I'm trying to understand why in the below position, where 3/1 2/1 is the right play, why 3/2 3/1 is such a poor play. Is it related to bear off efficiency, to some need to preserve playable 2's behind O's blot, or something about containment of O in the event of a hit? There must be something pretty routine here that I'm oblivious to.
Bob

XGID=--ABaBBBBBB--a--a-bbbbbb--:0:0:1:21:0:0:3:0:10

Score is X:0 O:0. Unlimited Game, Jacoby Beaver
+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 |
| | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| | | |
| X X X X | | X X X |
| X X X X | | X X O X X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 98 O: 96 X-O: 0-0
Cube: 1
X to play 21

1. Rollout¹ 3/1 2/1 eq:+0.569
Player: 67.89% (G:2.14% B:0.04%)
Opponent: 32.11% (G:3.75% B:0.07%)
Confidence: ±0.010 (+0.559..+0.580) - [100.0%]
Duration: 32.0 seconds

2. Rollout¹ 7/6 7/5 eq:+0.530 (-0.040)
Player: 66.48% (G:2.04% B:0.05%)
Opponent: 33.52% (G:3.29% B:0.05%)
Confidence: ±0.014 (+0.516..+0.543) - [0.0%]
Duration: 25.3 seconds

3. Rollout² 3/2 3/1 eq:+0.489 (-0.081)
Player: 64.84% (G:2.31% B:0.05%)
Opponent: 35.16% (G:4.31% B:0.06%)
Confidence: ±0.012 (+0.476..+0.501) - [0.0%]
Duration: 31.3 seconds

¹ 646 Games rolled with Variance Reduction.
Moves: 3-ply, cube decisions: XG Roller

² 643 Games rolled with Variance Reduction.
Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.10

On Friday, May 12, 2023 at 2:27:46 PM UTC+1, Robert Zimmerman wrote:
> I'm trying to understand why in the below position, where 3/1 2/1 is the right play, why 3/2 3/1 is such a poor play. Is it related to bear off efficiency, to some need to preserve playable 2's behind O's blot, or something about containment of O in the event of a hit? There must be something pretty routine here that I'm oblivious to.
> Bob
>
> XGID=--ABaBBBBBB--a--a-bbbbbb--:0:0:1:21:0:0:3:0:10
>
> Score is X:0 O:0. Unlimited Game, Jacoby Beaver
> +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 |
> | | | |
> | | | |
> | | | |
> | |BAR| |
> | | | |
> | | | |
> | | | |
> | X X X X | | X X X |
> | X X X X | | X X O X X |
> +12-11-10--9--8--7-------6--5--4--3--2--1-+
> Pip count X: 98 O: 96 X-O: 0-0
> Cube: 1
> X to play 21
>
> 1. Rollout¹ 3/1 2/1 eq:+0.569
> Player: 67.89% (G:2.14% B:0.04%)
> Opponent: 32.11% (G:3.75% B:0.07%)
> Confidence: ±0.010 (+0.559..+0.580) - [100.0%]
> Duration: 32.0 seconds
>
> 2. Rollout¹ 7/6 7/5 eq:+0.530 (-0.040)
> Player: 66.48% (G:2.04% B:0.05%)
> Opponent: 33.52% (G:3.29% B:0.05%)
> Confidence: ±0.014 (+0.516..+0.543) - [0.0%]
> Duration: 25.3 seconds
>
> 3. Rollout² 3/2 3/1 eq:+0.489 (-0.081)
> Player: 64.84% (G:2.31% B:0.05%)
> Opponent: 35.16% (G:4.31% B:0.06%)
> Confidence: ±0.012 (+0.476..+0.501) - [0.0%]
> Duration: 31.3 seconds
>

I think I know the answer to this one, but there are
stronger players and analysts on a few nearby planets,
so I may well be corrected.
First of all, your play is not a worse idea than using a
Cadbury's Creme Egg to make an omelette -- 0.08 errors
in this type of position can well be quite subtle.

I think the loss in equity is readily explained by the way your jokers
are so much worse, with no compensating factor.
The correct play gives pick-and-point (I'm not sure whether this is
standard terminology) jokers of 31/41/51/61.

