Refuting Mike Lawrence's theory on 4-card overcalls
metobillc
supporting 4-card overcalls was suspect. I received some email asking
for clarification, so I wanted to present some arguments that show his
claim is false. I have a lot of respect for Mr. Lawrence's writing,
and have read almost all of the bridge books he has written. In
almost all cases, I have found them to be well-reasoned, with the
exception of the claim below from his book "The Complete Book on
Overcalls in Contract Bridge".
1D: S 8 2
H A Q T 7
D T 7 6 5 4
C A Q
In the example, your RHO opens 1D, and you hold the hand above. Mr.
Lawrence recommends overcalling 1H on the good 4-card suit, reasoning
as follows:
"If the possession of five cards in the suit opened bothers you,
forget it. Your length in diamonds plus opener's length assures you
that your partner (and LHO) are also short. This means your partner
is likely to have heart support. The length in diamonds therefore is
not a minus but, rather, an asset."
The first clue that the reasoning might be suspect is the symmetry
between the two unknown hands, that of your partner and your LHO.
Anything that is true for one of them must be true for the other, as
acknowledged by Mr. Lawrence in the parenthetical remark. If partner
is likely to have heart support and diamond shortness (good for you),
equally likely is LHO to have hearts over you and diamond shortness
(good for them).
A rigorous way of looking at the problem is to count vacant places in
the 4 hands, and deal the remaining hearts at random into them to see
on average how much support partner will have for your suit. Let's
suppose that opener will typically have 5 diamonds for opening 1D
(although the very fact that you have 5 of them makes his expected
number lower). RHO has 8 (13 minus 5) vacant places in his hand, and
LHO and partner each have 13. So if we deal the 9 remaining hearts
into those places at random, we end up with partner and LHO each
holding 13*(9/(13+13+8))=3.44 hearts, and opener holding 8*(9/
(13+13+8))=2.12 hearts. On average partner will not have enough
support for an 8-card fit. Note that this analysis did NOT depend on
the opener opening diamonds on this hand -- it could have been any non-
heart suit, and the calculation would remain the same.
Now consider a similar hand where your club and diamond suit lengths
are switched. Let's suppose now that opener will have 6 diamonds for
opening 1D -- more than the previous case, because you hold fewer of
them (the exact number doesn't matter, just that it's more than in the
case where you held length in opener's suit). RHO has 7 vacant
places, and LHO and partner each have 13. So on average partner and
LHO will hold more hearts, 13*(9/(13+13+7))=3.55 hearts, and RHO will
hold fewer 7*(9/(13+13+7))=1.91 hearts. The true advantage is to hold
shortness in the opponents suit.
The second advantage to holding shortness in opener's suit is that you
now hold length in suits other than hearts and diamonds. If there is
no heart fit, it is more likely that you will be able to find a safe
place to play if partner cannot support hearts. There are many
possible advantages to making four-card overcalls, but if you are
going to make one, do it in spite of your length in opener's suit
rather than because of it.
Bill Campbell
brsri...@gmail.com
I might be naive on this one but I believe Mr. Lawrence is right. What
can I say, I am a believer :)
Boris
Phil
a bigger asset - that IF partner is short in diamonds and has heart
length, you can ruff your diamonds in dummy without being overruffed.
(You might be "pre-ruffed" by LHO but that's usually not a problem.)
To me, that potential makes this worth it.
Whereas, if you had a doubleton diamond and overcalled on this 4-card
heart suit (AQ10x) (rare, since most such hands would take-out-
double), you could be forced to ruff diamonds in hand, either being
overruffed by LHO (who may also be short) or be shortened in trumps to
the point where opponents have more trumps than you. Especially since
you have to use up your A, Q, 10 of hearts to ruff with.
OldPalooka
I believe your empty space arithmatic holds only if you can deal more
clubs into those empty spaces. If rho is hypothesized to hold
precisely 6 clubs, you have to deal the other 5 clubs into the 26
empty spaces of lho and cho as well as the remaining hearts etc.
