This article was first published in  vol. 8, nr. 2, April/May 1997

The GIB Zone (2) 'The Bet is Off'
  by Onno Eskes
 
 zia  Zia Mahmood

The new computer program 'GIB' (Goren in a box), developed by the American professor Matthew Ginsberg signals a breakthrough in bridge software. For the first time, a computer plays as well as an expert, and perhaps even better! In this series, IMP editor Onno Eskens lets GIB loose on a number of interesting deals, and arrives at surprising conclusions. IMP readers are invited to submit to us problems in declarer play they have encountered at the table, and have them tested by GIB.

 
1. W/both A K Q J 
5 4 3 2 
J 10 9 
Q J
- ----
e 5 4 3 2 
A Q 
K 2 
A K 10 9 8
 
 WEST 
  3 --- 
  Pass 
  Pass 
  NORTH -- 
  Pass 
  6NT 
  Pass
  EAST --- 
  Pass 
  Pass
  SOUTH 
  3NT 
  Pass
 West leads the  10. Looks pretty hopeless, or do you disagree?
 
2. N/both -
9 7 
A Q 8 5 3 
J 9 7 
K 9 7
A 10 3 
J 9 6 
A K 8 
A 6 5 4
e -
 
 WEST --- 
      -  -  
    2 
   3NT 
   Pass
  NORTH---  
-1 
 Pass 
 Pass 
 Pass
EAST --- 
1NT 
 2 
 4 
Pass
 SOUTH 
 Pass 
 Pass 
 Pass
 South starts with the 5, North plays the jack. You decide to duck. North switches to the 10. How do you proceed?
 
 
3. W/both 9 7 3 
9 7 6 2 
K 10 8 6 
9 2
- -
e A K 6 2 

A Q J 9 7 4 2 
A 8
 
  WEST --- 
  2 - 
  Pass 
NORTH --- 
Pass 
Pass
EAST --- 
4 
Pass
SOUTH 
6 
 2 was a weak two. West tracks the  K. You ruff and A fells the outstanding trumps. Now what?
 
San Francisco - 1 December, 1996. Zia is the big star of the Fall Nationals, having just triumphed in the premier event, the Reisinger. Smiling from ear to ear, he accepts the congratulations of his predominantly female admirers. Then he is accosted by a man he's never seen before. "Mr. Mahmood, my congratulations; and incidentally, may I ask you something?" "But, of course," replies the always amiable Pakistani, "what's it about?" "It concerns a one-million-pound bet." The Pakistani grows pale. "What is your name, sir?," he immediately asks. "Matthew Ginsberg," says the man. Suddenly there's little left of the great Zia with his aura of invincibility. He cringes, and mumbles, "The bet is off!," and walks out of the room. What's going on here?
 
1. W/both A K Q J 
5 4 3 2 
J 10 9 
Q J
10 9 
K J 10 9 8 7 6 
Q 4 3 
7
spade 8 7 6 

A 8 7 6 5 
6 5 4 3 2---
e 5 4 3 2 
A Q 
K 2 
A K 10 9 8
 
 
 Eugene, Oregon - 25 October, 1996. Late at night they're still playing bridge in a dusty university office. South has arrived at 6NT following a 3 -preempt by West, and the lead is the 10. Declarer thinks for about thirty seconds, plays a couple of rounds of clubs, and next a spade to dummy. This is followed by the J.

The east seat is occupied by a player who has never been caught making a defensive error. You can almost hear him count: "If I take the A and continue diamonds, declarer will cash his clubs and spades, leaving:
 

1. W/both


-

K J 

-
spade im- 
ma- 
teri- 
al-----
e
A Q 

-
 
 
 On dummy's last spade West will be subjected to a simple squeeze. Rising with the ace is no good, so I'll duck a round." And so he did. But South wasn't caught out; he played K, and rattled off his black-suit winners. This led to a rare position:
 
1. W/both

10 9 
-

K J 
Q 4 
-
spade

A 8 7 6 
-
e
A Q 

-
 
 
  The last of dummy's spades was played, and both East and South discarded a diamond. And West? Out of necessity he had to let go a diamond as well. The position was clear now: 10 from dummy presented the defenders with an insoluble problem. Should East rise with the ace, he then crashes his partner's queen, while promoting 9 for the twelfth trick. East ducked, West was thrown in and had to play a heart into South's tenace: six NT in the bag. The end position is known in the literature as the winkle. If you should recognise the hand, or the infallibility of the defenders, you're right. The south player played the deal against Bridge Master, Fred Gitelman's program with pre-dealt hands in which the opponents make no errors, and where you always have to rely on proper technique as declarer.

 In Zia's book, Bridge, My Way, which appeared a few years ago, he offered to take a one-million-pound bet that no computer would be able to beat him at the bridge table. The stunt seemed to work in that it produced a lot of publicity for his book. That is, until last month, when word reached him that bridge program GIB, brainchild of American professor Ginsberg, proved capable of incredible feats of declarer play. Indeed, the above winkle was carried out by a computer equipped with GIB….

Hold the press!
 