These seem much better than your corresponding jokers of 33/43/53/63,
plus the fact that your corresponding jokers occur 1/36 less often.

If you have the acepoint made, you'd rather keep it company with the 3 point than the 2 point.
Preserving 2's might also be a factor but I think this aspect of the play is a
(slightly) lesser scandal than Watergate, and that the main reason is what I say it is.

Note also that the difference between the two sets of jokers is often preserved on future rolls.

Not such a massive blunder, though.

Paul

On Friday, May 12, 2023 at 9:27:46 AM UTC-4, Robert Zimmerman wrote:
> I'm trying to understand why in the below position, where 3/1 2/1 is the right play, why 3/2 3/1 is such a poor play. Is it related to bear off efficiency, to some need to preserve playable 2's behind O's blot, or something about containment of O in the event of a hit? There must be something pretty routine here that I'm oblivious to.
> Bob
>
> XGID=--ABaBBBBBB--a--a-bbbbbb--:0:0:1:21:0:0:3:0:10
>
> Score is X:0 O:0. Unlimited Game, Jacoby Beaver
> +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 |
> | | | |
> | | | |
> | | | |
> | |BAR| |
> | | | |
> | | | |
> | | | |
> | X X X X | | X X X |
> | X X X X | | X X O X X |
> +12-11-10--9--8--7-------6--5--4--3--2--1-+
> Pip count X: 98 O: 96 X-O: 0-0
> Cube: 1
> X to play 21
>
> 1. Rollout¹ 3/1 2/1 eq:+0.569
> Player: 67.89% (G:2.14% B:0.04%)
> Opponent: 32.11% (G:3.75% B:0.07%)
> Confidence: ±0.010 (+0.559..+0.580) - [100.0%]
> Duration: 32.0 seconds
>
> 2. Rollout¹ 7/6 7/5 eq:+0.530 (-0.040)
> Player: 66.48% (G:2.04% B:0.05%)
> Opponent: 33.52% (G:3.29% B:0.05%)
> Confidence: ±0.014 (+0.516..+0.543) - [0.0%]
> Duration: 25.3 seconds
>
> 3. Rollout² 3/2 3/1 eq:+0.489 (-0.081)
> Player: 64.84% (G:2.31% B:0.05%)
> Opponent: 35.16% (G:4.31% B:0.06%)
> Confidence: ±0.012 (+0.476..+0.501) - [0.0%]
> Duration: 31.3 seconds
>
>
>
>
> ¹ 646 Games rolled with Variance Reduction.
> Moves: 3-ply, cube decisions: XG Roller
>
> ² 643 Games rolled with Variance Reduction.
> Moves: 3-ply, cube decisions: XG Roller
>
>
> eXtreme Gammon Version: 2.10

Turn the cube.

It's about how your board is often constructed over the next rolls. By that I mean do you want your hit and cover numbers/your doubles to make the 3pt leaving the 4pt and 2pt open or do you want them to make the ace point leaving the 4pt and 3pt open. Also more numbers can do this if the blot is closer to all your other checkers.

