Assumption #1
- If my opponent closes me out and ends up with a perfect distribution I will win somewhere between 2% to 2.5% of the time. (also rolled out with standard 3 ply giving the exact same win percentage of 97.7%)
      
    
         
 |  is Player 2
score: 0
pip: 45 | | 1 point match | pip: 60
score: 0
 is Player 1 |
|
| XGID=aBBBCCB--A-------------fh-:0:0:1:12:0:0:0:1:10 |
| to play 12 |
| 1. | Rollout1 | 9/6 | eq: +0.954 |
| Player:
Opponent: | 97.72% (G:4.85% B:0.06%)
2.28% (G:0.00% B:0.00%) | Conf: ± 0.001 (+0.953...+0.955)
Duration: 54.5 seconds |
|
| |
1 1296 Games rolled with Variance Reduction.
Dice Seed: 36438452
Moves: 5 ply, cube decisions: 4 ply
|
Assumption #2
- Given my opponent's current structure from the original problem it is highly unlikely he will achieve a perfect bearoff structure. Any normal distribution issues will cost him between 1% to 1.5% assuming these flaws do not have the opportunity to leave immediate shots. A somewhat standard example as I see it: (also rolled out 3 ply and yielded the same results)
| is Player 2
score: 0
pip: 45 | | 1 point match | pip: 62
score: 0
is Player 1 |
|
| XGID=aBBBCCB----A-----------fh-:0:0:1:61:0:0:0:1:10 |
| to play 61 |
| 1. | Rollout1 | 11/5 4/3 | eq: +0.929 |
| Player:
Opponent: | 96.46% (G:4.04% B:0.05%)
3.54% (G:0.00% B:0.00%) | Conf: ± 0.001 (+0.928...+0.930)
Duration: 45.9 seconds |
|
| |
1 1296 Games rolled with Variance Reduction.
Dice Seed: 36438452
Moves: 5 ply, cube decisions: 4 ply
|
Assumption #3
- From the original now posted below (assuming RT + autorewhip) the fact is the prime isn't rolled yet and White's checkers are so far advanced there may be hiccups in bringing it home. I estimated this is worth at least 1% but admit it's the part of the equation most prone to being lowballed. An immediate 55 risks the 61 or even an entering ace and then the inability to remake the bar or a shot. 44 could cause trouble if the front checker is hit and even 66 needs 'fixed' to put checkers where they belong. There are other rolls such as 63 or 54 or 53 that don't leave the gin position you're looking for.
  
| is Player 2
score: 0
pip: 42 | | 1 point match | pip: 104
score: 0
is Player 1 |
|
| XGID=---aBBBCBCA------------fh-:0:0:1:00:0:0:0:1:10 |
| |
So for sure we've worked our way up to at least 4.5% in the eyes of the person trailing -1 -8 if they pass in our original position. As I said in my first analysis, I don't feel this number could be too high, it could be too low though. (originally I had 5% at this point) The take point depending on MET of choice, either G11 or RK, is 6.5% or 6.75%. We are hovering around 5% in my analysis, perhaps a shade more. Now the question becomes if you are the weaker player how much weaker do you need to be to take this? First, let's look at the actual rollout of the position.
 
 | is Player 2
score: 0
pip: 46 | | 1 point match | pip: 104
score: 0
is Player 1 |
|
| XGID=---aBBBCBCA--------a---eh-:0:0:-1:42:0:0:0:1:10 |
| to play 42 |
| 1. | Rollout1 | 6/2 | eq: -0.883 |
| Player:
Opponent: | 5.83% (G:0.40% B:0.00%)
94.17% (G:4.27% B:0.08%) | Conf: ± 0.002 (-0.885...-0.881)
Duration: 4 minutes 12 seconds |
|
| |
1 1296 Games rolled with Variance Reduction.
Dice Seed: 36438452
Moves and cube decisions: 4 ply
|
So this rolls out to a whopping 5.8% for the person being cubed! A significant amount higher than my original estimates for such a decision. I also rolled this out 3 ply, 3 ply huge, and 5 ply and GNU 2 ply. Both 3 plys, 4 ply, and GNU 2 ply gave the same 5.8% for the trailer whereas 5 ply brought it down to 5.7%. We can imagine now that given even the weakest of elo difference in players, let's say 50 elo, the decision to take or pass and put the match on the line is quite close. If you widen the gap to 100 elo the take will be clear.
Stick