finesse or drop

[Douglas Newlands]

The club situation is

AJ9543

K2

The long clubs have no vestige of a side entry
and you need 6 club tricks.
Play the K and RHO plays the 10.
You look at RHO and see a strong player.
You play a small club and LHO plays small.
What are you going to do? Is it 50-50?

doug

[Dave Flower]

On Monday, 6 May 2013 14:11:29 UTC+1, Douglas, wrote:
> The club situation is AJ9543 K2 The long clubs have no vestige of a side entry and you need 6 club tricks. Play the K and RHO plays the 10. You look at RHO and see a strong player. You play a small club and LHO plays small. What are you going to do? Is it 50-50? doug

It would help to know whether the bidding has revealed your club doubleton. If so, I finesse, as RHO would have played Q from Q 10 doubleton !

Dave Flower

[Charles Brenner]

That's a nice point, which would be a great point if the dummy had
another club. With only 8 clubs we're finessing anyway so RHO's
possible trickiness doesn't matter.

[vsp...@hotmail.com]

In Lawrence's book on card combinations
it was not unusual for Lawrence to play
the suit 3 or more different ways depending
on the rest of the hands.

[spenser]

In article
<4edc2b7b-19f1-41f5...@o2g2000yqb.googlegroups.com>,

Of course it matters! With 8 cards still missing the 10, finessing the
9 picks up the suit when LHO started with 10xxx; however, if you're
still missing the Q, finessing won't pick up Qxxx. Therefore, if you
need to bring home the suit, it would be correct to play for Q10
doubleton on your right. Dave makes the excellent point that playing
the Q is obligatory to give declarer a losing option.

--
Dennis Cohen

[Nick France]

On May 6, 9:11 am, Douglas Newlands <douglas.newla...@gmail.com>
wrote:

Finesse especially if a good player. RHO could have, QT, T6, T7. T8
doubletons. If he has the singleton ten you can't make 6 club tricks,
if he has QTx you can't do anything. And if for some reason (which I
can't figure), he does it from Txx, the queen will show up on the next
trick so you dont have to guess.

On the face for the cases that count it looks like 3-1 but since he
had a choice of what to play on the three Tx hands the odds seem about
3-2 in favor of the finesse on the hands that matter.

Nick France

[Steve Willner]

On 2013-05-06 9:06 PM, Nick France wrote:
> RHO could have, QT, T6, T7. T8 doubletons.

...

> On the face for the cases that count it looks like 3-1 but since he
> had a choice of what to play on the three Tx hands the odds seem about
> 3-2 in favor of the finesse on the hands that matter.

As Dave pointed out, RHO also had a choice of what to play from QT.
However, if he knows the exact layout and knows you will know he is good
enough to play T from Tx, he's probably best off to play the T in all
cases, winning one time in four. At least I don't see any strategy that
does better, but maybe I'm missing one.

--
Help keep our newsgroup healthy; please don't feed the trolls.
Steve Willner Phone 617-495-7123 swil...@nhcc.net
Cambridge, MA 02138 USA

[Will in New Haven]

On May 6, 9:11 am, Douglas Newlands <douglas.newla...@gmail.com>
wrote:

I finesse. RHO could have any singleton or doubleton Ten. A good
player will certainly play it from Tx. he will also play it from QT if
he feels that you will give him credit for playing it from Tx. If he
would play you for a newbie, then you can eliminate QT because he
would make the "tricky" play but that just makes the case for the
finesse stronger.

So you don't have to go to level three, "what does this player think
of me?" but you can be complimented if he wins the Queen from QT.

Of course, if he has the singlton we are just deciding how many
undertricks we are going to make. Winning three IMP by taking the
finesse and going down one fewer is not to be despised, although I
sneer at it a bit.

--
Will in New Haven

[Will in New Haven]

On May 6, 7:51 pm, spenser <s...@dogbreath.com> wrote:
> In article
> <4edc2b7b-19f1-41f5-97d7-a7e78bf9b...@o2g2000yqb.googlegroups.com>,

if playing the Ten could be from Tx, then the Ten from QT gives
declarer a _different_ losing option and it is probably correct.

[Douglas Newlands]

On 8/05/13 1:22 AM, Will in New Haven wrote:
> On May 6, 9:11 am, Douglas Newlands <douglas.newla...@gmail.com>
> wrote:
>> The club situation is
>>
>> AJ9543
>>
>> K2
>>
>> The long clubs have no vestige of a side entry
>> and you need 6 club tricks.
>> Play the K and RHO plays the 10.
>> You look at RHO and see a strong player.
>> You play a small club and LHO plays small.
>> What are you going to do? Is it 50-50?
>
> I finesse. RHO could have any singleton or doubleton Ten.

