I happily tripped over this position giving lessons the other day. I had tried to find it before to no avail to use for the 4a 3a score and for the story but had been unable to hunt it down. Upon seeing it there was no doubt where it was from.
I was playing another Giant, I still can't recall who. I'm sure it wasn't Falafel and have no idea if he ever did comment on this position in real life. I'm positive it wasn't Michy either because I ended up betting a fair amount against the person I was playing and Jake on the correct cube action. OtB it went D/T. Afterwards, I don't remember the exact conversation, but I let it be known I thought it was a dead drop and yada yada yada, bet made.
I like this problem so much not because of the fond memory of Jake saying something along the lines of "The best way to tell me the answer tomorrow is by handing me the money" and then having him do just that the next day or that I get to tell the story like 3 years later but this problem should be the biggest next in the world to Giants and should be very confusing to your average player.
What you need to know going into this problem is a lot. There is more than one way to skin a cat and sure, we can all play by 'feel', but I 'feel' making estimates especially when they're this easy is the right way to go. What we need to know going into the problem:
- The 3a player's take point is ~21.5%.
- The gammon values for 4a 3a are 1.0 and .50 respectively. (for you picky people, yes, I know they're .968 and .449 but since they both round in the same direction and make much more usable OtB numbers those are the ones I use)
- How to estimate wins. (reference positions)
- How to estimate gammons. (reference positions/early game study)
- How to do the math OtB
So an odd point of interest that I knew from all my early game study is that when we get to the fourth roll there is a big slew of positions where the second roll was a 66 reply played 18(2) 7(2) and we are at least considering a cube the win breakdown always hovers around 65/35. Yes, always. If you're considering a cube for money and it has that possible feel to it, you're safe assuming the win breakdown is in that ballpark. So going in to this problem I felt the utmost confidence assigning the breakdown was 65/35.
Now, on to the gammons. For early game blitzes where it's DT for money the doubler's gammons fall between 22-33% and the taker's between 5-10%. The taker's gammons are more likely to be around 8% in the early game because there's so much game left to play it's hard for them to fall much lower. Knowing all this I estimated a 25/8 breakdown for the gammons.
The math. Take point of 21.5%. Four away's gammons at a value of 1.0 is 25 * 1 or 25%. Three away's gammons having only half that value is 8 * .50 or 4%. We subtract to get the difference to add to our take point, 25-4 = 21%. We add that to our original take point of 21.5% + 21 = 42.5% as our fully adjusted take point. We already knew this position was a 65/35 position so even if that estimate is off a bit there's no way in hades this can be a take. Pay dat man his money.
                          |
| ||||
| XGID=-aa---ECC-A-dA---bBdab----:0:0:1:00:7:8:0:11:10 | |||||
| Blue on roll, cube action? | |||||
| Analyzed in Rollout | No double | Double/Take |
| Player Winning Chances: | 65.85% (G:23.14% B:1.06%) | 66.08% (G:23.63% B:1.63%) |
| Opponent Winning Chances: | 34.15% (G:8.05% B:0.52%) | 33.92% (G:8.18% B:0.97%) |
| Cubeless Equities | +0.458 | +1.457 |
| Cubeful Equities | ||
| No double: | +0.865 (-0.135) | ±0.016 (+0.849..+0.881) |
| Double/Take: | +1.290 (+0.290) | ±0.029 (+1.261..+1.318) |
| Double/Pass: | +1.000 | |
| Best Cube action: Double / Pass | ||
| Rollout details | ||
| 648 Games rolled with Variance Reduction. Dice Seed: 57539203 Moves: 3-ply, cube decisions: XG Roller | ||
| Double Decision confidence: | 100.0% | |
| Take Decision confidence: | 100.0% | |
| Duration: 9 minutes 40 seconds | ||
eXtreme Gammon Version: 2.10, MET: Kazaross XG2