I was a bit surprised when I first looked at this on Snowie 3 ply. Snowie 3 ply has all 3 plays within' .01 of each other and making the 4 point as best. I personally thought that splitting was the only choice.
First of all, let's rule out making the 3 pt. as making the stronger pt. in this situation should be best. [the distribution left after making the 3 instead of the 4 is nothing to write home about, they both leave the same shot, so it reasons the 4 is better than the 3]
After that I cut out (and didn't roll out) 24/20 8/5. Granted, my fours are somewhat duplicated, but it's too loose. If you're going to make any split it should be 24/20 13/10, not one where you're going to be trying to come in on a stronger board than yours, hitting loose, either directly or indirectly, and then also magically covering your own 5 pt.
Now we're down to 3 plays, making the 4, the big split, and attempting a counteroffensive by slotting your own 5 and bringing a builder down from the mid. Making the 4 pt. add sixes to my good numbers, not that my sixes weren't ok to begin with, but I'd much rather hit than make the bar or 3 pt. with my sixes. It gives me too many good numbers and leaves your position stacked + stripped. The counteroffensive is also weak in my opinion because I have such a huge head start on you and your position is lackluster to keep making inner board points. You're ahead in the race and need to extricate those back men, the sooner the better before I have a Ted Nugent-esque stranglehold on them.
The split to me is a "if you aren't going to split now when in the f&*# do you think you'll be able to split!?" That's the best argument I can put forth. My offensive position will strengthen, I simply have too many good numbers, and I'll be making (hopefully) the 5 pt. or if not, the 3 or bar. Then what? I have one man I need to escape, you have two, I am behind in the race [which favors me in a priming battle], and as I said, I have a big head start on a priming battle anyway, the best offensive in this question is a good defense. Lower your helmets and hope I don't splat you on the 5 pt.
The Readers split barely won the voting with 13/9 8/5 coming in a close second.
| 143 |                          |
| 155 | |
| Board image courtesy of GO-Figure | |
| # | Ply | Move | Equity | |
| 1 | R | 24/20 13/10 | -0.547 | |
| 0.7% 9.9% 37.0% 63.0% 23.4% 1.3% | ||||
| Live cube rollout: -0.547 | ||||
| 95% confidence interval: - money cubeless eq.: -0.402 ±0.015, - live cube: -0.547 ±0.027. Rollout settings: Full rollout, 972 games (equiv. 28214 games), played 3-ply (precise), cube 3-ply, settlement 0.550 at 16 pts, seed 1, with race database. | ||||
| 2 | R | 13/9 8/5 | -0.588 (-0.041) | |
| 0.5% 9.4% 36.6% 63.4% 21.3% 2.0% | ||||
| Live cube rollout: -0.602 | ||||
| 95% confidence interval: - money cubeless eq.: -0.402 ±0.016, - live cube: -0.602 ±0.028. Rollout settings: Full rollout, 972 games (equiv. 27035 games), played 3-ply (precise), cube 3-ply, settlement 0.550 at 16 pts, seed 1, with race database. | ||||
| 3 | R | 8/4 7/4 | -0.606 (-0.059) | |
| 0.4% 10.0% 35.1% 64.9% 19.5% 1.5% | ||||
| Live cube rollout: -0.635 | ||||
| 95% confidence interval: - money cubeless eq.: -0.405 ±0.014, - live cube: -0.635 ±0.029. Rollout settings: Full rollout, 972 games (equiv. 32275 games), played 3-ply (precise), cube 3-ply, settlement 0.550 at 16 pts, seed 1, with race database. | ||||