FIBS Board backgammon forum

Backgammon => Fibsboard Forum Matches => FFM1 => Topic started by: blitzxz on May 17, 2009, 10:26:20 PM

Poll
Question: 5-1?
Option 1: bar/24 15/10 votes: 0
Option 2: bar/24 13/8 votes: 0
Option 3: bar/24 11/6 votes: 7
Option 4: bar/24 6/1 votes: 6
Title: Match 1, Game 3, Move 8, Forum
Post by: blitzxz on May 17, 2009, 10:26:20 PM
5-1?

Score:
Green(forum) 2, White(factotum) 0, match to 5

Pips:
Green 146 White 163

GnuID:
sN3gABMcZ8wMQA:QYmmAAAAEAAA
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: diane on May 17, 2009, 11:55:27 PM
Spoiler
I have chosen to move the 5 from the 11 to 6, although I would change my mind if someone can take the time to convince me that 6-1 isn't awful.  I looked at that first, but really don't like moving that checker so far forward and leaving us, er is it called stripped?  It just seems that if we move that one, we have nothing left to do, wheras if we move 11-6, and dont get hit, we have options in the future  ;)  If we do get hit, well we can look to anchor up behind that little prime and play a different game.
[close]
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: vikingblood80 on May 18, 2009, 01:06:34 AM
Quote from: diane on May 17, 2009, 11:55:27 PM
Spoiler
is it called stripped?  It just seems that if we move that one, we have nothing left to do, wheras if we move 11-6, and dont get hit,
[close]

Spoiler
Yes its called stripped. The problem is that now almost all our points are stripped (especially after 6/1) thats why I also went for 11/6 with the 5  ;)
[close]
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: ah_clem on May 18, 2009, 02:28:46 AM
Spoiler

The one is forced.   Where to play the 5?  6-1 is safe, but does nothing to advance our game.  We're not under the gun yet, so we don't have to play super safe.

But 15/10 leaves two blots and we shouldn't play that loose either.

That leaves the choice between leaving a blot on the 13, subject to a direct hit with a 1 plus indirect shots of nine and eleven (17/36) or leaving a blot on the 11 subject to  an indirect hit of 7 or 9 (10/36) .  The shot count should be enough, but observe that 13/5 allows a devastating double hit with 6-1, and many hits after 11/6 require factorum to abandon the anchor.

So, bar/1 11/6 seems clear.

[close]
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: playBunny on May 18, 2009, 12:15:57 PM
Quote from: ah_clem on May 18, 2009, 02:28:46 AM
but observe that 13/5 allows a devastating double hit with 6-1

I can't see the 13/5 or any devastating double hit! Must put away these wine glasses and put on some seeing glasses. :D
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: playBunny on May 18, 2009, 12:24:47 PM
Spoiler
Mine's the 13/8. I like that it gives more flexibility than 11/6 and that would make a difference if we get away with that midpoint blot. What I'm not sure of is whether it's enough advantage to counter the extra hits. 11/6 exposes us to 30.6% (11 shots) versus 13/8 to 44.4% (16) and that's a fair increase.

I'm certainly not keen to do 6/1, though. GnuBg does tell me off occasionally for not playing deep-but-safe but I have something of an aversion to sending my men to the Wall. If the vote looks to be going the way of 6/1 (as it is at present) then I will change mine to 11/6.
[close]
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: playBunny on May 18, 2009, 12:30:43 PM
Spoiler
Oops, that's 50% (18/36 shots ) for the midpoint blot, which is even worse than I thought. Now it's looking like too high a penalty so I'm changing my vote to 11/6.
[close]

ps. Why can't we modify a post more than once? One typo fix and that's that?
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: blitzxz on May 18, 2009, 12:46:32 PM
(spoiler contains my thoughts about possible double also)

Spoiler
One point is my choice. Any hit here is a likely game or even match loser so ultrasafe and let's hope for a miracle. Cube is looking horrible but propably we have to take it any way (I don't like it at all) so now our too early double really backfires. With one point move there is also a bigger chance that opponent misses double because there are no immediate hits.
[close]

And if somebody missed my post in other topic: if votes are tied there'll be addition day for voting/changing votes and if still tie then the coin flip. Sorry about that other flip I just made too hasty decision. And let's not try to make ties.   :)
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: ah_clem on May 18, 2009, 01:16:19 PM
Quote from: playBunny on May 18, 2009, 12:15:57 PM
I can't see the 13/5 or any devastating double hit!