Likewise if he hypothetically holds 5 diamonds you have to deal the
remaining 3 diamonds into their 26 empty spaces. This ignores that
opener is constrained in other suits, i.e., with 6 clubs he 'cannot'
hold 6 diamonds, or 7 spades or 6 hearts etc.
As a thought experiment, make N deals where rho holds 4 or 5
diamonds. Count the DD tricks in hearts for your side, then swap lho
and cho and count again. The positional value of partner's symmetric
holding over lho's should be significant. Repeat for rho holding 5 or
6 clubs. The difference in total tricks may be significant.
-- Bill Shutts
John Blubaugh
I happen to think Mike is correct and I used this method for years.
However, not everyone agrees with him. He and Frank Stewart had an
interesting round of debates about it in their columns in the
"Bulletin" years ago. I think that would have been sometime around
1990 or 1991. I once had a girl friend who was always quoting to me
what Mike would do in any instance. I got tired of this so I hired
Mike to play with her as a Christmas present at the year ending
regional in Reno many years ago just to find out what he would have to
say. They got along famously and they did well. Mike was wonderful to
her. But, I won the two pair events playing with two different
partners ;-)
I happen to think Mike is the best bridge writer of all time. But that
is just my opinion which I value highly ;-)
JB
Stephen Fischer
> In another thread, I stated that Mike Lawrence's reasoning
> supporting 4-card overcalls was suspect.
I agree that Lawrence's reasons do not support his theory, as you point out:
- partner is not more likely than LHO to have good hearts
- if partner does not have hearts, we're not terribly likely to have a
good place to play the hand
However, there are three reasons why I like the concept:
- A 4-3 fit generally plays extremely well here. If the opponents don't
lead trumps at trick one, it's not unusual to be able to score 6-7 trump
tricks even on a bad break.
- This is by far the safest time to show this suit, so you are likely to
lose it if you don't bid now.
- Your hand should have good values, and will not be a disappointment to
partner if they have a suit of their own.
In particular, the first point has worked in practice for many years.
Lawrence does allude to the play in his examples of hands that fit the
one you gave, but he spends too much time on the "expectation of fit", IMO.
He also gives other valid reasons:
- Opponents may misjudge their fit, both expecting shortage in partner's
hand
- It's always lead-directional
- Even a simple overcall reduces the precision in the opponents' auction
- They might not find their cold 3NT with xxx opposite xxx in your suit
metobillc
Those are irrelevant for determining how many hearts partner will
hold, which is at the (pardon) heart of the argument. The main points
were that from opener's bid, you have an estimate of how many of that
suit he holds (and may have an estimate of the maximum number he holds
in a different suit, but this is a small effect). The more cards he
holds in whatever suit he bids, the fewer empty spaces he has for
hearts. The hearts are as likely to be in any empty slot as any
other, so my method is exactly correct for the question I was
answering, which was how many hearts partner will hold on average
opposite my 4-card overcall (which is the same number as LHO will hold
on average).
> Likewise if he hypothetically holds 5 diamonds you have to deal the
> remaining 3 diamonds into their 26 empty spaces. This ignores that
> opener is constrained in other suits, i.e., with 6 clubs he 'cannot'
> hold 6 diamonds, or 7 spades or 6 hearts etc.
What you are saying is that we have a tiny bit more information than
just opener's average diamond length. We also know that he does not
have, e.g., 6 or more of any other suit. These distributions, though,
are extremely unlikely a priori, so neglecting them changes the
average number of hearts held by much less than 0.1%. An excellent
exposition of this principle is given in Kelsey & Glauert's "Bridge
Odds for Practical Players".
> As a thought experiment, make N deals where rho holds 4 or 5
> diamonds. Count the DD tricks in hearts for your side, then swap lho
> and cho and count again. The positional value of partner's symmetric
> holding over lho's should be significant. Repeat for rho holding 5 or
> 6 clubs. The difference in total tricks may be significant.
Yes, a more accurate analysis would look at not just the average
length of the trump fit, but at the par score in hearts for the
overcalling side. I can perform a simulation of that for the given
hand, but we would have to agree on a set of parameters for opener's
1D call in advance. I would be surprised, though, if having a shorter
fit on average will lead to better scores on average.