The two hands are
Butler, none vul, Dealer North
S 42
H 853
D 32
C AJ9632


S KQ6
H AQ106
D KJ105
C K5

and after the JS lead (from AJ10xx) there are few options in the play!


The argument against finessing is that you need the suit to be 3-2
if you are going to make 6 tricks so, at the point of the decision,
you need them to be Qxx - 10x or xxx - Q10.
Crudely, there is one space left in each hand so it is a 50-50 decision.
This suggests you should play for the drop rather than finesse.
Do you like this argument.

doug

[Nick France]

On May 7, 7:32 pm, Douglas Newlands <douglas.newla...@gmail.com>
wrote:

No you are missing using Vacant spaces. There are not just one card
missing in the suit but 2 cards. Also in any case if RHO started with
QTx it doesn't matter what you do. So there are only two cases that
matter, LHO started with Qxx or xxx. At the start there were 6 cases
for Qxx and only 4 for xxx. That makes the finesse a 3 to 2 favorite
for those casses where it matters.

To say that there was some overbidding would be a true statement but
now that you are in the contract it looks like you need the club suit
and the best line is to finesse the jack.

Nick France

[Charles Brenner]

On May 6, 4:51 pm, spenser <s...@dogbreath.com> wrote:
> In article
> <4edc2b7b-19f1-41f5-97d7-a7e78bf9b...@o2g2000yqb.googlegroups.com>,
>
>
>
>
>
>
>
>
>

My comment that it "doesn't matter" of course referred to declarer's
play given that defender has played the 10. Is there really a reading
under which I could have been referring to what matters for the
defender? Perhaps, though I don't see it.

[Will in New Haven]

On May 7, 7:32 pm, Douglas Newlands <douglas.newla...@gmail.com>
wrote:

No. The Ten didn't fall at random. There are more cases of Tx than
there are of QT and your opponent is probably capable of playing the
Ten from each of them, as well as from T, Txx and, most cruelly, QTx.
The play from QTx is absolutely brilliant in a long, grueling match.
Now you will sit and sit and stew and eventually go down and you have
used a ton of energy that you would not have used had he played the x.
The Q from QTx, a Grovner coup, is a little too rich for my blood but
it would work and be hilarious. Unless you had another Club, of
course.

--
Will in New Haven

>
>
> >A good player will certainly play it from Tx. he will also play it from QT if
> > he feels that you will give him credit for playing it from Tx. If he
> > would play you for a newbie, then you can eliminate QT because he
> > would make the "tricky" play but that just makes the case for the
> > finesse stronger.
>
> > So you don't have to go to level three, "what does this player think
> > of me?" but you can be complimented if he wins the Queen from QT.
>
> > Of course, if he has the singlton we are just deciding how many
> > undertricks we are going to make. Winning three IMP by taking the
> > finesse and going down one fewer is not to be despised, although I
> > sneer at it a bit.
>
> > --

> > Will in New Haven- Hide quoted text -
>
> - Show quoted text -- Hide quoted text -
>
> - Show quoted text -

[Charles Brenner]

> > On 8/05/13 1:22 AM, Will in New Haven wrote:
>
> > > On May 6, 9:11 am, Douglas Newlands <douglas.newla...@gmail.com>
> > > wrote:
> > >> The club situation is
>
> > >> AJ9543
>
> > >> K2
>
> > >> The long clubs have no vestige of a side entry
> > >> and you need 6 club tricks.
> > >> Play the K and RHO plays the 10.
> > >> You look at RHO and see a strong player.
> > >> You play a small club and LHO plays small.
> > >> What are you going to do? Is it 50-50?

> The Q from QTx, a Grovner coup, is a little too rich for my blood but
> it would work and be hilarious. Unless you had another Club, of
> course.

In my experience it still works. I played with a lady once -- I don't
remember why -- who had Qxxx opposite dummy's AJ10xxx, finessed once
and RHO decided to duck. No matter. She led another club from hand,
LHO showed out, and she finessed again anyway.

[jfn]

Supposing that the club doubleton is known or must be assumed
opponent first level : has QT or T , drop (100/100 for you)
opponent second level : always plays T with Tx : finesse (67/100 for you at least)
opponent third level : plays 50-50 T or x with Tx ; drop or finesse 50-50

[Charles Brenner]

I like the idea of the analysis but am not able to reproduce the
second and third results.

