Cube question
in a match to 5 with the score 0-0 my opponent offered a recube. i have a huge pip lead, but he has a good 4pt board, 4 builders pointing at my blot on the ace point, and controls the outfield. my own home board is crunched, and may crunch further. the position looked hopeless, and i declined. the bot had a different opinion, saw me winning 34% of the time, and slapped me with a 0.436 blunder. over the board, how do i get to a take, how do i estimate my winning chances?
2 Replies
First off, we need to know your take point. Dropping gives you 34% MWC; taking either 95 or 15. You risk 19 to gain 51, so your take point is 27%, close to money play.
The key is that with your race lead you are not out of it no matter what your opponent does. Consider groups of possible rolls for your opponent:
1. He makes the one point on your blot - 54, 53, 51, 43,41, 31, 11 - 13 rolls
We aren’t too great in this spot, but if we can roll a 5”3 and escape we likely will win a good many games. We probably wouldn’t take a double had he waited and double after these rolls, so we likely are under 25%, but not too far under (consider that 1/6 of the time we enter and escape; we win most of those). I’ll give us 2.5 wins in this scenario
2. He hits our blot but doesnÂ’t make a 5th point 61, 63, 64, 65, 66, 55, 44, 33, 52, 42, 32, 21, 22 - 21 rolls. Here we are better off. We are a favorite to reenter and when we do we are a favorite to win. Overall we are probably not a favorite, but we could probably take a double so I give us about 6 wins (30%)
3. Our blot is not hit - 62, 2 rolls. Obviously we are well off here and a strong favorite. Give us 1.5 wins.
Adding up we get about 10 wins out of 36 which is about 28% by my estimate. It looks like my estimates are a bit conservative since the bot says we actually win about 34%, but it got us to the right cube decision.
FWIW I might have botched this one too; I find I play a LOT better in hindsight 😀
thank you for your answer, that helps. i guess the lesson is you have to go though all the numbers and make estimates, even if it looks bad. thanks again.