Bill Campbell
Andrew
Another reason is the positional value of honors. When you hold
length, your honors in RHO's suit are well placed. When you have
shortness, partner's honors in RHO's suit are often worthless.
Andrew
Kieran Dyke
"John Blubaugh" <jblu...@yahoo.com> wrote in message
news:788139b1-138f-4b98...@v57g2000hse.googlegroups.com...
JB
Just out of interest, what has Frank won?
Tiggrr
metobillc
Andrew
Yes, I agree, but that was not what Mr. Lawrence said. He said" Your
length in diamonds plus opener's length... means your partner is
likely to have heart support. The length in diamonds therefore is not
a minus but, rather, an asset." To me, this means that the
longer my diamonds are, the more length partner will have in hearts.
That is the claim that I have refuted. If partner having fewer hearts
on average is more than compensated by other factors, I'd be
surprised, but I'm not denying that possibility. As I said before,
there are plenty of good reasons to overcall, but shortness in
opener's suit should spur you on rather than deter you, and vice
versa.
Bill Campbell
Kieran Dyke
"metobillc" <meto...@gmail.com> wrote in message
news:b0c7e522-809b-48f6...@v13g2000pro.googlegroups.com...
Except, of course, that hands tend to play rather well when you have length
over length and shortage over shortage, so that you can be the overruffor
instead of the overruffee.
Tiggrr
Stephen Fischer
>
>
> Just out of interest, what has Frank won?
>
> Tiggrr
"Over 40 tournament events", according to the only description I could
find in his books. He's certainly better known as a bridge writer.
Carl
>
> I happen to think Mike is the best bridge writer of all time. But that
> is just my opinion which I value highly ;-)
>
> JB
He's certainly one of the most re-readable writers of all time. I
started on his material as a novice and got very useful lessons. When
I re-read the same books a few years later, they were like new
material to me. I could read another layer into the concepts.
Obviously there's still stuff in there waiting for me to get, when I'm
ready for it.
Very much like Kelsey in that respect. Neither of them are "easy
reads", but each book is like three books in one.
Cheers,
Carl
Andrew
If you read my earlier response to Raija in the original thread you
will see that I completely agree with you. Mike's contention that
length in RHO's suit increases partner's expected length in your suit
is wrong for all the reasons you state.
Mike is no mathematician and he would be the first one to admit that.
And although Mike got this one wrong, but he got so many others right
I think we can forgive him this error.
Andrew
Andrew
> "John Blubaugh" <jbluba...@yahoo.com> wrote in message
To my knowledge, Frank has not won any major bridge championships. But
he certainly is a fine bridge author who has written several good
books for advancing players.
Andrew
Will in New Haven
Belief and authority are no substitute for evidence and logic. I think
Lawrence is right that overcalling a four-card Major is often a
reasonable shot but not because partner is more likely to fit your
suit. It is right because WHEN partner fits your suit, he or she will
often be able to ruff the long cards in your RHO's suit without being
over-ruffed. And he's right because partner will have _length_ if he
bids another suit, which you won't be able to depend on if you make an
offshape takeout double. And he's right because low-level penalty
doubles are so out of favor that nothing bad will happen when partner
doesn't fit your suit, and you will have often directed a good lead.
However, his stated reasoning is mathematically incorrect. It's not
that I disagree with him or the OP disagrees with him. It is that the
math disagrees with him.
--
Will in New Haven
henry...@yahoo.com
I ran a small simulation (300 deals) in which I gave North the actual
hand and West the following:
12-20 hcps, 0-4 spades, 0-4 hearts, 3-13 diamonds, 0-13 clubs,
diamonds as long or longer than clubs, excluding 15-17 balanced.
I wanted to see how many hearts South would have in this sim, and here
are the results;
0-1 hearts = 21 deals, or 7%
2 hearts = 75 hands, or 25%
3 hearts = 90 hands, or 30%
4+ hearts = 114 hands, or 38%
Perhaps a bigger sim would come up with results substantially
different than this one, but from where I sit, the 1h overcall is
going to hit 3+ heart support about 60-ish percent of the time.