"First level" I understand -- you are saying that against an opponent
who plays the 10 only from Q10 (or a singleton), the drop wins in 100%
of the cases where there is a winning play.

"Second level": There are three 10x holdings (i.e. x can be any of
6,7,8) and only one Q10 holding. That's four holdings against which we
have a winning play, so the finesse wins in 75/100 cases.

"Third level": Now the opponent's 10 play might be from the one Q10
holding or from half of the three 10x holdings, hence is 3:2 or 60/100
to be from 10x, assuming we have a winning play at all.

[David Goldfarb]

In article <2795b534-c5da-42d2...@d6g2000yqi.googlegroups.com>,
Will in New Haven <willre...@yahoo.com> wrote:
>a Grovner coup

Pronounced "Grovner" but spelled "Grosvenor".

--
David Goldfarb |
goldf...@gmail.com | "End of the universe. Have fun. Bye-bye!"
gold...@ocf.berkeley.edu |

[Steve Willner]

On 2013-05-07 7:32 PM, Douglas Newlands wrote:
> if you are going to make 6 tricks so, at the point of the decision,
> you need them to be Qxx - 10x or xxx - Q10.

Yes: three cases of the former and one of the latter.

> Crudely, there is one space left in each hand so it is a 50-50 decision.
> This suggests you should play for the drop rather than finesse.
> Do you like this argument.

I haven't a clue what you mean by "one space left." The four relevant
cases are equally probable. Unless you think RHO will play the T from
Tx less than 1/3 of the time (and will never play Q from QT), the odds
say to finesse.

[Douglas Newlands]

On 9/05/13 3:09 AM, Charles Brenner wrote:
> On May 8, 7:26 am, jfn <jean-francois.noi...@wanadoo.fr> wrote:

>> Le lundi 6 mai 2013 15:11:29 UTC+2, Douglas, a �crit :

I received the following email quibbling with Charles Brenner's comments
and I need to understand who is right and why. This commentator, DB,
does not read rgb, hence his initial remarks.

\begin{quote}

Not sure if the Charles Brenner comments are the one's starting with

"I like the idea of the analysis but am not able to reproduce the second
and third results."

or if there is another set of comments which haven't come through..
Which Blog were you on?

First level - Agree

Second level - Disagree
He is ignoring the fact that you have seen 3 of the opponents cards at
the key moment in time.
Assume opponents are sound and play small cards at random (not
necessarily true)
If W has played the 6 and then 7, and assuming that he has a third card
(otherwise there is no winning line) then he was dealt with 678 or 67Q
and East 8T or TQ. (Note that the problem you gave me originally did not
specify that the T had been played.) If E always plays the T from T8
then it is equally likely that he has T8 or TQ. Having seen the 6 and 7
played by West then East cannot have T6 or T7, only T8 or TQ (assuming
he will never play the Q on the first round - if there is any chance
that he will do that then the finesse is better.) Finesse and drop are
equally likely.

Third Level - Disagree again because the 6 and 7 have been played.
We are still in the position that E has T8 or TQ and will assume that he
does not play Q from TQ. Initially these holdings are equally likely.
However half the time he was dealt T8 he will play the 8 and half the
time the T. This is an example of Bayes Theorem. P(dealt TQ given that
he plays T) = P(dealt TQ and plays T)/ {P(dealt TQ and plays T) +
P(dealt T8 and plays T)} = a/{a+a/2) = 2/3
There is a 2/3 chance that he has TQ doubleton and so play for the drop.
Again, if he may play the Q from QT then this reduces the chances of the
drop and if he plays it more than 50% of the time then you should finesse.

\end{quote}

doug

[Will in New Haven]

On May 8, 11:05 pm, Douglas Newlands <douglas.newla...@gmail.com>
wrote:

I disagree that the play from T8 is at all random. People play the Ten
in this case much more than half the time.

[Charles Brenner]

On May 8, 8:05 pm, Douglas Newlands <douglas.newla...@gmail.com>
wrote:

[CB] I'll mark my comments with [CB]. Yes, that's the beginning of my
comments.

> or if there is another set of comments which haven't come through..
> Which Blog were you on?

[CB] "Blog"? And who is "you"?

[DB]

> First level - Agree
>
> Second level - Disagree
> He is ignoring the fact that you have seen 3 of the opponents cards at
> the key moment in time.

[CB] More precisely, I'm ignoring what they are since, according both
to the jfn post to which I was replying as well as to your original
post, we explicitly do not know what they are.