Stick

On Friday, May 12, 2023 at 9:13:21 PM UTC+1, Stick Rice wrote:
> On Friday, May 12, 2023 at 9:27:46 AM UTC-4, Robert Zimmerman wrote:
> > I'm trying to understand why in the below position, where 3/1 2/1 is the right play, why 3/2 3/1 is such a poor play. Is it related to bear off efficiency, to some need to preserve playable 2's behind O's blot, or something about containment of O in the event of a hit? There must be something pretty routine here that I'm oblivious to.
> > Bob
> >
> > XGID=--ABaBBBBBB--a--a-bbbbbb--:0:0:1:21:0:0:3:0:10
> >
> > Score is X:0 O:0. Unlimited Game, Jacoby Beaver
> > +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 |
> > | | | |
> > | | | |
> > | | | |
> > | |BAR| |
> > | | | |
> > | | | |
> > | | | |
> > | X X X X | | X X X |
> > | X X X X | | X X O X X |
> > +12-11-10--9--8--7-------6--5--4--3--2--1-+
> > Pip count X: 98 O: 96 X-O: 0-0
> > Cube: 1
> > X to play 21
> >
> > 1. Rollout¹ 3/1 2/1 eq:+0.569
> > Player: 67.89% (G:2.14% B:0.04%)
> > Opponent: 32.11% (G:3.75% B:0.07%)
> > Confidence: ±0.010 (+0.559..+0.580) - [100.0%]
> > Duration: 32.0 seconds
> >
> > 2. Rollout¹ 7/6 7/5 eq:+0.530 (-0.040)
> > Player: 66.48% (G:2.04% B:0.05%)
> > Opponent: 33.52% (G:3.29% B:0.05%)
> > Confidence: ±0.014 (+0.516..+0.543) - [0.0%]
> > Duration: 25.3 seconds
> >
> > 3. Rollout² 3/2 3/1 eq:+0.489 (-0.081)
> > Player: 64.84% (G:2.31% B:0.05%)
> > Opponent: 35.16% (G:4.31% B:0.06%)
> > Confidence: ±0.012 (+0.476..+0.501) - [0.0%]
> > Duration: 31.3 seconds
> >
> >
> >
> >
> > ¹ 646 Games rolled with Variance Reduction.
> > Moves: 3-ply, cube decisions: XG Roller
> >
> > ² 643 Games rolled with Variance Reduction.
> > Moves: 3-ply, cube decisions: XG Roller
> >
> >
> > eXtreme Gammon Version: 2.10
> Turn the cube.
>
> It's about how your board is often constructed over the next rolls. By that I mean do you want your hit and cover numbers/your doubles to make the 3pt leaving the 4pt and 2pt open or do you want them to make the ace point leaving the 4pt and 3pt open. Also more numbers can do this if the blot is closer to all your other checkers.
>
> Stick

This is exactly the same point I made.

Paul

On Friday, May 12, 2023 at 4:52:20 PM UTC-4, peps...@gmail.com wrote:
> On Friday, May 12, 2023 at 9:13:21 PM UTC+1, Stick Rice wrote:
> > On Friday, May 12, 2023 at 9:27:46 AM UTC-4, Robert Zimmerman wrote:
> > > I'm trying to understand why in the below position, where 3/1 2/1 is the right play, why 3/2 3/1 is such a poor play. Is it related to bear off efficiency, to some need to preserve playable 2's behind O's blot, or something about containment of O in the event of a hit? There must be something pretty routine here that I'm oblivious to.
> > > Bob
> > >
> > > XGID=--ABaBBBBBB--a--a-bbbbbb--:0:0:1:21:0:0:3:0:10
> > >
> > > Score is X:0 O:0. Unlimited Game, Jacoby Beaver
> > > +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 |
> > > | | | |
> > > | | | |
> > > | | | |
> > > | |BAR| |
> > > | | | |
> > > | | | |
> > > | | | |
> > > | X X X X | | X X X |
> > > | X X X X | | X X O X X |
> > > +12-11-10--9--8--7-------6--5--4--3--2--1-+
> > > Pip count X: 98 O: 96 X-O: 0-0
> > > Cube: 1
> > > X to play 21
> > >
> > > 1. Rollout¹ 3/1 2/1 eq:+0.569
> > > Player: 67.89% (G:2.14% B:0.04%)
> > > Opponent: 32.11% (G:3.75% B:0.07%)
> > > Confidence: ±0.010 (+0.559..+0.580) - [100.0%]
> > > Duration: 32.0 seconds
> > >
> > > 2. Rollout¹ 7/6 7/5 eq:+0.530 (-0.040)
> > > Player: 66.48% (G:2.04% B:0.05%)
> > > Opponent: 33.52% (G:3.29% B:0.05%)
> > > Confidence: ±0.014 (+0.516..+0.543) - [0.0%]
> > > Duration: 25.3 seconds
> > >
> > > 3. Rollout² 3/2 3/1 eq:+0.489 (-0.081)
> > > Player: 64.84% (G:2.31% B:0.05%)
> > > Opponent: 35.16% (G:4.31% B:0.06%)
> > > Confidence: ±0.012 (+0.476..+0.501) - [0.0%]
> > > Duration: 31.3 seconds
> > >
> > >
> > >
> > >
> > > ¹ 646 Games rolled with Variance Reduction.
> > > Moves: 3-ply, cube decisions: XG Roller
> > >
> > > ² 643 Games rolled with Variance Reduction.
> > > Moves: 3-ply, cube decisions: XG Roller
> > >
> > >
> > > eXtreme Gammon Version: 2.10
> > Turn the cube.
> >
> > It's about how your board is often constructed over the next rolls. By that I mean do you want your hit and cover numbers/your doubles to make the 3pt leaving the 4pt and 2pt open or do you want them to make the ace point leaving the 4pt and 3pt open. Also more numbers can do this if the blot is closer to all your other checkers.
> >
> > Stick
> This is exactly the same point I made.
>
> Paul
I think it was synchronicity, judging by the time stamps.
Thanks to you both. Makes sense and I was definitely overlooking the point. Next time something of the same vein comes along, there's a slightly lesser chance of my eating a creme egg omelet.
Bob

