Readers vs. Woolsey
RemarQ User
I'm a little surprised about our choice of move yesterday ( 7/8 ). Anchor
the three point this early in the
game? I might agree if he had a builder on the 8 point, is the blitz danger
really severe enough at
this point to anchor now, on the 3 point, rather than leaving them split
thus making it difficult for him
to build his board?
Look at his roll of 61. With the back men split this would have been a very
poor roll. Now, he can slot
the 7 point and all we can hit with is a 4 or double 2's.
Well, Mr. Woolsey agrees with the move, I guess that's why he's a master and
I'm just advanced, that is
unless he was sort of on the fence and leaned toward being friendly with the
30 folks who voted to anchor
on the 3.
It seems to me there is a similar situation in 'New ideas in Backgammon' in
the anchors away chapter
where the value of split men early was explained......I'll have to look
tonight.
Any comments, especially from someone who voted to anchor?
Steve
-**** Posted from RemarQ, http://www.remarq.com/?a ****-
Search and Read Usenet Discussions in your Browser - FREE -
Chuck Bower
| X O O | X | O X | O = Kit Woolsey
| O O | | O X |
| O | | X |
| O | | X |
| | | |
| | | |64
| | | |
| X | | O |
| X | | O |
| X | | X O |
| O X | | X O |
| O X | | X X O | X = GoL subscribers
+-1--2--3--4--5--6-+---+-7--8--9-10-11-12-+
X leads 1-0 in an 11-point match.
X to play 32.
For those who have not subscribed to GammOnline journal
(edited by Kit Woolsey), one of the features is an ongoing
match between Kit and the subscribers. Two plays (one by
the masses and one by Kit) are made each day, and Kit comments
on the moves of BOTH sides when appropriate. The above
position was our decision for July 8.
In article <aOuh3.332$td2.166150@WReNphoon3>,
RemarQ User <sham...@hp.com> wrote:
>Hi,
>
>I'm a little surprised about our choice of move yesterday ( 7/8 ).
(Subscribers voted 30-6 to make the anchor instead of the 11-point.
All agreed to start the roll bar/22.)
>Anchor the three point this early in the game? I might agree if
>he had a builder on the 8 point, is the blitz danger really severe
>enough at this point to anchor now, on the 3 point, rather than
>leaving them split thus making it difficult for him to build his
>board?
>
>Look at his roll of 61.
(This was Kit's actual response after we made the 22-point. Kit
played 13/7, 24/23.)
You need to look at ALL 36 of O's responses before deciding...
>With the back men split this would have been a very poor roll. Now,
>he can slot the 7 point and all we can hit with is a 4 or double 2's.
>
>Well, Mr. Woolsey agrees with the move, I guess that's why he's a
>master and I'm just advanced, that is unless he was sort of on the
>fence and leaned toward being friendly with the 30 folks who voted
>to anchor on the 3.
If that is the case, then one of us doesn't know Kit very well...
>
>It seems to me there is a similar situation in 'New ideas in
>Backgammon' in the anchors away chapter where the value of split
>men early was explained......I'll have to look tonight.
>
>Any comments, especially from someone who voted to anchor?
Yes. I voted for the anchor. Presumably you read Kit's explanation
of why he thought we made the better move. Here was my thinking:
X's choices boil down to an offensive point (11-point) with back
checkers split or the defensive point (22-point) with a builder
(also a blot) on the 11-point. The race is close (even X moves,
but O is on roll). White has made a home board point and we,
until now, have made nothing. With the better board, white will
be in an attacking mood, and if he is lucky enough to point on
our head, he will have even a bigger home board vs. home board
advantage.
Making EITHER point (22- or 11-) is constructive for us. Both
are good moves, and under different conditions--location of
O's checkers, I might choose offense. O's better home board
swings my decision for the defensive point.
I'll be quite interested to know the "New Ideas..." position
you refer to. Please report back with your findings.
Chuck
bo...@bigbang.astro.indiana.edu
c_ray on FIBS
JP White
<snip>
> Well, Mr. Woolsey agrees with the move, I guess that's why he's a master and
> I'm just advanced, that is
>
> Any comments, especially from someone who voted to anchor?
