Bridge after Deep Finesse
(Page of 4)

(Initial NOTE: The original of the following article was written in Turkish some time during Ankara Open Teams League Matches in 2007-2008 season, and later put on the web site of ‘Jeofizik Briç Spor Kulübü’ in Ankara. Recently I rewrote it in English making small changes and replacing one of the deals with a newer one. So, it is a little different from the original, but the main theme is the same. Though outdated, I hope it gives you an idea about early impacts of Deep Finesse in some bridge circles in another country.)

As we know, there is a small table affixed to each deal in the hand records. The numbers in that table indicates how many tricks can be taken in each contract by each direction. Or, at least, the computer program “Deep Finesse” says so.

A bridgeur working in “Oracle Corp.” as a programmer, an American named William Bailey, decided to take a break from his job for one year in 1998 in order to pursue one of his dreams. Mr. Bailey, whose greatest passions in life were bridge and computer, developed Deep Finesse (DF) in 1999. With his own words, the creator of DF was not even a good bridge player; but a fanatic in analyzing hands that he played, and could not peacefully sleep until he felt that he completely understood everything about a problematic hand.

When all hands are given to DF, it yields the number of tricks available in all denominations and also indicates which cards should be played or not played to each trick for success.

I am writing these because I have observed that we, bridge players, may often find ourselves in trouble because of DF. I think almost all of us have been the subject or witness of the following discussions.

“It says here that 4 makes, how did you go down in 3?”

“11 tricks in ’s, 10 tricks in ’s, 9 tricks in ’s; you guys found 3, bravo!”

“This 4 should go down; didn’t you lead a  to that?”

Number of examples can be augmented. Before DF came into the scene, we did not hear all these so often; because, in order to find out which contract could be made on which lead or defense, it was necessary to analyze deals using our own brain, and such analysis required both mental labor and time; furthermore, it was never this easy to be adequately sure. Some difficult hands had to be reviewed again; so, the analyses could be completed usually until the next week’s match. Now, there is no need to analyze; DF simply tells us all possible outcomes.

Are there some people among us who still resist DF? I think and hope there are. I remember that, when results of DF were first put in hand records some time ago, the bridgeurs had been roughly divided into two: ‘Resisters’ and ‘Supporters’.

Eventually, the supporters won, but we owe a lot to the resisters. And I believe it is now time for all of us to resist DF. Why? DF has already eliminated the diversity of views on numbers of tricks available, causing uniformity in that area; moreover, if we get into the habit of accepting its results without investigating, how are we going to impede the process toward mental laziness?

Here are some deals, which, one way or another, involved DF in discussions after games.

1. Resisting Deep Finesse

West
10865
Q752
5
J1043
North
4
10
Q10983
AKQ965
East
K9
AJ983
KJ42
87
South
AQJ732
K64
A76
2
D

As the declarer of 4 who went down from South, while you compare the scores with your teammates, you see that 4 was made at the other table; and there is a 10-imp loss. There is no time and need to analyze the deal; you slip away from your teammates to get a hand record.

You find the hand; and get relieved upon seeing the number ‘9’ in the box of ‘-South’ in the DF table, which means 3 only can be made by South. Now, it is time to question how this 4 could have been made at the other table; swiftly you go back and join your teammates.

“Didn’t you lead a  to this 4?”

“I led a , but we couldn’t set it. Does it go down on a  lead?”

You continue with the confidence coming from DF: “They led a  to me, too, but I couldn’t make it. Anyway, only 3 can be made. How did the declarer play at your table?”

“He went to A, discarded a  on K, finessed in ’s, cashed A, J and gave me a . I returned a .”

Not needing to think even for a second, you flounder to keep on interrogating. “If you return a , it makes of course; why didn’t you play a ?”

“A  … (??) … 10 tricks are there: 5 , 3 , 2 .”

Aha, everything he says seems to be accurate. At this point, you get to feel that something is wrong. Your teammate, probably smelling a rat, now asks:

“How did YOU go down?”

“I tried to cash three ’s, East ruffed the third … … Anyway, Deep Finesse says that only 3 is makable.”

“Who is this ‘DIP FINES’? I don’t give a hoot what HE says; come on, set me, I am playing.”

That’s it! Here is a genuine resister.

I don’t know how long the above chat goes on; however, by virtue of that, it surely occurs to someone that 4 may not be going down with a  lead but with some other leads. (If there were no dissidence, who knows how many people would go home with wrong/deficient information.)

