For all that Magriel has contributed, I'm pretty convinced (and I think it's been said by others) that his UTG rule -- Don't expose a blot in your opponent's
inner board to three active builders -- should be rejected.

But what is the reason(s) for this?
Is it 1) 3 is not a high enough threshold -- it should be 4 or 5.
2) If your opponent is preparing to make an inner-board point, it often pays off when you do put your blot there, so that you can fight for it.

I think 2) is a stronger consideration than 1) but I do sometimes think
a bit Magrielishly when a candidate play would be subjecting myself
to 4 active builders instead of 3.

Any (other) opinions?

Thanks,
Paul

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 4/29/2022 4:16 AM, [email protected] wrote:

For all that Magriel has contributed, I'm pretty convinced (and I think it's been said by others) that his UTG rule -- Don't expose a blot in your opponent's
inner board to three active builders -- should be rejected.

But what is the reason(s) for this?
Is it 1) 3 is not a high enough threshold -- it should be 4 or 5.
2) If your opponent is preparing to make an inner-board point, it often pays off when you do put your blot there, so that you can fight for it.

I think 2) is a stronger consideration than 1) but I do sometimes think
a bit Magrielishly when a candidate play would be subjecting myself
to 4 active builders instead of 3.

Any (other) opinions?

The way I would frame the question is, given that there is so little
support for the rule, how did Magriel come to formulate the rule in
the first place?

Over the years, I've come to the conclusion that there is really just
one situation in which Magriel's under-the-gun rule is useful. You
know you want to split, but you have a choice between splitting to
(say) the 3pt vs. the 4pt, or the 4pt vs. the 5pt. Splitting to the
higher point puts you under the gun because of your opponent's
configuration of checkers in the outfield. Otherwise, there's not a
strong reason to prefer one to the other, besides the fact that usually
the higher point is better for you (easier to escape, a better anchor
if you anchor). In such situations, it's often better not to come UTG.

My guess is that Magriel noticed this and came up with the rule to help
him remember what to do in such situations. But then somehow it got
turned into a heuristic principle for deciding *whether* to split, and
it's just not a useful rule for that purpose. Deciding whether to split
should be based on other criteria (does your opponent have an advanced
anchor, what's the pip count, are you more in danger of getting primed
or getting attacked, etc.).

In short, UTG is sometimes useful for deciding *how* to split, but don't
use it (or any version of it) to decide *whether* to split.

---
Tim Chow

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

"[email protected]" <[email protected]> writes:

his UTG rule -- Don't expose a blot in your opponent's inner board to
three active builders -- should be rejected

Walter Trice has a reformulated version in "Backgammon Boot Camp",
roughly it goes like this:

Don't be afraid to split your back anchor in the face of less than
10 checkers.

So UTG is not considered, it is rather number of opponent checkers in
"the zone". IMHO this works much better.

3 is not a high enough threshold -- it should be 4 or 5.

4 I use to call coming "Under The Machine Gun". This is a far more
serious condition, I agree.

As I am as usual eagerly searching for rules of thumb to be employed in speedgammon, I was thinking about how to combine "10 in the zone" with
"UTMG". Not sure whether boolean "and" or boolean "or" works better.

Best regards

Axel

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 4/29/2022 4:11 PM, Axel Reichert wrote:

3 is not a high enough threshold -- it should be 4 or 5.

4 I use to call coming "Under The Machine Gun". This is a far more
serious condition, I agree.

There are still plenty of situations where (for example) splitting
to the opponent's bar point is perfectly fine even when there are
four separate numbers that hit. Below is a basic example where there
are even 10 checkers in the zone.

