[luke 0]

Hi all, I play a lot of the free daylong tournaments and the most common GIB bug/inaccuracy is that it will frequently lead singletons and doubletons vs NT, and not in a suit that its partner has mentioned (or even in an implied suit). I think that it overprioritizes not giving up a finesse with holdings like KJTx or AJxxx or Kxxx.

Anyone else experience this problem?

[ahydra]

There may be some logic behind this - I recall from reading Bird/Anthias' book on opening leads that low leads from holdings like AJxx, K10xx were the ones most likely to give away tricks, and GIB may well be finding the same things in its simulations. Of course, those sims are not going to take into account the fact partner may well be misled by a short-suit lead.

ahydra

[iandayre]

Luke, your perception is correct. But it is not a bug, it is a deliberate programming decision. There is an explanation in the GIB system notes, quoting some reference book.

[Stefan_O]

iandayre, on 2016-August-31, 13:41, said:

Luke, your perception is correct. But it is not a bug, it is a deliberate programming decision. There is an explanation in the GIB system notes, quoting some reference book.

Really?
Isn't it the simulation that decides?

I figured this was because it trusts declarer's NT-bid too hard.
So "if I have the king in pd's suit, declarer must surely have AQ for his NT-bid" --- it thinks...

[Stefan_O]

Quote

"GIB usually leads passively against NT (read the book Winning Notrump Leads to understand why)."

I have seen the above in the system-notes before, but understood it more like an explanation why Gib simulations frequently come up with passive leads,
rather than a deliberate programming decision.

I have a high preference for passive leads, myself, but pd's suit is definitely an exception