If we are not too lazy to think over the deal, we can see that 4 goes down on a  lead (for example). Lead of a  does not only prevent declarer from making three  tricks, but also ensures that the entry problem goes on.

(NOTE: All persons and dialogs aforesaid are fictitious; any resemblance to real persons or dialogs is purely coincidental.)

2. Supporting Deep Finesse

After first session of a pairs event, I and my partner that time (Uğur Taş) were going over the hands. On the deal below, opponents had bid 3NT by South and took 9 tricks after my  lead.

West
K10843
J985
AK
J6
North
2
AK10
QJ9875
A103
East
Q65
Q74
104
Q9542
South
AJ97
632
632
K87
W
N
E
S
1
2
2
2NT
P
3NT
P
P
P
D
3NT South
NS: 0 EW: 0

Our score was below average, as expected. It seemed that there was nothing to do about it, until my partner noticed the DF report: 8 tricks in NT by South.

We looked again and again, but could not find a way to prevent South from taking 9 tricks.

For a second, we even thought of the probability that DF might be wrong; maybe there was a bug in it? We were about to give up, then reconsidered: “DF has never reported something wrong so far. If it says that there are 8 tricks, then it should be so. It is the correct time to advocate DF”; and we started all over again…

DF was right, of course.

Did you find the killing lead?

K: if it holds, switch to a .

Believe me; we did not consult DF to find it. As I implied above, it sometimes kills the fun. Yet, here, thanks to DF; if it did not awaken us, we would never discover this oddly logical (!) lead.

3. Deep Finesse cannot think.

West
7542
103
AJ8654
5
North
AJ103
QJ75
73
982
East
86
K8642
Q109
J76
South
KQ9
A9
K2
AKQ1043
D

DF says: “by South: 12 tricks in and NT; by North: 11 tricks in  and , 7 tricks in NT.”

(NOTE: We all know the importance of declaring contracts by the correct side; still, the difference in tricks in NT by North and South is noteworthy here.)

In this deal, it is very easy to obtain the results of DF by using human brain only; no complication whatsoever. However, can DF reach the conclusion that the declarer below will reach shortly? Impossible, because DF is just a perfect double-dummy analyzer, with no ability to think; not yet… (I wonder, if it learns to think, how often it will go down.)

South in first chair opens strong, artificial 2, West tucks a 2 (!?). Finally N-S reaches 6. West leads 5.

Let us put ourselves in declarer’s place (South). If  finesse is on, we are home. Yet, West overcalled 2 vulnerable; can we make 6 if West has both K and A? Let us think…

Yes, just replace West’s 3 with K, all the rest being the same. We cash all the ’s and three ’s, meanwhile counting West’s hand. He shows up with a singleton , four ’s, probably six ’s; then, he must have two ’s.

Here is the four-card ending (note that West discarded four ’s and one  on the ’s).

West
K10
AJ
North
J
QJ
7
East
8x
Q10
South
A9
K2
D

We cash J and discard 2. What will West discard? If he discards a , we exit with a ; if he discards a , we cash A.

At the table, West discards T, and we play a  to the Ace, saying to West: “give me the K.” … West produces 3.

I believe someone playing 6 like this deserves appreciation. Perhaps it is West who deserves more appreciation here.

I suppose any computer program with capability of playing single-dummy would not go down in this 6, because, determining that East has more ’s, it would play according to the odds.

Sometimes, it only takes a human to go down.

(Final NOTE: This was about 8 years ago. Today, Deep Finesse still cannot play bridge, but it has taken its place as an indispensable bridge companion for many of us, to help fully analyze double-dummy deals and/or to verify one’s findings.

Let me add that there are several software programs which can play bridge: Jack, WBridge5, Ginsberg’s Intelligent Bridgeplayer (GIB), Bridge Baron, Shark Bridge… For further information, see “Bridge Playing & Simulation Software Review” by Pete Matthews: (http://web.mit.edu/mitdlbc/www/articles/Bridge_Playing_Software_Review.pdf ).

Since 1997, every year, competitions among computer programs were organized; in addition, some ‘computer-versus-human’ tournaments/matches were played, and ‘the computer’ obtained quite successful results. I hope someone experienced/knowledgeable in this area considers writing an article about this interesting theme, including a few hands from previous matches that made a difference between computer and human.)

EDIT: On the last deal East's 's were T3, corrected.