On Saturday, May 13, 2023 at 2:00:48 PM UTC+1, Robert Zimmerman wrote:
> On Friday, May 12, 2023 at 4:52:20 PM UTC-4, peps...@gmail.com wrote:
> > On Friday, May 12, 2023 at 9:13:21 PM UTC+1, Stick Rice wrote:
> > > On Friday, May 12, 2023 at 9:27:46 AM UTC-4, Robert Zimmerman wrote:
> > > > I'm trying to understand why in the below position, where 3/1 2/1 is the right play, why 3/2 3/1 is such a poor play. Is it related to bear off efficiency, to some need to preserve playable 2's behind O's blot, or something about containment of O in the event of a hit? There must be something pretty routine here that I'm oblivious to.
> > > > Bob
> > > >
> > > > XGID=--ABaBBBBBB--a--a-bbbbbb--:0:0:1:21:0:0:3:0:10
> > > >
> > > > Score is X:0 O:0. Unlimited Game, Jacoby Beaver
> > > > +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 |
> > > > | | | |
> > > > | | | |
> > > > | | | |
> > > > | |BAR| |
> > > > | | | |
> > > > | | | |
> > > > | | | |
> > > > | X X X X | | X X X |
> > > > | X X X X | | X X O X X |
> > > > +12-11-10--9--8--7-------6--5--4--3--2--1-+
> > > > Pip count X: 98 O: 96 X-O: 0-0
> > > > Cube: 1
> > > > X to play 21
> > > >
> > > > 1. Rollout¹ 3/1 2/1 eq:+0.569
> > > > Player: 67.89% (G:2.14% B:0.04%)
> > > > Opponent: 32.11% (G:3.75% B:0.07%)
> > > > Confidence: ±0.010 (+0.559..+0.580) - [100.0%]
> > > > Duration: 32.0 seconds
> > > >
> > > > 2. Rollout¹ 7/6 7/5 eq:+0.530 (-0.040)
> > > > Player: 66.48% (G:2.04% B:0.05%)
> > > > Opponent: 33.52% (G:3.29% B:0.05%)
> > > > Confidence: ±0.014 (+0.516..+0.543) - [0.0%]
> > > > Duration: 25.3 seconds
> > > >
> > > > 3. Rollout² 3/2 3/1 eq:+0.489 (-0.081)
> > > > Player: 64.84% (G:2.31% B:0.05%)
> > > > Opponent: 35.16% (G:4.31% B:0.06%)
> > > > Confidence: ±0.012 (+0.476..+0.501) - [0.0%]
> > > > Duration: 31.3 seconds
> > > >
> > > >
> > > >
> > > >
> > > > ¹ 646 Games rolled with Variance Reduction.
> > > > Moves: 3-ply, cube decisions: XG Roller
> > > >
> > > > ² 643 Games rolled with Variance Reduction.
> > > > Moves: 3-ply, cube decisions: XG Roller
> > > >
> > > >
> > > > eXtreme Gammon Version: 2.10
> > > Turn the cube.
> > >
> > > It's about how your board is often constructed over the next rolls. By that I mean do you want your hit and cover numbers/your doubles to make the 3pt leaving the 4pt and 2pt open or do you want them to make the ace point leaving the 4pt and 3pt open. Also more numbers can do this if the blot is closer to all your other checkers.
> > >
> > > Stick
> > This is exactly the same point I made.
> >
> > Paul
> I think it was synchronicity, judging by the time stamps.

