BG is a different game than when I left it 7 years ago....this is sick.

How big a factor is black's anchor in the outfield here too?

In my limited experience often these games seem to break down into the side that controls the outfield best coming out on top - simply because even if red anchors it will be so hard to get its pieces back around without getting hit the minute black gets his men moving out and controls the out field. Multiple hits usually follow, and I can easily see red ending up with numerous men on the rail later even if they anchor very quickly.

Still way above my head this problem. I one minute grunched it as a double/take (missed it was a 3 point match which prob would have made a difference to this), although with some thought and reading the analysis itt I'm now pretty convinced it must be a double/drop - and maybe even a too good to double/pass.

XG 1296 trial 3 ply with VR for money gives a .989 take.

I crancked up to 100% confidence, stopped at 9077 trials. Look at the take decision confidence [Confidence Double: ± 0,019 (+1,017<E<+1,055)]. 1296 trials are not enough.

No Double
Player Winning Chances: 66,11% (G: 37,16% B: 5,73%)
Opponent Winning Chances: 33,89% (G: 12,35% B: 1,60%)
Double/Take
Player Winning Chances: 67,72% (G: 37,79% B: 5,11%)
Opponent Winning Chances: 32,28% (G: 12,05% B: 1,80%)

Cubeless Equities: No Double=+0,322, Double=+1,290

Cubeful Equities:
No Double: +0,798 (-0,202)
Double/Take: +1,036 (+0,036)
Double/Drop: +1,000

Best Cube action: Double / Drop

Rollout
9077 Games rolled with Variance Reduction.
Moves and cube decisions: 3 ply
Confidence No Double: ± 0,011 (+0,787<E<+0,809)
Confidence Double: ± 0,019 (+1,017<E<+1,055)

Double Decision confidence: 100,0%
Take Decision confidence: 100,0%


I have to check up my post with the market windows. I wrote Xavier an E-Mail. Something with the cube information window. I don't know if there is a fault be me or XG or mishandling it.

And here for match:

No Double
Player Winning Chances: 66,73% (G: 37,36% B: 7,14%)
Opponent Winning Chances: 33,27% (G: 12,67% B: 3,21%)
Double/Take
Player Winning Chances: 67,51% (G: 37,47% B: 14,34%)
Opponent Winning Chances: 32,49% (G: 13,66% B: 5,22%)

Cubeless Equities: No Double=+0,813, Double=+1,487

Cubeful Equities:
No Double: +0,935 (-0,065)
Double/Take: +1,253 (+0,253)
Double/Drop: +1,000

Best Cube action: Double / Drop

Rollout
1296 Games rolled with Variance Reduction.
Moves and cube decisions: 3 ply
Confidence No Double: ± 0,038 (+0,898<E<+0,973)
Confidence Double: ± 0,059 (+1,194<E<+1,312)

Double Decision confidence: 100,0%
Take Decision confidence: 100,0%

This one is another remarkable norwegian position.

LOL. Yeah, lets all be happy that we actually have players that take pictures and notes about the really interesting positions that we all can probably learn something from(even though this is about as complex as it gets).

Actually, it's not as bad as all that.

There is backgammon knowledge which is very useful and which makes you a better player, such as knowing where to put your checkers, knowing what sort of structures play well and what sort are awkward, and knowing how to do a pip count. All of these things are important and wil make you much better when you master them.

Then there is backgammon knowledge which is very arcane and makes almost no difference in your play. A lot of the mathematical tournament strategy, such as the post you were referring to, fits this category. It matters a little, and if you want to be a very top player you'll need to look at this stuff at some point, if only to develop some simple rules you can apply over the board.

Imagine you went to a party and met a couple of mechanical engineers. One knew the value of pi to 5 decimal places. The other knew pi to 100 decimal places. Who's the better engineer? Answer -- no difference. It's possible to have lots of very specialized knowledge which doesn't affect your results much. The trick is knowing what's important and what's not, so you can concentrate on the important stuff and actually play better.