XGID=-b----E-C---cEaa-c-e----B-:0:0:1:62:0:0:3:0:10

X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game, Jacoby Beaver
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O | | O X |
| X O | | O X |
| X O | | O |
| X | | O |
| X | | O |
| |BAR| |
| | | X |
| | | X |
| O X | | X |
| O X | | X O |
| O X | | X O |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 167 O: 162 X-O: 0-0
Cube: 1
X to play 62

1. Book¹ 24/18 13/11 eq:-0.155
Player: 47.07% (G:12.99% B:0.77%)
Opponent: 52.93% (G:18.24% B:0.86%)
Confidence: ±0.004 (-0.159..-0.151) - [100.0%]

2. Book¹ 24/16 eq:-0.185 (-0.030)
Player: 46.27% (G:11.86% B:0.65%)
Opponent: 53.73% (G:17.76% B:0.73%)
Confidence: ±0.004 (-0.189..-0.181) - [0.0%]

3. Book² 24/22 24/18 eq:-0.197 (-0.042)
Player: 46.14% (G:12.05% B:0.71%)
Opponent: 53.86% (G:18.70% B:0.72%)
Confidence: ±0.007 (-0.204..-0.189) - [0.0%]

4. Book² 13/11 13/7 eq:-0.198 (-0.043)
Player: 45.67% (G:13.61% B:0.79%)
Opponent: 54.33% (G:17.85% B:1.29%)
Confidence: ±0.008 (-0.205..-0.190) - [0.0%]

5. Book² 13/5 eq:-0.204 (-0.049)
Player: 45.51% (G:13.41% B:0.75%)
Opponent: 54.49% (G:18.00% B:1.12%)
Confidence: ±0.007 (-0.211..-0.196) - [0.0%]

7. Book³ 24/22 13/7 eq:-0.214 (-0.060)
Player: 45.02% (G:12.03% B:0.60%)
Opponent: 54.98% (G:17.23% B:0.83%)


¹Generated by David Rockwell on 4/8/2012 using eXtreme Gammon 2.00
20736 Games rolled with Variance Reduction.
Dice Seed: 78709853
Moves and cube decisions: 3-ply

²Generated by GameSite 2000, Ltd on 2/26/2011 using eXtreme Gammon 2.00
10368 Games rolled with Variance Reduction.
Dice Seed: 77390818
Moves: 2-ply, cube decisions: 3-ply Red

³Generated by GameSite 2000, Ltd on 2/26/2011 using eXtreme Gammon 2.00
Analyzed in XG Roller++

eXtreme Gammon Version: 2.19.207.pre-release

---
Tim Chow

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Timothy Chow <[email protected]> writes:

There are still plenty of situations where (for example) splitting
to the opponent's bar point is perfectly fine even when there are
four separate numbers that hit. Below is a basic example where there
are even 10 checkers in the zone.

Thanks. 32D-62S is indeed correct. So maybe a further qualification is
called for, because getting pointed on head on the bar point is less
terrible than getting pointed on head in the opponent's home board. I
checked the opening book for these situations:

- 10 checkers in the zone require a "D" opening.

- "Under The Machine Gun" in opponent's home requires a "large enough"
"S" reply.

There is no such situation without equity loss.

Best regards

Axel

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 4/30/2022 10:22 AM, Axel Reichert wrote:

Thanks. 32D-62S is indeed correct. So maybe a further qualification is
called for, because getting pointed on head on the bar point is less
terrible than getting pointed on head in the opponent's home board.

Yes, in fact part of the motivation for splitting here is that it
distracts O from making a home-board point.

checked the opening book for these situations:

- 10 checkers in the zone require a "D" opening.

- "Under The Machine Gun" in opponent's home requires a "large enough"
"S" reply.

There is no such situation without equity loss.

Generally speaking, the problem I see with any kind of "under the gun"
rule is that it encourages you to focus on only one aspect of the
position. But any decision about whether to split must take into
account several different factors, and weigh them against each other.
In particular, the under-the-gun rule takes into account only the
opponent's attacking potential, and not the opponent's priming
potential, and most splitting decisions come down to an assessment
of priming versus attacking. Focusing only on the attacking aspect
is going to lead to a lot of misjudgments.

---
Tim Chow

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Timothy Chow <[email protected]> writes:

the under-the-gun rule takes into account only the opponent's
attacking potential, and not the opponent's priming potential, and
most splitting decisions come down to an assessment of priming versus attacking

Yes, good point. Some time back I was sorting and grouping a list of
larger errors in checker play and was surprised how often the mistake
was related to staying back sitting on a deep anchor rather than moving
up and trying to "secure an escape number".

Best regards

Axel

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)