Yes, I totally overlooked the time stamps despite referring to the concept in an earlier post.
Stick has offered (and maybe still does offer) paid tuition in backgammon so I'm sure he
can explain these things better than I can. For some reason, I missed the "hit and cover"
phrase and chose "pick and point" which is far less standard.

Paul

On 5/12/2023 9:27 AM, Robert Zimmerman wrote:
> I'm trying to understand why in the below position, where 3/1 2/1 is the right play, why 3/2 3/1 is such a poor play. Is it related to bear off efficiency, to some need to preserve playable 2's behind O's blot, or something about containment of O in the event of a hit? There must be something pretty routine here that I'm oblivious to.

Usually, a blot on the 1pt is harder to cover than a blot on the 2pt.
If you're ultimately aiming for a position with no blots then it makes
sense to leave yourself with an easier task than a harder task.

In specific situations, of course, it might be easier to cover a blot
on the 1pt than a blot on the 2pt, but this doesn't seem to be one of
those exceptional situations.

Also if you make the 1pt then you can't prime your opponent; that
consideration comes into play sometimes, but not here really.
---
Tim Chow

On May 12, 2023 at 7:27:46 AM UTC-6, Robert Zimmerman wrote:

> I'm trying to understand why in the below position,
> where 3/1 2/1 is the right play, why 3/2 3/1 is such
> a poor play.....

> XGID=--ABaBBBBBB--a--a-bbbbbb--:0:0:1:21:0:0:3:0:10
> X to play 21
> 1. Rollout¹ 3/1 2/1 eq:+0.569
> 2. Rollout¹ 7/6 7/5 eq:+0.530 (-0.040)
> 3. Rollout² 3/2 3/1 eq:+0.489 (-0.081)

3/2 3/1 is indeed a poor play but you must be given
credit for questioning why 3/1 2/1 is the right play,
because it isn't and hence your being stumped...

I'm kind of stumped also because I would've played
7/6 7/5 but Noo-Bg ranks 9/8 9/7 as better at times.

GNUbg ID: tm0TAQSabdsAAA:cAkFAAAAAAAA

Cubeful 4-ply hint
3/1 2/1 +0.518
9/8 9/7 +0.497 (-0.021)
3/2 3/1 +0.412 (-0.107)
7/6 7/5 +0.407 (-0.111)

What's very interesting is that in "standard", (biased),
rollouts, 3/1 2/1 comes up as best, but 9/8 9/7 and
7/6 7/5 alternate ranks in cubeful, cubeless rollouts
(until you go down to 0-ply in my trials but that may
be due to the sequence of dice rolls which may end
up differently, i.e. in line with higher ply rollouts, with
different sequences of dice rolls, but I don't put much
importance on this anyway).

Cubeful 3-ply 1296 trials
3/1 2/1 +0.562
9/8 9/7 +0.522 (-0.040)
7/6 7/5 +0.505 (-0.057)
3/2 3/1 +0.483 (-0.079)

Cubeless 3-ply 1296 trials
3/1 2/1 +0.345
7/6 7/5 +0.317 (-0.028)
9/8 9/7 +0.300 (-0.045)
3/2 3/1 +0.285 (-0.060)

Cubeful 2-ply 1296 trials
3/1 2/1 +0.572
9/8 9/7 +0.526 (-0.046)
7/6 7/5 +0.497 (-0.075)
3/2 3/1 +0.486 (-0.086)

Ccubeless 2-ply 1296 trials
3/1 2/1 +0.348
7/6 7/5 +0.319 (-0.029)
9/8 9/7 +0.303 (-0.045)
3/2 3/1 +0.286 (-0.062)