>
> Steve
>
I'd just like to point out to those not fortunate enough to be GammOnline
subscribers, that the match Kit is having with the readers is one of the few
opportunities for less than expert players to play against a top flight player
*and* have him give you insights into the best moves for both sides as you play.
This feature alone is worth more than the annual subscription fee in my opinion.
An opportunity that should not be passed up lightly.
BTW Steve, I did vote to anchor, but as a mere beginner/intermediate player, you
can put it down to beginners luck, rather than impressive deductive analysis.
JP
--
JP White
Mailto:jp.w...@nashville.com
RemarQ User
It's the first problem in the book! Problem #1.
Whites position is exactly the same as Mr. Woolsey's in the move we are
discussing. Black's position is very
similar, the only difference being instead of a checker on the bar and one
on Woolsey's 1 point black has two
checkers on the opponents 1 point, and instead of 4 checkers on the midpoint
there are 3, with a blot each
on the opponents 15 and 16 points. Everything else is the same and black is
on roll with a 2 1. The question
is whether to make the bar point or split the back men and make the 9 point.
The correct move according to the book is to split the back men and make the
9 point, eshewing the bar point.
The reasons have little to do with the value of the 9 point, rather the
value of splitting the back men is
emphasised. The book lists 3 reasons.........
1. No blot dangling on the 10 point. ( this corrosponds with our blot on the
11)
2. SPLITTING THE BACK CHECKERS, QUITE IMPORTANT WITH WHITE'S EIGHT POINT
STRIPPED AND
WHITE HAVING A BUNCH OF CHECKERS STACKED ON THE MIDPOINT. ( which
corrosponds to our
position exactly)
3. Fluidity of the position.
Of course there was not an opportunity to anchor the 3 in this example, but
if you have the book look at it,
the situations are very much the same.
It just seemed to me that posting up on his 3 point this early in the game
with his 8 point stripped and a bunch
of checkers on his midpoint was just an invitation for him to build his
board, rather than splitting the backmen
as a DEFENSE to him building his board. I wasn't scared at all of his two
point board. BIG DEAL. If he wanted
to hit loose let him!! We would simply send his checkers back to the other
side of the board.
RemarQ User
JP wrote..
> I'd just like to point out to those not fortunate enough to be GammOnline
subscribers, that the match Kit is having with the readers is one of the
few opportunities for less than expert players to play against a top flight
player *and* have him
> give you insights into the best moves for both sides as you play. This
feature alone is worth more than the annual
> subscription fee in my opinion. An opportunity that should not be passed
up lightly.
itself with the first issue. I highly
recommend it.
Steve
p.s. I intended no disrespect in questioning Mr. Woolseys agreement with the
readers choice. I am enjoying
learning from him immensley.
Chuck Bower
(From NEW IDEAS IN BACKGAMMON by Kit Woolsey and Hal Heinrich, Gammon
Press, 1996, p 2.):
+24-23-22-21-20-19-+---+18-17-16-15-14-13-+
| X O O | | O X |
| X O O | | O X |
| O | | X |
| O | | |
| | | |
| | | |64
| | | |
| X | | O |
| X | | O |
| X | | X O |
| O X | | X O |
| O X | | X X X O |
+-1--2--3--4--5--6-+---+-7--8--9-10-11-12-+
X to play 21.
>Whites (O's) position is exactly the same as Mr. Woolsey's in the move we are
>discussing. Black's (X's) position is very similar, the only difference being
>instead of a checker on the bar and one on Woolsey's 1 point black has two
>checkers on the opponents 1 point, and instead of 4 checkers on the midpoint
>there are 3, with a blot each on the opponents 15 and 16 points. Everything
>else is the same and black is on roll with a 2 1. The question is whether
>to make the bar point or split the back men and make the 9 point.
>
>The correct move according to the book is to split the back men and make the
>9 point, eshewing the bar point. The reasons have little to do with the
>value of the 9 point, rather the value of splitting the back men is
>emphasised. The book lists 3 reasons.........
>
>1. No blot dangling on the 10 point. ( this corrosponds with our blot on the
>11)
>
>2. SPLITTING THE BACK CHECKERS, QUITE IMPORTANT WITH WHITE'S EIGHT POINT
>STRIPPED AND WHITE HAVING A BUNCH OF CHECKERS STACKED ON THE MIDPOINT.
>( which corrosponds to our position exactly)
>
>3. Fluidity of the position.