Since you have posted in this thread Robertie, but also seen the rollouts unfortunately. What would you play over the board? The play was double take in the actual game. We have seen through rollouts that this probably is wrong but the rollouts vary a lot too. I feel this is a very good practical double, what are your thoughts?

Preposterous to pretend that this is a proposition that tests the position itself. If it goes double and take, an extended middle game is likely to result which will afford both players multiple opportunities to make big errors.

If you are offered such a "proposition", think about whether it's simply a test of who is the better player. With the cube in a disadvantageous spot.

atrifix=MCG=Matt Cohn-Geier would give you 1 point and the cube, and i think, for every amount, that you can afford. So i think this is a pass and a good prop for practical purposes to lure someone in.

And Xavier wrote me:

Hello,



The Number displayed in the Cube information are different than the ones in the analyze: in the cube information it is the (back)gammon ratio.

In this case for the bottom player: the analyze show 66.73% game won, 37.36% are gammon (or better) and 7.14% are backgammon

The values in the cube information a hence:

· Backgammon ratio= 7.14/66.73=10.70% of game won that are backgammons

· Gammon ratio =(37.36-7.14)/66.73=45.29&% of game won that are Gammons



So anything right, if you look in the cube information window for a centered cube, and this is the information for the decision to double to 2.

And last, but not least, a rollout for match or for money at this score must vary. 1296 trials was not enough in the light of the standard deviation (confidence).

Dear Bill,

i looked at work at this position, made a guess, and then i began the workout, like problemchess. At this time, luncheon, i reread Kits "How to play tournament backgammon" and your game against him Reno 1993, given as example. If you play tournament backgammon, you have to work this over the board, and i think you can do the math approximately over the board. So it was not only 2003 but even 1993 an issue, which you have to manage for successful matchplay. And no one who will play a major tournament is successful, without this considarations. What "sick" is: 42.000k RO for 3ply opening theory, databases about the flaws and biases of your opponents to adjust checkerplay. Modern Backgammon is fighting very hard against the variance of the dice.

An interesting point about this position for money: I gave white a 2 cube and did a long rollout. It was almost as big of a double as the center cube, and a much bigger take (1.138)? It seemed to me that a lot of the value of the initial cube (Jacoby on, of course) was in activating gammons, but this is apparently not the case, nor was it a major factor to keep an owned cube in a very volatile position. Also, in such a swingy position it would seem that owning a 4 cube would be fairly strong. Although I suspected that the best play would not change, I'm surprised at the numbers. Why is it so much worse for black to own a 4 cube, and why does the fact that it is not a center cube not matter as much as I thought it would (no double equity .832)?

I don't understand. Do you mean, that black owns the cube? It white owns the cube, there is no double/beaver.

Sorry it's white on my XG lol...yes the stronger side owns the cube.

No, I'm the fish. That's why I said "this fish is hooked." Prop is open.

No Double
Player Winning Chances: 66,11% (G: 37,16% B: 5,73%)
Opponent Winning Chances: 33,89% (G: 12,35% B: 1,60%)
Double/Take
Player Winning Chances: 67,72% (G: 37,79% B: 5,11%)
Opponent Winning Chances: 32,28% (G: 12,05% B: 1,80%)

Cubeless Equities: No Double=+0,322, Double=+1,290

Cubeful Equities:
No Double: +0,798 (-0,202)
Double/Take: +1,036 (+0,036)
Double/Drop: +1,000

Best Cube action: Double / Drop

Rollout
9077 Games rolled with Variance Reduction.
Moves and cube decisions: 3 ply
Confidence No Double: ± 0,011 (+0,787<E<+0,809)
Confidence Double: ± 0,019 (+1,017<E<+1,055)

Double Decision confidence: 100,0%
Take Decision confidence: 100,0%

So you think 0,036 isn't enough? Or so borderline that you think you can win from both sides against most of the fishes as fish under fishes?!

I think many people will make some mistakes playing Red and apparently I would have .036 the best of it to start with anyway. I didn't see the rollout until now, though.

I guess I was a little off about the beaver, but I still like this as a take for a prop. As atrifix said, there is a lot of room for red to make mistakes, but I think thats true for black also. bearing in against this position could create its share of errors.