Right.  Here's what I meant to say:
observe that 13/5 13/8 allows a devastating double hit with 6-1.

With blots on 12 and 24, 6-1 allows 13/12* 6/1*. 
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: vikingblood80 on May 18, 2009, 01:18:43 PM
Quote from: ah_clem on May 18, 2009, 01:16:19 PM
With blots on 12 and 24, 6-1 allows 13/12* 6/1*. 

13/12* 7/1*
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: stog on May 18, 2009, 01:19:11 PM
Spoiler
i thought as blitz too -1point damage limitaTION OR POSSIBLE RECYCLE!
[close]
and np blitzxz you are doing great (r u messaging factotum when we have played or when he has new dice or cube to consider - everything seems to be going quite fast - so thx again  to both you and factotum)
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: lewscannon on May 18, 2009, 06:02:01 PM
Spoiler
I prefer the 11-6, avoiding the direct shot.
[close]
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: factotum on May 19, 2009, 07:28:56 PM
you guys are toast.
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: NIHILIST on May 19, 2009, 07:49:24 PM
A real man would redouble.

Bob
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: Zorba on May 19, 2009, 08:30:22 PM
Spoiler
11/6 leaves sevens and nines to hit (but 3-3 is blocked) , that's 10 shots, almost like a direct 1-shot. I don't think we can afford odds like that on getting a second checker sent back. Escaping that 4-prime is much easier with one checker than with two. Unfortunately, 6/1 is terribly ugly, so I'm not sure, but I voted for safe and hope factotum rolls something terrible, or we roll a miracle next turn.
[close]
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: lewscannon on May 19, 2009, 10:49:41 PM
Quote from: Zorba on May 19, 2009, 08:30:22 PM
Spoiler
11/6 leaves sevens and nines to hit (but 3-3 is blocked) , that's 10 shots, almost like a direct 1-shot. I don't think we can afford odds like that on getting a second checker sent back. Escaping that 4-prime is much easier with one checker than with two. Unfortunately, 6/1 is terribly ugly, so I'm not sure, but I voted for safe and hope factotum rolls something terrible, or we roll a miracle next turn.
[close]

Spoiler
10 hits out of 36 rolls? Easy decision. 28% chance of getting hit. You'll make a lot of money in life with those odds.
[close]
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: ah_clem on May 20, 2009, 02:40:29 AM
Quote from: factotum on May 19, 2009, 07:28:56 PM
you guys are toast.

Hey, the ball's in your court, dude.

Make my day.
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: playBunny on May 20, 2009, 02:51:11 AM
 
Rollout 1296 trials, 2-ply for 12 plies, 0-ply thereafter, 2-ply cube throughout

Spoiler
The two top moves are separated by less than the error value and so are pretty much equivalent.

# .. Move ......... Equity .. (Diff) .... Win ..... WinG .. WinBg -- Lose ... LoseG . LoseBg
1 .. bar/24 11/6 .... -0.946 ............ 40.4% .. 18.0% ... 1.8% -- 59.6% .. 17.3% ... 2.1%
2 .. bar/24 6/1 ..... -0.951 (-0.006) ... 40.7% .. 18.6% ... 0.9% -- 59.3% .. 18.6% ... 3.2%
3 .. bar/24 13/8 .... -0.991 (-0.045) ... 39.1% .. 16.2% ... 1.2% -- 60.9% .. 16.6% ... 2.7%
4 .. bar/24 15/10 ... -1.000 (-0.054) ... 36.7% .. 10.1% ... 0.3% -- 63.3% .. 20.8% ... 1.1%
5 .. bar/24 8/3 ..... -1.000 (-0.054) ... 35.3% ... 9.8% ... 0.3% -- 64.7% .. 22.3% ... 1.2%
Std Err .............. 0.009 ............. 0.2% ... 0.5% ... 0.2% --- 0.2% ... 0.2% ... 0.2%

Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 859164089 and quasi-random dice
Play: supremo 2-ply cubeful prune [world class]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
Cube: 2-ply cubeful prune [world class]
Different evaluations after 12 plies:
Play: 0-ply cubeful prune [expert]
Cube: 2-ply cubeful prune [world class]
[close]
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: Zorba on May 20, 2009, 03:47:18 PM
Quote from: lewscannon on May 19, 2009, 10:49:41 PM
10 hits out of 36 rolls? Easy decision. 28% chance of getting hit. You'll make a lot of money in life with those odds.

