The reason I posted this is I often speak of simplification in these and other types of positions when I'm trying to make estimates. The more often they come up the better I recognize them and bank that a certain method works or does not. Here you can do all the math as some of you did and it's not that hard for this particular problem but there's really no need.
You count not getting hit as a win, you count getting hit as a loss so you win 16/36 or ~44% after running. Don't tell me there's no way of knowing when this will and won't work, it works a lot. The only way for someone to use it though is to constantly test it - similar to estimating recube vig by reasoning instead of strict mathematics. If you practice at it you will find a comfort zone. I will try to post more of these as I come across them.
                         |
| ||||
| XGID=-BBAACB-A---cBa-ab-bbbAb--:0:0:1:43:0:0:0:1:10 | |||||
| Blue to play 43 | |||||
| 1. | Rollout1 | 22/15 | eq: -0.122 | |||
| ||||||
| 1 1296 Games rolled with Variance Reduction. Dice Seed: 95629089 Moves: 3-ply, cube decisions: XG Roller | ||||||
Stick