Introducing 3 card suits as "twice rebiddable"

[Leo LaSota]

On ACBL tournament #66, hand #1, the auction went P-P-1C-1S-P-2D-P-P-2H**, which was a "twice rebiddable" suit on AQ109xx 109x Kx Qx.

[Bbradley62]

Instructions for inserting a hand diagram to show the problem:

 

http://www.bridgebase.com/forums/topic/44979-how-do-i-upload-a-lin-file-how-to-insert-a-hand-diagram/

 

[hv=lin=pn|Leo LaSota,~~M5958,~~M5956,~~M5957|st%7C%7Cmd%7C3S47H6JAD3TQC267JA%2CS368JKH4QD246C4TK%2CS259TQAH39TD7KC9Q%2C%7Crh%7C%7Cah%7CBoard%201%7Csv%7Co%7Cmb%7Cp%7Cmb%7Cp%7Cmb%7C1C%7Can%7CMinor%20suit%20opening%20--%203%2B%20C%3B%2011-21%20HCP%3B%2012-22%20total%20points%7Cmb%7C1S%7Can%7COne-level%20overcall%20--%205%2B%20S%3B%208-17%20HCP%3B%209-19%20total%20points%20%7Cmb%7Cp%7Cmb%7C2D%7Can%7C5%2B%20D%3B%202-%20S%3B%2011-%20HCP%3B%209-12%20total%20points%20%7Cmb%7Cp%7Cmb%7Cp%7Cmb%7C2H%7Can%7C4%2B%20H%3B%2011-%20HCP%3B%208-12%20total%20points%20%7Cmb%7Cp%7Cmb%7Cp%7Cmb%7Cp%7Cpc%7CC3%7Cpc%7CC2%7Cpc%7CCK%7Cpc%7CC9%7Cpc%7CD6%7Cpc%7CD7%7Cpc%7CDA%7Cpc%7CD3%7Cpc%7CD9%7Cpc%7CDT%7Cpc%7CD4%7Cpc%7CDK%7Cpc%7CH9%7Cpc%7CH2%7Cpc%7CH6%7Cpc%7CHQ%7Cpc%7CSJ%7Cpc%7CSQ%7Cpc%7CH7%7Cpc%7CS4%7Cpc%7CDJ%7Cpc%7CDQ%7Cpc%7CD2%7Cpc%7CCQ%7Cpc%7CCA%7Cpc%7CCT%7Cpc%7CS2%7Cpc%7CC8%7Cpc%7CCJ%7Cpc%7CC4%7Cpc%7CS5%7Cpc%7CC5%7Cpc%7CS7%7Cpc%7CS6%7Cpc%7CS9%7Cpc%7CH8%7Cpc%7CD8%7Cpc%7CC6%7Cpc%7CS8%7Cpc%7CH3%7Cpc%7CSA%7Cpc%7CH5%7Cpc%7CHJ%7Cpc%7CS3%7Cpc%7CHA%7Cpc%7CH4%7Cpc%7CHT%7Cpc%7CHK%7Cpc%7CC7%7Cpc%7CSK%7Cpc%7CST%7Cpc%7CD5%7C]360|270[/hv]Where does it say anything about "twice rebiddable"?

[Leo LaSota]

I posted this hand after I finished the 18 boards and making sure that everyone else had completed the board. I was unable to see what the description of the bid was wehn I posted this before the tournament was complete. Nevertheless, if the bid is supposed to show 4+ hearts, I do not think that 109x qualifies for this explanation.

[nathan2008]

wowwwwww..... u start to complain about GIB lol. Interesting. Don't think it is the first time u have met this kind of bidding..... But frankly speaking, if GIB had 4H, he would X 1s, so this 2h must be 3 cards, he just wants to play a little hehe.

[nathan2008]

As u are the king of roboters, maybe u have noticed what i've said :).

[Stephen Tu]

Nevertheless, if the bid is supposed to show 4+ hearts, I do not think that 109x qualifies for this explanation.

 

In general, in situations where GIB is allowed to simulate, it's allowed to bid things within 1 card of description. So if 2h shows 4, it thinks it's OK to bid on 3. On this hand it thinks it's too strong to pass, it has to do something, but apparently double is defined wrong, probably it shows diamond length instead of strength + *spade* length. And for whatever reason it doesn't like 2nt either (or 3nt for that matter). So it can't find anything else better, and bids 2h.

 

Now, I don't think we really want GIB to be able to bid on 4 cd hearts here; if it had that hand it should have neg-x earlier. So 2h should probably be fixed to show at least 5, then at least it will not bid 3 cd suits here though it might bid on 4. But really fixing 2nt and double should give it better bid options.

 

Edit: I checked on my home GIB (old database, but some possibility of still matching current BBO GIBs), and it thinks 2nt is unusual, showing 5+ clubs and 5+ hearts! 3nt also shows unusual 3nt. This should definitely be fixed. 3nt should never be unusual IMO. And 2nt unu should only apply when partner has previously been silent.

[Leo LaSota]

Instead of balancing with 2H on a 3 card suit, the robot should be programmed to double, which shows maximum passed hand values (good 9 to 11), plus a spade stack. In general, penalty doubles of low levels should be based on strength in suit bid on your right. In other words, penalty double here should not show diamond penalties, it should show spade penalties. I would convert the double and take likely down 2.

[barmar]

That seems to be the problem, GIB plays double in this situation as diamond penalty, not spade penalty.

 

As a result, the book bid is Pass, and uses simulations to choose between this and 2♥. They're pretty close, and 4 out of 10 MP simulations choose 2♥.