Noncontact cube action with wastage

[Timothy Chow]

This is the sort of position that can arise after a holding
game in which both players prefer crunching their board to
volunteering a shot---and then they disengage without hitting.

XGID=-DDC-B---AA--a---a-abcbcb-:0:0:1:00:0:0:0:0:10

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 |
| | | O O |
| | | |
| | | |
| |BAR| |
| | | |
| | | X X |
| | | X X X |
| | | X X X X |
| X X | | X X X X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 50 O: 62 X-O: 0-0
Cube: 1
X on roll, cube action

---
Tim Chow

[ah....Clem]

Should be a straightforward race cube decision, right?

Adj pipcount is 55 to 63. 7 pips is the point of last take, so D/P. QF
says it must be a close take but I don't do QF.


--
Ah....Clem
The future is fun, the future is fair.

[Timothy Chow]

XGID=-DDC-B---AA--a---a-abcbcb-:0:0:1:00:0:0:0:0:10

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 |
| | | O O |
| | | |
| | | |
| |BAR| |
| | | |
| | | X X |
| | | X X X |
| | | X X X X |
| X X | | X X X X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 50 O: 62 X-O: 0-0
Cube: 1
X on roll, cube action

The Keith Count gets this one right.


Analyzed in Rollout
No double
Player Winning Chances: 77.34% (G:0.00% B:0.00%)
Opponent Winning Chances: 22.66% (G:0.00% B:0.00%)
Double/Take
Player Winning Chances: 77.30% (G:0.00% B:0.00%)
Opponent Winning Chances: 22.70% (G:0.00% B:0.00%)

Cubeful Equities:
No double: +0.804 (-0.148)
Double/Take: +0.952
Double/Pass: +1.000 (+0.048)

Best Cube action: Double / Take

Rollout:
1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 4-ply, cube decisions: XG Roller+
Search interval: Huge
Confidence No Double: ± 0.002 (+0.802..+0.806)
Confidence Double: ± 0.004 (+0.949..+0.956)

eXtreme Gammon Version: 2.19.207.pre-release

---
Tim Chow

[peps...@gmail.com]

On Friday, July 22, 2022 at 2:16:10 PM UTC+1, Tim Chow wrote:
> XGID=-DDC-B---AA--a---a-abcbcb-:0:0:1:00:0:0:0:0:10
>
> 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 |
> | | | O O |
> | | | |
> | | | |
> | |BAR| |
> | | | |
> | | | X X |
> | | | X X X |
> | | | X X X X |
> | X X | | X X X X |
> +12-11-10--9--8--7-------6--5--4--3--2--1-+
> Pip count X: 50 O: 62 X-O: 0-0
> Cube: 1
> X on roll, cube action
> The Keith Count gets this one right.

That doesn't sound like a very comradely observation!
I think Axelisation also works perfectly here.

Our additions are:
+2 for the open 4 and 6 points.
+4 for the acepoint stack
+2 for the twopoint stack
to give us an adjusted count of 58.

Opponent's addition is:
+1 for the two point stack
to give the opponent an adjusted count of 63.

Now add 1/6 of 58 to 58 to get 67 2/3.
Since 67 2/3 - 63 > 2, we get a (correct) take.
Since 67 2/13 - 63 < 5, we get a redouble and an initial double.

You didn't roll out the redouble case, but with the equity being > 0.95,
it must surely be a redouble as well as an initial double.

So Axelization is my tool of choice here.
Why praise the Keith Count and ignore the fine methodology of
a forum regular?

Did you not try Axelisation? Or am I wrong that it also works here?
I'm a bit surprised at you not mentioning it.

Paul

[Timothy Chow]

On 7/22/2022 1:20 PM, peps...@gmail.com wrote:
> Did you not try Axelisation? Or am I wrong that it also works here?
> I'm a bit surprised at you not mentioning it.

I figured someone else would work that out. The Keith count is
implemented in XG, so I didn't have to do any work to report its
verdict.