Cubeful 1-ply 1296 trials
3/1 2/1 +0.546
9/8 9/7 +0.511 (-0.035)
7/6 7/5 +0.501 (-0.045)
3/2 3/1 +0.475 (-0.071)

Cubeless 1-ply 1296 trials
3/1 2/1 +0.340
7/6 7/5 +0.314 (-0.026)
9/8 9/7 +0.294 (-0.046)
3/2 3/1 +0.282 (-0.058)

Cubeful 0-ply 1296 trials
3/1 2/1 +0.503
7/6 7/5 +0.487 (-0.016)
9/8 9/7 +0.456 (-0.047)
3/2 3/1 +0.362 (-0.141)

Cubeless 0-ply 1296 trials
3/1 2/1 +0.324
7/6 7/5 +0.301 (-0.023)
9/8 9/7 +0.277 (-0.047)
3/2 3/1 +0.269 (-0.055)

Now, when we do some "clinical rollouts", i.e with
almost random checker/cube decisions, even ten
times more trials than usual, i.e. 129,600, these are
the results that we get:

Cubeful 0-ply, max cube+checker noise12960 trials
9/8 9/7 -0.316
7/6 7/5 -0.433 (-0.117)
3/1 2/1 -0.483 (-0.167)
3/2 3/1 -0.536 (-0.220)

Cubeless 0-ply, max cube+checker noise 12960 trials
9/8 9/7 -0.349
7/6 7/5 -0.492 (-0.143)
3/1 2/1 -0.566 (-0.217)
3/2 3/1 -0.608 (-0.259)

Cubeful 0-ply, max cube+checker noise129600 trials
9/8 9/7 -0.296
7/6 7/5 -0.423 (-0.127)
3/1 2/1 -0.501 (-0.205)
3/2 3/1 -0.516 (-0.220)

Even while playing gamblegammon, my backgammon
trained brain first considers the cubeless play, thus
correctly decides that 7/6 7/5 is better than 3/1 2/1
and 3/2 3/1 but I have no idea why the bots says that
9/8 9/7 is better than 7/6 7/5.

I am stumped... :(

MK

On May 13, 2023 at 7:59:56 AM UTC-6, Timothy Chow wrote:

> Usually, a blot on the 1pt is harder to cover than....

This is a combined response to what Paul, Stick
and you have said about 3/1 2/1 vs 3/2 3/1. You
all are somewhat right in your arguments but all
of you three are also wrong about what the right
play is because you all focus on 3/1 2/1 being
the best play just because the bots say so. See
my response to Zimmer on this. I hope that you
all will eventually appreciate what I am doing for
your and for gamblegammon's sake, even though
I neither care so much about any of you mentally
ill gamblers nor about your gamblegammon bots.

MK

On Tuesday, May 16, 2023 at 11:32:15 AM UTC+1, MK wrote:
> On May 13, 2023 at 7:59:56 AM UTC-6, Timothy Chow wrote:
>
> > Usually, a blot on the 1pt is harder to cover than....
>
> This is a combined response to what Paul, Stick
> and you have said about 3/1 2/1 vs 3/2 3/1. You
> all are somewhat right in your arguments but all
> of you three are also wrong about what the right
> play is because you all focus on 3/1 2/1 being
> the best play just because the bots say so.
....

But it isn't "just because the bots say so".
The "just" is wrong.
It's "because the bots say so and the world's best
bg players are bots."

I remember a time when the best bots were abysmally
bad players -- far worse than any competent human.
At that time, no one would argue for a play on the basis
of a bot's verdict.

Paul

On May 16, 2023 at 6:16:37 AM UTC-6, peps...@gmail.com wrote:

> On Tuesday, May 16, 2023 at 11:32:15 AM UTC+1, MK wrote:

>> ... you three are also wrong about what the right
>> play is because you all focus on 3/1 2/1 being
>> the best play just because the bots say so.

> It's "because the bots say so and the world's best
> bg players are bots."

I understand that you are not objecting to what I said
but to how I said it. Feel free to clarify if not so.

I didn't mean that you all have no reasons to believe
what the bots say. However your trust in the bots is
unfounded, undeserved. There has never been any
empirical evidence establishing bots as the world's
best bg players.

