[svengolly]

Why on this bidding and with this hand would the GIB lead the 3♣? Just about anything else holds the contract to 3 including the normal lead of a spade. The club lead gives them an extra trick.

Hopefully the screenshot shows up.

[1eyedjack]

According to a double dummy analysis, any lead concedes 10 tricks except for a Heart, which concedes 11.

Furthermore, on a superficial examination, a Club lead is the only one that does not immediately concede a trick:

A spade lead away from the Queen through Axx up to KJx provides declarer with a trick with the J, where left to his own devices a straight forward finesse of the J of Spades would lose to the Q

The same applies to the 10 of Hearts.

The Diamond 10 finesse is always working for declarer but he does not know that unless you lead a Diamond up to KTx, thus ensuring that he makes the 10 without having to worry about whether it is safe to finesse it.

A Club lead is the only one that forces declarer to work for his living.

We may find the logic of the lead to be a bit of a mystery, but there is no denying that on the lie of the cards it is the best chance for the defence.

[svengolly]

Double dummy I get but it still doesn't seem logical to me that N would lead a club into West's first bid suit. After all, it could just spear something in South's hand.

[smerriman]

svengolly, on 2016-December-26, 19:49, said:

After all, it could just spear something in South's hand.

.. which declarer was almost certainly going to win anyway; the chances it actually gives away a trick are far lower than chances when leading any other suit.

GIB doesn't (always) lead long suits like humans - all calculations are based on double dummy analysis, so assumes declarer will 'figure out' who has what and the best line to take, like experts often can via other means. Computer analysis has shown leading from suits like its clubs are superior (https://www.amazon.c...d/dp/1554947596)

[svengolly]

smerriman, on 2016-December-26, 20:03, said:

.. which declarer was almost certainly going to win anyway; the chances it actually gives away a trick are far lower than chances when leading any other suit.

GIB doesn't (always) lead long suits like humans - all calculations are based on double dummy analysis, so assumes declarer will 'figure out' who has what and the best line to take, like experts often can via other means. Computer analysis has shown leading from suits like its clubs are superior (https://www.amazon.c...d/dp/1554947596)

Thanks. Learning something here. Now if I could only figure out why GIB pards will often lead a suit, strike gold in my hand, then switch if they regain the lead. That drives me crazy.

[Stephen Tu]

svengolly, on 2016-December-26, 20:12, said:

Thanks. Learning something here. Now if I could only figure out why GIB pards will often lead a suit, strike gold in my hand, then switch if they regain the lead. That drives me crazy.

That's because GIB has pretty much zero knowledge of signalling. And it doesn't do deep enough reasoning like "partner must have this because he tried to do such and such, otherwise he would have done something else", or "declarer can't have this because he would try a different line".

It mostly just deals hands to fit the auction and previous plays, and plays the best card on average over all such deals.

The actual opening leads, are mostly OK on average when it understands the auction. And not doing silly stuff like not banging down 2 aces vs. 6nt. But after that it is pretty horrible compared to better humans.

[svengolly]

Stephen Tu, on 2016-December-26, 21:18, said:

That's because GIB has pretty much zero knowledge of signalling. And it doesn't do deep enough reasoning like "partner must have this because he tried to do such and such, otherwise he would have done something else", or "declarer can't have this because he would try a different line".

Heck sometimes I win, return the suit and it will still switch.