2. N/both K Q J 4 2 
K 7 2 
10 
Q 10 3 2
9 7 
A Q 8 5 3 
J 9 7 
K 9 7
A 10 3 
J 9 6 
A K 8 
A 6 5 4
e 8 6 5 
10 4 
Q 6 5 4 3 2 
J 8
 
 WEST --- 
      -  -  
    2 
   3NT 
   Pass
  NORTH---  
-1 
 Pass 
 Pass 
 Pass
EAST --- 
1NT 
 2 
 4 
Pass
 SOUTH 
 Pass 
 Pass 
 Pass
This hand occurred at the 1997 Dutch Mixed Pairs Championship. Tjeerd Kootstra was the only player making ten tricks in 4 without assistance from the opponents. South led 5, taken by North's Jack, and the 10 was returned. Kootstra rightly concluded that this had to be either a singleton or a doubleton. Therefore, South had Q, which made it clear that North had to posses  K for his opening bid. Thus, there was a potential loser in each of the suits. Kootstra took the A , played a heart to the ace, and exited in hearts. North took his king and switched to spades, taken by the leader's ace. The last trump was pulled, and it was clear that three club tricks were required. Tjeerd Kootstra clearly recognised that his combined club holding lent itself to an 'intrafinesse'. He played a small club from the east hand, and covered South's 8 with the 9. North won his 10, and played another spade, ruffed by West. Now, K dropped South's J, and with Q 3 under dummy's A 6, Kootstra picked up the rest of the suit by finessing against Q, and made his contract.

 Very well played, and appropriately rewarded by a favourable club position. I used this hand in my Saturday newspaper column of the General Commercial Daily, and the day after submitting the piece, I gave GIB this problem to solve. I was curious if GIB could match Kootstra's good line of play. To a certain extent GIB duplicated Kootstra's line: spade ducked, diamond won, trump ace, other trump, spade ace taken. But then… GIB ruffs the 10 and pulls the last trump with  J. He next plays a small club, sees South follow suit with the eight, and plays the king. A booboo, I'm thinking to myself. GIB now continues with 9, small, small, to the Jack, and South is endplayed! He only has diamonds left and is forced to play into the tenace. Then it starts to dawn on me. This line also works if South has a third club, as the clubs are then breaking three-three and the fourth round will provide a pitch for the losing diamond. Stop the press!

It was too late already - the column had been printed. What a mistake. Just imagine the layout had been slightly different:
 

- K Q J 4 2 
K 7 2 
10 4 
10 3 2
9 7 
A Q 8 5 3 
J 9 7 
K 9 7
A 10 3 
J 9 6 
A K 8 
A 6 5 4
e 8 6 5 
10 4 
Q 6 5 3 2 
Q J 8
 
   
 Kootstra's line (club to the nine and ten, club king, run the seven) would have resulted in the loss of two club tricks, whereas GIB's line still works. It's true that in layout North doesn't have much of an opening bid, but nowadays, at matchpoints, they don't shy away from justabout anything as long as they've got a reasonable suit…

 Hand of the year?

Geir Helegemo is one of the more competent bridge players around - that's a well-known fact. At the latest American Nationals he was partnered by The Bridge World's publisher Edgar Kaplan, and he took advantage of the occasion to shine in his customary fashion:
 

3. W/both 9 7 3 
9 7 6 2 
K 10 8 6 
9 2
8 4 
K Q J 8 4 

K 10 7 5 3
Q J 10 5 
A 10 5 3 

Q J 6 4-
e A K 6 2 

A Q J 9 7 4 2 
A 8
 
WEST ---- 
  
2 - 
 Pass
NORTH -- 
Kaplan 
Pass 
Pass
EAST --- 
- 
4 
Pass
SOUTH 
Helgemo 
6 
Helgemo ruffed the heart lead, pulled trumps and led a spade to the seven! East took the ten and forced declarer into taking a crucial decision by returning the 5. Helgemo ducked, allowing dummy's 9 to win the trick. The losing club was subsequently discarded on the fourth spade.

 Had East played back a club rather than a spade, then Helgemo would have taken the ace, crossed to dummy in trumps, and returned the  9, pinning the eight, and picking up the rest of the suit by a repeated finesse. Nicely played by the Norwegian Kootstra: an intrafinesse followed by a pinning finesse and a regular one. This hand was immediately nominated for the Hand of the Year Award. Incidentally, West could have made Helgemo's life a lot tougher by playing 8 in the first round, suggesting a holding of honour-eight doubleton.

 I submit the hand to GIB, telling him that West owns five or six hearts and less than an opening bid. GIB ruffs the heart lead, pulls trumps, and ponders the situation for well over a minute. Lo and behold, he proceeds with a spade to the seven! Again I let East take the ten and return the five-spot. GIB takes the ace in tempo! One down. I ask him why he didn't 'just' let the spade run to the nine. GIB shrugs his shoulders, saying, "Ah well, I don't need to get my name in the paper. I'm just a simple and unassuming computer program. Sure, I noticed the possibility of an intrafinesse, but that's for the macho's among us. After the weak two, the probability of a three-three spade break is about the same as that of a doubleton. But the intrafinesse works only one time in three (honour-eight or eight-small), even if West has a doubleton. I simply take the superior line (spades three-three) - that's what I studied for."

 Straight talk, that was, but it seems GIB is right. It reminds me somewhat of the young Hague expert Roefi Nurmohamed who, after awhile, was no longer capable of taking a normal finesse. The intrafinesse had become the be-all and end-all. This was, I believe immediately following his underruffing period, but before he started preempting on semi-forcing hands. It costs a few points, but one does get one's name in the paper from time to time.

Of course, Helgemo possesses something the computer doesn't: table presence. The moment your intuition tells you that the spades are four-two for sure, then Helgemo's line is best. I next tell GIB that I've peeked in East's hand and noticed four spades. Then I let him play the hand again: spade to the seven and ten; and 5 comes back. GIB plays low, wins the trick and makes his contract. "Yeah, that way it's a cinch," I hear him mutter.


  To IMP Bridge Index