Open ThePokerDude opened 8 months ago
The neural network agree but simulation doesn't
The better the net, the more we should trust the net, but it is still to difficult to decide when - because a high score can be because we have few examples or because it is the right bid.
I won't do anything by this for now.
instead of 4s Ben should bid 4h
[Event ""] [Site ""] [Date ""] [Board "172"] [Dealer "W"] [Vulnerable "NS"] [Deal "W:KJ63.J.AK53.AQT2 T.Q96542.JT4.J94 Q982.A83.986.865 A754.KT7.Q72.K73"] [Scoring ""] [BCFlags "1f"] [Generator "BridgeComposer Version 5.104.1"] [South "WBridge5"] [West "ben0308p5dd"] [North "WBridge5"] [East "ben0308p5dd"] [Declarer "E"] [Contract "4S"] [Result "10"] {PAR of the deal: 4S = played by East : 420 points} {PAR computation time: 00'03"} {Feasability: 4 4 8 3 4 9 9 5 10 8 4 4 8 3 4 9 9 5 10 8 } [Auction "W"] 1D Pass 1S Pass 4S Pass Pass Pass [Play "S"] sa SK st S2 s4 S6 h6 S8 d2 DA d4 D8 h7 HJ h2 HA ht S3 h4 H8 d7 DK dt D6 dq D5 dj D9 c3 CQ c4 C6 s5 SJ h5 SQ s7 C2 h9 S9 c7 CA c9 C5 hk D3 cj H3 ck CT hq C8