[mikl_plkcc]

---

The robot refused to set the contract with the ♣A, and let them made the doubled game!

[jillybean]

Why torture yourself with bots?

[johnu]

Good defense by GIB to hold the contract to making 3. Lesser robots would have allowed an overtrick.

[smerriman]

There really should be a forum for GIB posting about all of the countless things its human partners do wrong.

[msheald]

Hello! For my own education, I'd be curious to know ops change in position should Gib have taken the setting trick. I have not looked at the hand closely, but it is possible that taking the ace would have opened up other lines of play. In the event that it would have been the setting trick, knowing the point change between the two outcomes would be helpful in knowing how to play with Gib in the future.

Lastly, it would be helpful for those with program simulators to run 5000 deals to see what average results would be with different lines of play. As David Bird notes in "Winning Duplicate Tactics," a single deal means nothing, but the good player plays to maximize outcome over many deals. Best regards.

Mike

[msheald]

No One?

[smerriman]

msheald, on 2025-July-17, 06:30, said:

Hello! For my own education, I'd be curious to know ops change in position should Gib have taken the setting trick. I have not looked at the hand closely, but it is possible that taking the ace would have opened up other lines of play. In the event that it would have been the setting trick, knowing the point change between the two outcomes would be helpful in knowing how to play with Gib in the future.

Lastly, it would be helpful for those with program simulators to run 5000 deals to see what average results would be with different lines of play. As David Bird notes in "Winning Duplicate Tactics," a single deal means nothing, but the good player plays to maximize outcome over many deals. Best regards.

Mike

GIB is not capable of deciding what tables other contracts are likely to be playing in (or how their play would have gone). Here, obviously, you'd expect all other humans to be playing 4♥. If that makes, anything but an impossible down 3 is a bottom. If it's going down, all that matters is that you beat 3♦; it makes no difference by how many tricks, even at MPs where the difference of a single trick is usually crucial to the score.

GIB instead assumes everyone else is in the same 3♦x contract it is, so if there's a line which gets down 2 52% of the time and lets 3♦ make 48% of the time, it considers that better in the long run than a guaranteed down 1.

It also doesn't understand South's double (though who does), or carding or declarer play logic, so thinks South is a slight favorite to hold the ♣J, despite the fact that leading the ten and declarer not covering would both be very unusual, among other inferences from the play.

And if South holds the ♣J, ducking beats 3♦ by more.