I know the bots don't like penalty doubles

[BunnyGo]

But with an Ace and Qx of trump...it can't sit for one at the 5 level?

 

 

http://tinyurl.com/6sfbyob

[barmar]

I haven't been able to reproduce this. The book bid is Pass. It considers 5♥ as an alternative, but in all my simulation runs it preferred Pass.

 

Was this a basic or advanced robot?

[BunnyGo]

It was an ACBL robot tourney. It happened at least one other time. Here's the traveler:

 

http://www.bridgebase.com/myhands/hands.php?traveller=9657-1321116601-51205548&username=bunnygo

[BunnyGo]

In fact, everyone who doubled 5♣ had it pulled. One person managed to hit 4♣.

[barmar]

That helps. I was doing my tests in IMP mode. When I switched to MP, 4 out of 11 simulation runs bid 5♥.

 

I think there's two things going on:

 

1) The bidding seems confused it. None of the hands it dealt in the simulations match it well. In most cases, West doesn't have hearts stopped. And in almost all hands, South has a diamond void; the few exceptions have a club void. So the robot thinks it's probably making 5♥, possible even doubled.

 

2) If you're making, the vulnerability makes declaring better than defending at matchpoint. You need to set them 3 to do better.

 

I think the root cause is that the last double shows 19+ total points. With all the bidding that's going on, the only way for it to make sense of it is that there's wild distribution. The comment in the bidding database above this rule says:

 

; With no forcing P available, if we don't have enough offense to bid

; on, must X just to show our GF-values.

 

That may be appropriate below game, but it shouldn't apply at the 5 level, or if we've already denied extra strength by passing.

[barmar]

Another bug in this auction is West's 4♣ bid. It looks like the rule for retreating from 3♠ is picking any suit with a partial fit, rather than the best fit.