X own cube
X on roll, cube action

[Axel Reichert]

Isight gives a double and close take (some hand-waving might induce me
to pass as O, due the ugly gap on point 3 surrounded by large stacks and
without opportunity to get filled from point 6).

My approximation of EPCs together with Trice's "hidden" criterion for
doubling with EPCs gives not even a double. See page 20 and 21 of my
Isight article about why EPCs are an elegant concept in theory, but hard
to employ in practice.

The subject lines urges me to ask a meta question: Some people like to
have several tools in their boxes: After all, at least one of them is
likely to give the correct answer. The problem, though, is to come up
with clear-cut criteria that help in deciding on which tool is the right
one for the job. The criteria in itself introduce another source for
errors, which is why I normally abstain from them.

Axel

[Timothy Chow]

This might be a close take but I don't like the fact that O is going
to miss with every 3 for the rest of the game. I would pass.

---
Tim Chow

[ah...Clem]

On 9/20/2021 10:45 AM, ah...Clem wrote:

When I apply Isight, I add two to each player's pipcount for the spares
on the ace and deuce to get 49-55. Add 1/6th to X's count for 57.Since
that's only two more than O's pipcount it's a cube and a close take.

Applying Trice, the same adjustment to 49-55, subtract 5 from 49 to get
42, divide by seven to get 6 for the point of last take. With O exactly
6 behind again it's a cube and a close take.

Rollout says otherwise it's a whopper to take this one.

Analyzed in Rollout
No redouble
Player Winning Chances: 80.93% (G:0.00% B:0.00%)
Opponent Winning Chances: 19.07% (G:0.00% B:0.00%)
Redouble/Take
Player Winning Chances: 80.87% (G:0.00% B:0.00%)
Opponent Winning Chances: 19.13% (G:0.00% B:0.00%)

Cubeless Equities: No Double=+0.619, Double=+1.235

Cubeful Equities:
No redouble: +0.910 (-0.090)
Redouble/Take: +1.135 (+0.135)
Redouble/Pass: +1.000

Best Cube action: Redouble / Pass

Rollout:
1296 Games rolled with Variance Reduction.
Moves: 3-ply, cube decisions: XG Roller
Confidence No Double: ± 0.001 (+0.909..+0.912)
Confidence Double: ± 0.004 (+1.131..+1.138)

Double Decision confidence: 100.0%
Take Decision confidence: 100.0%

Duration: 9.5 seconds

eXtreme Gammon Version: 2.10