                        |
| ||||
| XGID=-a---aE-CA--eD---cAd-a---A:0:0:1:21:0:0:3:0:10 | |||||
| Blue to play 21 | |||||
| 1. | Rollout1 | Bar/23 6/5* | eq: -0,080 | |||
| ||||||
| 2. | Rollout1 | Bar/24 18/16 | eq: -0,130 (-0,050) | |||
| ||||||
| 1 1296 Games rolled with Variance Reduction. Moves: 3-ply, cube decisions: XG Roller | ||||||
eXtreme Gammon Version: 2.10
If we are searching for a metapher, a phenomen, perhaps the difference of ‚Anatomy/Physiology’ can help. There is the given anatomy of the board, the actual move to play and the inherent virtuality, to develop. There are restrictions of the anatomy, which are not really to overcome, later in the game. Danny Kleinman worked on this, I remember the ‘prime field’ and the ‘evil prime forest’. Taking this into account, the 5 men on the 6-point are crippled men, in the beginning. Their freedom to move, given the meta-concept ‘efficiency’, is limited due to the sub-concepts ‘flexibility’ and purity’. The concept ‘purity’ ends for them at the 3-point, this is the coastline, where suddenly the ‘evil prime forest’ appears. If you’ve made the 2-point, you have to break the 8-point, to rebuild the prime, given in the skeleton of the beginning, containg two already points made, 8 and 6, there is the 3 for the edge of this prime. Time-consuming, to rebuild. And the ace-point is the graveyard of purity and prime value. The ace-point is valuable only for Blitz/Attack Value. So given the anatomy, numbers for the 6-points ends with the 3, crippling their value of freedom in a ‘pure’ sense to three numbers.
Part of the great book from Nack Ballard and Paul Weaver were the insights, coming into value, to see the given and the virtual at once, for the actual move to play. The anatomy is influencing the physiology and vice-versa.
So hitting with the ace from the 6-point addresses this heavy stacked point, which has almost no freedom for reaching the purest point, instead with an ace. This is an argument, to put more weight on the need, to hit here. Given his limited freedom to develop, unloading on the most valuable point to reach, it should be a strong argument, despite the price, you have to pay for the hit.
The ‘neural net’ confirmes the considerations of the ‘human neural net’. It can be regarded as an expression of correlation between the two ways of thinking. We just don’t know, how the bot in his black box is coming up with his move. But his reinforcement learning gave him the ‘right’ answer. He also backed up an ancient view, the beginning of backgammon theory, Danny Kleinman. It could be, that the new framework, if there is something new in this, it will give a restart, to rethink backgammon theory, in the way of a ‘human neural net’.