[goffster]

If I hold


AJTxx
opposite
xxxx

If I lead a low to AJTxx, and lefty does not fly queen, then i know it does not have KQx.
(from experience)

Therefore it is always right for me to play the ace with the current incarnation of GIB

[mgoetze]

Yes, that is the way GIB works, it assumes double-dummy play. A matter of fundamental design.

What's your point?

[shyams]

goffster, on 2015-October-11, 14:10, said:

If I hold


AJTxx
opposite
xxxx

If I lead a low to AJTxx, and lefty does not fly queen, then i know it does not have KQx.
(from experience)

Therefore it is always right for me to play the ace with the current incarnation of GIB

GIB also finds plays of honor in 2nd seat from Hx. I've had a situation when RHO gib did that, and I assumed RHO started with QJxx so chose a repeat finesse in favour of another (possibly better) line. Guess what, LHO held Jxxx.

So I don't think that one quirk alone gives declarer a definitive edge on all deals.

[broze]

mgoetze, on 2015-October-11, 21:16, said:

Yes, that is the way GIB works, it assumes double-dummy play. A matter of fundamental design.

What's your point?

Does it do this completely? i.e. When I lead to AQT in dummy AFAIK GIB doesn't randomly play honours with King Jack as it would if it thought you were always going to double finesse.

In any case I think it's pretty obvious what OP's point is.

[goffster]

GIB does not *have* to assume double dummy play.
There could be some cases where it recognizes certain combinations
where splitting honors is inane.

[iandayre]

goffster, on 2015-October-11, 14:10, said:

If I hold


AJTxx
opposite
xxxx

If I lead a low to AJTxx, and lefty does not fly queen, then i know it does not have KQx.
(from experience)

Therefore it is always right for me to play the ace with the current incarnation of GIB

A corollary to this is that, in 4th hand, GIB never falsecards with equal honors. If you are playing to the AJT and RHO wins the K, you can be 100% certain that the second finesse will succeed. This applies equally to QJ and JT combinations. Interesting that the programmers allowed this weakness while GIB makes falsecards extensively - sometimes to its own detriment - in other situations

[barmar]

I recall seeing something in the code where it randomizes which card of equals to play, to force declarer to use restricted choice. But for some reason I don't think it's done in the case where it's winning in 4th seat.

I've occasionally seen it not split. Last night I was in 4♥ with JT85 in dummy opposite A96 in hand. East had KQ7432. I ran dummy's 8 and it held, East playing the 7. Of course, when West showed out, this was a pyrrhic victory.

[johnu]

mgoetze, on 2015-October-11, 21:16, said:

Yes, that is the way GIB works, it assumes double-dummy play. A matter of fundamental design.

I would call it a fundamental misdesign Maybe you can find examples where sacrificing honors and high spot cards is necessary, but my guess is that it is at least 10 to 1 a losing play.

[iandayre]

Well, so much for that theory. I also thought your point was valid.

http://tinyurl.com/q9yr37x

Note E did not split the C honors.

[1eyedjack]

iandayre, on 2015-October-20, 14:19, said:

Well, so much for that theory. I also thought your point was valid.

http://tinyurl.com/q9yr37x

Note E did not split the C honors.

Maybe East was worried that West might hold Jackleton, and swapping the J and T it might have split. Be interesting to know if it covers J lead from North