[steve2005]

My link
Gib is notorious for not leading suit indicated by partner
Here 2♣ Stayman is doubled to show clubs
On lead there isn't an obvious lead from the cards held.
But is obvious to lead partner's suit against NT
Of course the only lead which defeats 3N double dummy is a club

[smerriman]

Are you suggesting robots should be forced to lead partner's suit? I don't think that's a good idea. If a simulation (which takes into account what you've shown in the bidding) tells it that another suit is better, then it should lead accordingly.

Of course, it often simulates poorly, but that's a more widespread and unrelated issue.

[johnu]

smerriman, on 2022-March-17, 16:01, said:

Are you suggesting robots should be forced to lead partner's suit? I don't think that's a good idea. If a simulation (which takes into account what you've shown in the bidding) tells it that another suit is better, then it should lead accordingly.

Of course, it often simulates poorly, but that's a more widespread and unrelated issue.

Obviously if GIB East wanted a club lead it should have bid diamonds (or a major). Clearly leading from a broken J fourth through dummy's known 4 card suit Smolen hand is much better than leading from 3 small to partner's lead directing double suit. Do we really need a simulation to tell us that?

[steve2005]

smerriman, on 2022-March-17, 16:01, said:

Are you suggesting robots should be forced to lead partner's suit? I don't think that's a good idea. If a simulation (which takes into account what you've shown in the bidding) tells it that another suit is better, then it should lead accordingly.

Of course, it often simulates poorly, but that's a more widespread and unrelated issue.

no trump defence is about setting up your longest suit.
We have at least 8 clubs, odds are clubs is our longest suit
I would wager any simulation would show leading a club is best. or at least equal to the best lead.
Not sure if the robot is simulating here it was just a tcr 80 game. Do they use the cheap robots?

[smerriman]

steve2005, on 2022-March-17, 20:37, said:

I would wager any simulation would show leading a club is best. or at least equal to the best lead.

The latter is sometimes the problem; if it only uses a small sample and they're all exact equal, then it wouldn't know to lead a club. But I calculate about a 58% chance of a spade being strictly worse, so even a sample of 10 would result in a club 99.99% of the time. So yeah, can't even blame a poor simulation here.

steve2005, on 2022-March-17, 20:37, said:

Not sure if the robot is simulating here it was just a tcr 80 game. Do they use the cheap robots?

Robots always simulate during the play, it's the only way they know what card to play. (Cheap robots don't simulate during the *bidding*). Though the only explanation for some of these hands is that it failed to simulate and ended up playing a card at random, or something along those lines. Nobody knows..