100% GIBug

[virgosrock]

[hv=http://www.bridgebase.com/tools/handviewer.html?lin=pn%7Cstefan_o%2CVirgosRock%2CRobotN%2CRobotE%7Cst%7C%7Cmd%7C1S268JQH79QKD48KCQ%2CS3479TH36AD3C257K%2CS5H48TD2567JQAC6A%2C%7Crh%7C%7Cah%7CBoard+3%7Csv%7Ce%7Cmb%7C1S%7Can%7CMajor+suit+opening+--+5%2B+%21S%3B+11-21+HCP%3B+%7Cmb%7Cp%7Cmb%7C1N%7Can%7CForcing+one+notrump+--+3-+%21S%3B+6%2B+HCP%3B+12-+total+points+%7Cmb%7Cp%7Cmb%7C2H%7Can%7CNew+suit+--+4%2B+%21H%3B+5%2B+%21S%3B+11%2B+HCP%3B+12-18%7Cmb%7Cp%7Cmb%7C3D%7Can%7CNew+suit%2C+bad+S+--+3-+%21H%3B+2-+%21S%3B+6%2B+HCP%3B+twice+rebiddable+%21D%3B+10-+total+points+%7Cmb%7Cp%7Cmb%7Cp%7Cmb%7Cp%7Cpg%7C%7Cpc%7CSA%7Cpc%7CS2%7Cpc%7CS3%7Cpc%7CS5%7Cpg%7C%7Cpc%7CH2%7Cpc%7CHK%7Cpc%7CHA%7Cpc%7CH4%7Cpg%7C%7Cpc%7CS4%7Cpc%7CDJ%7Cpc%7CSK%7Cpc%7CS6%7Cpg%7C%7Cpc%7CD2%7Cpc%7CD9%7Cpc%7CDK%7Cpc%7CD3%7Cpg%7C%7Cpc%7CCQ%7Cpc%7CCK%7Cpc%7CCA%7Cpc%7CC4%7Cpg%7C%7Cpc%7CDA%7Cpc%7CDT%7Cpc%7CD4%7Cpc%7CC2%7Cpg%7C%7Cpc%7CH8%7Cpc%7CH5%7Cpc%7CHQ%7Cpc%7CH6%7Cpg%7C%7Cpc%7CSQ%7Cpc%7CS7%7Cpc%7CC6%7Cpc%7CC3%7Cpg%7C%7Cpc%7CSJ%7Cpc%7CS9%7Cpc%7CHT%7Cpc%7CCJ%7Cpg%7C%7Cpc%7CS8%7Cpc%7CST%7Cpc%7CD6%7Cpc%7CCT%7Cpg%7C%7Cpc%7CD7%7Cpc%7CC9%7Cpc%7CD8%7Cpc%7CC5%7Cpg%7C%7Cpc%7CH9%7Cpc%7CH3%7Cpc%7CDQ%7Cpc%7CHJ%7Cpg%7C%7Cpc%7CD5%7Cpc%7CC8%7Cpc%7CH7%7Cpc%7CC7%7Cpg%7C%7C%0A%0A]399|300[/hv]

 

vrock

[steve2005]

What's the bug?

Looks more like Gib can't evaluate it's hand.

[Stephen Tu]

There are several problems here that contribute to this sequence, that they ought to tighten up IMO:

 

1. It seems that the bot wants 14+ total points to 2/1. With the Gib walrus TP counting essentially this means it needs to turn the T of hearts into the J for it to bid 2d. 14+ seems like a lot to me, 13+ ought to be enough if 11+HCP. Unless one is playing some like super light opening system or whatever.

 

2. It doesn't upgrade for the 7th card in a good suit, I've long complained it doesn't upgrade for good long suits enough. If I had my way the bot would stop counting shortness in partner's suit, and maybe shortness in general before a fit is found, but count more for good long 6-8 cd suits. This might qualify North even if still requiring 14+ for the 2/1.

 

3. GIB plays strong jump shifts for 1s-3d. This is a big problem because now 1s-1nt-2M-3d has to cover such a huge range, essentially 6-13 total points since 2d requires 14+ TP. But they have defined it as 10- total points. So now basically 11-13 total points has fallen into a black hole. There are basically 2 ways to fix this:

- get rid of strong jump jump shifts, play invitational jump shifts, so 1s-3d is like some approximation of 11-12 total points, 6+d, 2-s, 3-h. Hands with x46x are still awkward, prob still have to bid 1nt, and guess later, but at least hands of original post's shape are covered.

-play 1s-2d-2x-3d as invitational (old style Aces/Eastern scientific / old Lawrence 2/1)

 

I'm strongly in favor of using the intermediate jump shift approach. The strong jump shift to the 3 level is just really space consuming and not nearly as indispensable when 2/1 is GF. And the 2/1 rebid inv has fallen off in popularity and is probably harder to program into the bots as to when you can pass below game after a 2/1.

[steve2005]

I agree intermediate js are an improvement.

Been trying to convince partners one by one of this is hard.

Everyone has fallen in love with Bergen so use 3m for that, which is anther story!

[virgosrock]

What's the bug?

Looks more like Gib can't evaluate it's hand.

 

Bug is easy. Can't show drop bid with 11 hcp and 7 card suit in such a sequence. Bid 2 NT if you have to.

 

vrock