Not with the pay-off you get on this bet: getting hit costs you more than being missed gains you... :unhappy:
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: lewscannon on May 20, 2009, 06:22:48 PM
Quote from: Zorba on May 20, 2009, 03:47:18 PM
Not with the pay-off you get on this bet: getting hit costs you more than being missed gains you... :unhappy:

But you can only move what the dice give you, therefore it's an easy best option. If he gets his 28% hit, then you move to the next roll and the next decision. And when all else fails, it's time to access the dice generating program.  ;) I don't know if factotum knows yet that we can program all the rolls.
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: Zorba on May 21, 2009, 04:48:35 PM
Quote from: lewscannon on May 20, 2009, 06:22:48 PM
But you can only move what the dice give you, therefore it's an easy best option.

Not sure if I follow you here. I voted for 6/1, which does not leave a shot at all. That's a lot less risk than this 28%. So I disagree strongly that there's an easy best option. The downside from 6/1 is that the "gain" from never getting hit is also much lower, as it leaves a more awkward position than after 11/6.

playBunny's rollout has a too high error margin to tell which play is best, but it's enough to tell that both plays are close. I'm running a longer one at higher settings but that takes days to complete. To be continued.

BTW, have we all donated to blitzxz already?  B)
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: lewscannon on May 21, 2009, 06:20:25 PM
Quote from: Zorba on May 21, 2009, 04:48:35 PM
Not sure if I follow you here. I voted for 6/1, which does not leave a shot at all. That's a lot less risk than this 28%. So I disagree strongly that there's an easy best option. The downside from 6/1 is that the "gain" from never getting hit is also much lower, as it leaves a more awkward position than after 11/6.

playBunny's rollout has a too high error margin to tell which play is best, but it's enough to tell that both plays are close. I'm running a longer one at higher settings but that takes days to complete. To be continued.

BTW, have we all donated to blitzxz already?  B)

Scuse me Zorba, my bad. I thought your vote was to bring down one of the outfield pieces. The 6-1 doesn't do much for me either.
Title: Re: Match 1, Game 3, Move 8, Forum
Post by: Zorba on May 21, 2009, 07:32:45 PM
Ah got it, np! All makes more sense now  :laugh:
Title: ** ROLLOUT - Match 1, Game 3, Move 8, Forum
Post by: Zorba on May 22, 2009, 11:34:31 AM
Full 2-ply rollout:

Spoiler

The plays are very close, and given the standard errors it's too close to call which is best. The 6/1 play wins some more games, but it also loses more gammons.

The equities also point out that we have a small take, this time with good statistical significance: dropping would technically be a 0.0275 error after the 11/6 play.


        Full cubeful rollout with var.redn.
        2592 games, Mersenne Twister dice gen. with seed 857820087 and quasi-random dice
        Play:  2-ply cubeful prune [world class]
        keep the first 0 0-ply moves and up to 8 more moves within equity 0,12
        Skip pruning for 1-ply moves.
        Cube: 2-ply cubeful prune [world class]

    1. Rollout          bar/24 11/6                  Eq.:  -0,9725
        39,81  17,32   1,75 -  60,19  17,41   3,16 CL  -0,3193 CF  -0,9725
      [  0,15   0,42   0,18 -   0,15   0,18   0,26 CL   0,0032 CF   0,0070]
    2. Rollout          bar/24 6/1                   Eq.:  -0,9768 ( -0,0043)
        40,27  18,95   1,42 -  59,73  19,09   3,06 CL  -0,3213 CF  -0,9768
      [  0,15   0,43   0,17 -   0,15   0,18   0,23 CL   0,0034 CF   0,0075]
[close]