[DB]

> Assume opponents are sound and play small cards at random (not
> necessarily true)
> If W has played the 6 and then 7, and assuming that he has a third card
> (otherwise there is no winning line) then he was dealt with 678 or 67Q
> and East 8T or TQ.

[CB] Thus DB has changed the problem by pretending that the spots are
known. This isn't necessarily a big deal, but it is a way to introduce
confusion and make the problem more difficult. I know what he will
say: "Ok, I understand we don't know the actual spots played but we
know SOME spots were played and in order to talk about them
conveniently I'll assign them names '6' and '7', whatever they
actually may be." A small problem with that is that some spots may be
treated differently than others by the defender, so to get around that
our correspondent had to stipulate random play. As I analyzed that
stipulation was unnecessary. The big problem is that DB has in effect
stipulated that from 876 defender always conceals one particular card
-- namely the one that is for convenience being called '8'. This
stipulating can be dealt with by the so-called "restricted choice"
argument, but some people find it hard to grasp and anyway if we don't
invent difficulties from the beginning we don't need to deal with
them.

[DB]

> (Note that the problem you gave me originally did not
> specify that the T had been played.)

[CB] That seems to be a private note from DB to Doug. On
rec.games.bridge we knew.

[DB]

> If E always plays the T from T8
> then it is equally likely that he has T8 or TQ.

[CB] Now DB has fallen into real error, because by assigning a
specific name "108" for the holding that is actually "10x", he
automatically overlooks the need to consider restricted choice.
Vaguely put, the point is that 108 is one holding while 10x is three
holdings and the consequence is that if you insist on labeling the
actual spot played (e.g. as "8") then you must consider the restricted
choice argument in order to back out of the confused start and come to
the right answer.

[CB] The easier approach in my opinion (opinion re the best way to
understand, not opinion about mathematics which of course isn't a
matter of opinion) is to think about "policy". By that I mean imagine
a declarer who adopts a particular policy and then calculate how
successful that declarer is.

[CB] In this case -- "second level" -- we are assuming that RHO has
played the 10, always plays 10 from 10x and also from Q10, and since
those are the only cases of interest when the 10 appears, we may as
well assume that one of them is the case.

[CB] If declarer's policy is always to finesse when faced with this
situation, declarer will succeed whenever RHO was dealt 108, 107, or
106, and fail when RHO is dealt Q10. Hence this policy succeeds 3/4 of
the (relevant) time, and of course the alternative policy of "drop"
succeeds 1/4. Note that this analysis doesn't require any assumption
of random spot play by the defender.

[CB] In summary, DB complicated the problem by changing it, then
failed to cope by taking into account LHO's restricted choice of spot
plays from Q76 versus freedom of choice from 876.

[pgmer6809]

On May 8, 8:05 pm, Douglas Newlands <douglas.newla...@gmail.com>
wrote:

1) Suitplay (and hence Roudi's Rules) says to finesse. RHO could play
the T from Tx which is 10.17% or QT which is 3.3%.
So the finesse is the clear favorite.
Roudinesco in his book discusses this exact combination with West
dropping the T on page 18.

2) When it comes to Vacant Places, and Restricted Choice these are
calculation short cuts, and you need to be very clear about when they
apply and when they don't. When it doubt go back to first principles
which usually means lots of C(n,r) math!
When I try to do this, my head hurts.!

Here the T from Tx and from QT is 'forced'. The choice in both cases
is 'restricted', in the sense that if West is a good player the T from
Tx is a 'mandatory' false card.
What of East's play of the 6 and 7 of clubs? From Qxx he MUST play the
two x's so his choice in that case is restricted. From xyz he could
play any one of three doubletons (xy, xz, yz) so the fact that he
played xy say implies that he is more likely to have the Qxy.
So from both of these considerations we conclude that West is more
likely to have Qxx and East is more likely to have Tx.

[Douglas Newlands]

We have had enough of this hand so to wind it up, here is the hand.

Butler, none vul, Dealer North
S 42
H 853
D 32
C AJ9632

S AJ1073 S 985
H 972 H KJ4
D 98 D AQ764
C 874 C Q10

S KQ6
H AQ106
D KJ105
C K5

After the JS lead, I played KC and finessed and ended up 4 down.
It was suggested that finessing was wrong, hence this thread.

At another table, two pro-client pairs were facing off with the pros
sitting North and East. The contract and lead were the same but when
South led the KC at trick 2, East dropped the QC!
At the table, South was oblivious to all this and played the AC next
so North could smile contentedly to East who could only wish that North
had been sitting South!

doug