---
Tim Chow

[peps...@gmail.com]

Ok. That's a good reason.
Besides the binary right/wrong indicator, there's also the question of
what the adjusted counts say about the marginality of the position,
and whether that verdict is correct.
If that verdict is wrong, then it is grounds for scepticism when that same count
is applied to similar positions.
From this consideration, Axelisation actually seems to do rather poorly.
The adjusted difference is 4 2/3 which (wrongly) indicates a big take rather
than a marginal take.

Let's see how the Keith count does from this perspective.
My adjustments are:
+6 for acepoint stack.
+3 for twopoint stack.
+2 for high gaps.

Adjusted count is 61.
Opponent's adjustments give + 4 to give the opponent
an adjusted count of 66.

I add the floor of 61/7 to my count to get 69.
The difference 69-66 is 3.
Ideally, we would hope for this difference to be 2 indicating that
it's still a take but a very close one.

Here, rounding down helps the count.
If we rounded-nearest rather than down, we would get no redouble
which is almost certainly wrong.

I don't see a performance difference between Tom and Axel when
it comes to this specific position. Both seem to (wrongly) understand
the position as (marginal redouble/ big take).

Paul

[peps...@gmail.com]

Ironically, despite getting the answer wrong, this might be the best understanding of the position.
This finds a marginal drop which is more accurate than the (slight redouble/ big take) given by the
other counts on this thread. There's a fair amount of luck involved in getting the answer right, if we're
limited to naively replicating counts.

Which count did you use? And can I buy some? I actually like it. Although it didn't work here, the
closeness gives me confidence that it will serve me well in all similar positions.
I'd like to buy it for all cases where at least one of the sides has high wastage.
Since I get a race decision in that situation about once per day, I'd like to buy 400 to cover me for one year.

Paul

[ah....Clem]

I used iSight to calculate the adjusted pipcount and then Trice's rule
of 63 to calculate the point of last take. I'm not good at doing
arithmetic in my head, and the Trice metric is easier to use than iSight
while iSight has an easier adjusted pipcount calculation.

Call me lazy, but it's not the weakest part of my game, so I'll work on
other things before refining my race cube strategy.

[peps...@gmail.com]

On Sunday, July 24, 2022 at 9:27:11 PM UTC+1, ah....Clem wrote:
> On 7/23/2022 4:45 AM, peps...@gmail.com wrote:
> > On Wednesday, July 20, 2022 at 11:15:12 PM UTC+1, ah....Clem wrote:
>

> > Which count did you use? And can I buy some? I actually like it. Although it didn't work here, the you

> > closeness gives me confidence that it will serve me well in all similar positions.
> > I'd like to buy it for all cases where at least one of the sides has high wastage.
> > Since I get a race decision in that situation about once per day, I'd like to buy 400 to cover me for one year.
> I used iSight to calculate the adjusted pipcount and then Trice's rule
> of 63 to calculate the point of last take. I'm not good at doing
> arithmetic in my head, and the Trice metric is easier to use than iSight
> while iSight has an easier adjusted pipcount calculation.
>
> Call me lazy, but it's not the weakest part of my game, so I'll work on
> other things before refining my race cube strategy.
> --
> Ah....Clem
> The future is fun, the future is fair.

I think a verdict of marginal pass is fine. It's wrong but only by around 5% and I'm not sure that any humans can
do better than that.
So, yes, don't bother refining your race cube strategy.
Refining your sailing strategy might be a better idea.

Paul

[peps...@gmail.com]

On Sunday, July 24, 2022 at 9:27:11 PM UTC+1, ah....Clem wrote:

> Call me lazy, ...

Wow, we've already been given three things to call you:
Walt Askew
Ah..Clem
Paul Townsend.

I'm not sure I can take any more additions to this list so no,
I don't think I'll be calling you "lazy" as well.

Paul

[Stick Rice]

The fact that Keith got it right is no excuse for using the Keith Count here. Something more fitting to the position should have been used. I'd have used EPC.

Stick