15 point match
Green-White:
Score 8-10
|
|
Pip: 143 | |
| Game 1 15 point match Green-White: Score 8-10 |
||
| Pip: 182 |
| # | Ply | Move | Equity |
|---|---|---|---|
| 1 | R | bar/23 8/5 | -0,737 |
| 95% confidence interval: - money cubeless eq.: -0,537 ±0,017. Rollout settings: Truncated rollout, depth 11, 144 games (equiv. 18158 games), played 2-ply (precise), random seed, with race database. | |||
| 2 | R | bar/22 7/5 | -0,901 (-0,163) |
| 95% confidence interval: - money cubeless eq.: -0,593 ±0,018. Rollout settings: Truncated rollout, depth 11, 144 games (equiv. 16395 games), played 2-ply (precise), random seed, with race database. | |||
| 3 | R | bar/22 6/4* | -0,969 (-0,232) |
| 95% confidence interval: - money cubeless eq.: -0,626 ±0,015. Rollout settings: Truncated rollout, depth 11, 144 games (equiv. 23239 games), played 2-ply (precise), random seed, with race database. | |||
| 4 | R | bar/22 13/11 | -1,000 (-0,263) |
| 95% confidence interval: - money cubeless eq.: -0,705 ±0,018. Rollout settings: Truncated rollout, depth 11, 144 games (equiv. 17061 games), played 2-ply (precise), random seed, with race database. | |||
| 5 | R | bar/22 8/6 | -1,000 (-0,263) |
| 95% confidence interval: - money cubeless eq.: -0,665 ±0,017. Rollout settings: Truncated rollout, depth 11, 144 games (equiv. 18294 games), played 2-ply (precise), random seed, with race database. |
| Cube action |
|
If one moves White's three spares on the 6 point back to the midpoint, forming position b [above], the strategic features change radically. Now B/23 8/5, which was a mega-blunder in position a [the position in the initial problem], becomes by far the best move -- and the best move in position a, B/22 6/4 *, becomes here a big blunder. It's not difficult to guess why: In position b, Green has 'gained' in race terms but in fact has lost in timing terms 3x7=21 pips. This means that, only 34 pips down after completing his roll, he can't restore his timing for a backgame by any means. Now his main plan should be to try to contain White's back checker. Excluding jokers, Green's back checkers are not in big danger as in position a, when they were desirable and easy targets for the spares on White's 6 point, and can be used to make the 22 point or even White's bar point with a sound game. Interestingly enough, Extreme Gammon [3 ply with Variance Reduction Rollout, 1296 games] more or less agrees with Snowie's analysis in both positions a and b. I believe that this problem would be a relatively easy one for good players in the 70's and 80's, when lots of backgames were being played and more difficult for today's good players, who don't have as much experience in backgames due to the style of modern bots. |