>
>Of course there was not an opportunity to anchor the 3 in this example,
details, details...
>but if you have the book look at it, the situations are very much the same.
No. There are a LOT of similarities and a FEW differences.
>It just seemed to me that posting up on his 3 point this early in the game
>with his 8 point stripped and a bunch of checkers on his midpoint was just
>an invitation for him to build his board, rather than splitting the backmen
>as a DEFENSE to him building his board. I wasn't scared at all of his two
>point board. BIG DEAL. If he wanted to hit loose let him!! We would simply
>send his checkers back to the other side of the board.
...if the dice allowed us to do so. Details, details...
(A piece of advice for ALL posters to this newsgroup when discussing
positions: "a diagram is worth a thousand words.") Here is the situation
which occurred in the online match:
http://www.gammonline.com/index.html
+24-23-22-21-20-19-+---+18-17-16-15-14-13-+
| X O O | X | O X | O = Kit Woolsey
| O O | | O X |
| O | | X |
| O | | X |
| | | |
| | | |64
| | | |
| X | | O |
| X | | O |
| X | | X O |
| O X | | X O |
| O X | | X X O | X = GoL subscribers
+-1--2--3--4--5--6-+---+-7--8--9-10-11-12-+
X leads 1-0 in 11-point match. X to play 32 (from the bar).
I review: readers chose b/22, 24/22 over b/24, 13/11 by a 30-6
count. Kit's analysis (which is worth reading--please subscribe!)
agreed with the subscribers.
I applaud Steve for "doing his homework" and looking up the position
in NEW IDEAS IN BACKGAMMON. That puts him ahead of a lot of r.g.bg
readers who just want to post their ideas but not back them up.
Some good things to keep in mind:
1) Just because the subscribers prefers one play by a 30-6 margin
doesn't mean it is necessarily the right play.
2) Just because a world-class human player (Kit) prefers one play
over all others doesn't mean it is necessarily the right play.
3) Just because a world-class robot player (JF or Snowie) prefers
one play over all others doesn't mean it is necessarily the right play.
JF rollouts (from the book) preferred building the 9-point and
splitting over merely building the bar point by 0.04 units of cubeless
equity in Problem 1. SW2.0 evaluation agrees with the choice but says
the difference is 0.03.
For the GoL problem (second of the two positions above), SW likes
the majority play by 0.02.
It looks like the choice of plays in both positions is pretty close
between the top two candidates. This makes answering "which play is
best" a lot tougher than if there were a large disparity. Maybe Steve's
play is better. I don't know. It's certainly reasonable. In a chouette,
with Steve captain, I would respect his decision (assuming he looked and
listened to my side).
The real lesson here, IMO is not actually about these positions in
particular. It is about "reference positions". I have never been a
huge fan of this principle, although in the right hands it can be valuable.
Allow me to summarize:
A 'reference position' is one which you "know" the "correct" action
(doubling decision or checker play). The idea is to memorize the exact
position and use it in the future.
No one (that I know of) who believes in reference positions expects these
to come up EXACTLY in real play, unless they occur either in the first
couple of rolls or late in the bearoff. There are so many BG positions
that most will never recur in your lifetime. (Surely some have never
recurred in the history of backgammon!)
Reference positions are to be used similar to what Steve has done--if
a position is close, see if you can deduce the best move by imagining how
the situation changed between the reference position and the actual one.
Here one side had the identical position. The other side was different.
The decisions (according to the robots) are close in both instances. What
seemed to be a minor difference to Steve was enough to flip Snowie's
choice of moves. So was it worth all the effort to memorize the reference
position. In this case I think the answer is 'no'.
For someone with a photographic memory and good analyzing skills,
reference positions might be quite useful in general. For the typical
player, I'm not so sure. First you must remember the details of the
reference position (location of checkers and cube, match score, equity, etc.)
EXACTLY. Then the position must come up which is identical or very close
to the reference position. Finally you must make the correct deductions
in the differences of the positions in order to reach the correct conclusions.
Sure, it is easy to create some reference positions (again, mostly from
the first couple rolls or the bearoff) which an expert can show to be useful.
But how many refernence positions can your brain hold? And if the answer
is 'a lot', are you capable of makeing the subtle adjustments in equity
to find the right choice in an over-the-board decision?