Was there ever any prizes offered for humans to beat
the bots, substantial enough sums of money to make
it worthwhile for gamblegammon giants like mockey,
mickey, sticky, etc. to give it a go, (at the risk of being
ridiculed by their peers for daring to try and to fail)?

I hope I haven't missed out of any such opportunity to
not win money but to show that the past/present bots
aren't necessarily the world's best bg players.

In fact, I believe that some players like me who deviate
from the bot play are most likely to beat the bots than
all the gambler giants who strive to play like the bots.

But all that said and saved aside, what do you think of
the results of my *cubeful* "clinical rollout", (i.e. almost
random checher+cube play) of 129 thousand 6 hundred
trials?

Do you still "believe" that the bots' top pick 3/1 2/1 is
the correct play or do you accept that the 9/8 9/7 of
my rollout is the correct play?

MK

On Wednesday, May 17, 2023 at 8:51:37 AM UTC+1, MK wrote:
> On May 16, 2023 at 6:16:37 AM UTC-6, peps...@gmail.com wrote:
>
> > On Tuesday, May 16, 2023 at 11:32:15 AM UTC+1, MK wrote:
> >> ... you three are also wrong about what the right
> >> play is because you all focus on 3/1 2/1 being
> >> the best play just because the bots say so.
> > It's "because the bots say so and the world's best
> > bg players are bots."
> I understand that you are not objecting to what I said
> but to how I said it. Feel free to clarify if not so.
>
> I didn't mean that you all have no reasons to believe
> what the bots say. However your trust in the bots is
> unfounded, undeserved. There has never been any
> empirical evidence establishing bots as the world's
> best bg players.
....

I don't have time to substantiate this fully, but there is strong empirical
evidence that bots are the world's best bg players.
The evidence is that bots generally beat humans in actual play, whether money
or match.

Paul

On May 17, 2023 at 4:58:04 AM UTC-6, peps...@gmail.com wrote:

> On Wednesday, May 17, 2023 at 8:51:37 AM UTC+1, MK wrote:

>> I didn't mean that you all have no reasons to believe
>> what the bots say. However your trust in the bots is
>> unfounded, undeserved. There has never been any
>> empirical evidence establishing bots as the world's
>> best bg players.

> I don't have time to substantiate this fully,

"Partially" would be good enough to start with
if you could spare time for that much.

> but there is strong empirical evidence that
> bots are the world's best bg players.

Where? What kind of evidence? Could you or
anyone else provide at least some pointers to
where I can look for any "evidence" at all..??

I'm also very disappointed that you and others
shy away from commenting on whether 3/1 2/1
or 9/8 9/7 is the correct play. :(

And I'm even more disappointed nobody offered
any opinions on why Noo-BG finds 9/8 9/7 better
than 7/6 7/5 (regardless of 3/1 2/1 vs 3/2 3/1).

You can't possible be all stumped also, can you..?

MK

On Thursday, May 18, 2023 at 1:51:25 AM UTC+1, MK wrote:
> On May 17, 2023 at 4:58:04 AM UTC-6, peps...@gmail.com wrote:
>
> > On Wednesday, May 17, 2023 at 8:51:37 AM UTC+1, MK wrote:
>
> >> I didn't mean that you all have no reasons to believe
> >> what the bots say. However your trust in the bots is
> >> unfounded, undeserved. There has never been any
> >> empirical evidence establishing bots as the world's
> >> best bg players.
> > I don't have time to substantiate this fully,
> "Partially" would be good enough to start with
> if you could spare time for that much.

I can't spare time for that much.

Paul

On May 18, 2023 at 1:55:02 AM UTC-6, peps...@gmail.com wrote:

> On May 18, 2023 at 1:51:25 AM UTC+1, MK wrote:

>> On May 17, 2023 at 4:58:04 AM UTC-6, peps...@gmail.com wrote:

>>> I don't have time to substantiate this fully,

>> "Partially" would be good enough to start with
>> if you could spare time for that much.

> I can't spare time for that much.

I understand... :) Fully... ;)

MK