[pescetom]

It's full moon and so I give all at table some rope... but still.

My link

I concede that my strong hand 2NT and the 4♠ bid may not be to everyone's taste or on the (poorly documented) radar of GiB, and that opponent was optimistic or nuts in passing my Double, but what was his GiB partner thinking when he did the same?
Both robots are fully paid "Advanced" if that matters.

[johnu]

Well declared by GIB East. A lesser bot might have gone down 6 tricks.

[Huibertus]

So, after you have played ♠ Ace and King you are 100% sure that IF declarer has ♠ J he has 6 ♣ and ♠ J for 7 tricks max. So you cash all your entries just to make sure you can not hold him to the 4 tricks of which you are sure in case partner holds ♠ Jack as a pleasant surprise and then you make fun of the things bots and opponents did wrong?

Yeah, right.

[pescetom]

Huibertus, on 2025-August-10, 00:38, said:

So, after you have played ♠ Ace and King you are 100% sure that IF declarer has ♠ J he has 6 ♣ and ♠ J for 7 tricks max. So you cash all your entries just to make sure you can not hold him to the 4 tricks of which you are sure in case partner holds ♠ Jack as a pleasant surprise and then you make fun of the things bots and opponents did wrong?

Yeah, right.

Opponent was neither named nor ridiculed.
This forum is about things bots may have done wrong.

If your point is that the hand could have been played seriously and at least with a minimum of concentration to pick up all the spades, sure. I was too busy laughing about the final contract.
I also (already) pleaded guilty to a dubious 2NT if for that, Alvin Roth would be rolling in his grave but then he played with Tobias Stone, not GiB.
The question here is "why did East robot pass?".

[smerriman]

pescetom, on 2025-August-10, 15:11, said:

The question here is "why did East robot pass?".

The sequence is far too messed up / varies too much from the old version of GIB to know.

GIB appears to play a non-keycard ROPI over the double (despite it being basically undefined), but for the current version:

- redouble is alerted as "DOPI-none", and a grand total of 0 aces combined is apparently enough for it to leap to 6♣ slam
- pass is "DOPI-one", which it passes
- 5♣, 5♦, and 5♥ are DOPI-two/three/four, but it apparently 2 aces is even better than 0 because over all three of these bids it follows up with 5NT querying kings, confirming all keycards

Those are the database definitions / basic GIB bids, but with your 4♠ showing 25+ points it's going to struggle to overrule these with simulations. Old GIB has different followups so something has clearly changed here, but expects if it bids again after partner has shown 1 ace, you'll almost always end up defending a slam doubled, often making.. so there's that too.