That would appear to support Lawrence's view, but one sim of 300 hands
is of course far from conclusive. Perhaps someone can run a longer
sim as a double check against these results.
Henrysun909
Will in New Haven
wrote:
I think there must be better mathematical methods than sims but I am
unfamiliar with these things in a bridge context. It is often, when
looking at card situations, possible to work out a percentage of times
when x, y or z will happen that is independant of simulations and
superior to any small simulation, certainly.
But it would be work, so I won't likely do it.
Will in New Haven
> "John Blubaugh" <jbluba...@yahoo.com> wrote in message
Turn off your targeting computer. Use the Mike.
I think Mike's _reasoning_ is wrong even though overcalling a four-
bagger with length in the opponent's suit is often right. This is a
brute-force math problem and I would not be shocked if Mike were
correct that you were more likely to fit your suit, for some value of
"more likely," but I would be surprised.
Andrew
wrote:
You can't draw that conclusion from one sim. Run another one with the
minor suit holdings reversed. If Bill is correct, then you will find a
fit slightly more often when overcaller has short diamonds, If Mike is
correct, you will find a fit more often with length in RHO's suit.
Andrew
henry...@yahoo.com
I think that part of the problem is going to be that the number of
available diamonds will have an effect on the likelihood of South
having 3+ hearts. To take an extreme example:
if West has 6 diamonds, thus leaving 2 diamonds outstanding, then East
and South are pulling cards from a deck that will have very few
diamonds and consequently very many spades, hearts, and clubs. But if
west is 4=4=3=2, then East and South will be pulling from a deck with
relatively few hearts (5) and relatively more spades and clubs.
So while I think that it could be done statistically for each red suit
permutation (West is 1=4, 2=4, 3=4, 4=4, 0=5, 1=5, 2=6, 3=5, 4=5, 0=6,
etc.) I think it would be a tedious calculation that might not give us
better average information that, say, a 1000 hand simulation. I only
chose 300 because DmP does not calcuate that particular statistic. I
suppose Deal might, if someone were to program it to list the number
of hearts held by South.
Henrysun909
Andrew
You don't have to. Bill has done the work at the top of this thread.
Will in New Haven
Looks pretty convincing to me. Much more convincing than "Mike
Lawrence said."
Andrew
Think of it this way. For the purpose of calculating partner's
estimated heart length the deck consists of two types of cards:
* Hearts
* Non-hearts
There is nothing special about the diamond suit. Diamonds are simply
non-hearts. From our own hand, we know the location of 5 hearts and 8
non-hearts. No matter what distribution we hold in spades,clubs and
diamonds our hand will still have 8-non-hearts so the fact that we
hold 5 diamonds provides no more useful information than if we held 5-
clubs (for the purpose of calculating fit).
We have one additional significant piece of information: RHO is known
to hold 4+ diamonds (e.g., 4+ non-hearts). This information slightly
increases the number of hearts we can expect to find with both LHO and
partner. It stands to reason that if holding 4+ diamonds increases our
likely heart fit, then 6 diamonds, would make the inference even
stronger. And so it is. If RHO opens 2D, 3D, or 4D, the inference that
partner has a fit for our 5-card heart suit gets increasingly strong.
But note--this effect has nothing to do with how many diamonds we hold
in our hand. No matter how many diamonds we hold, the more diamonds
RHO shows, the more likely it is that we have a heart fit.
So what is the impact of our diamond length? it impacts the diamonds
we can expect to find in RHO's hand. The more diamonds we hold, the
less likely it becomes that RHO has extra diamond length.
--When we have a singleton diamond, RHO is a favorite to hold 5+
diamonds (increasing our chance of a heart fit).
--When we have 5 diamonds, RHO is a favorite to have minimum length
(4-5 diamonds) decreasing our chance of catching a fit.
So when we have length in RHO's suit, RHO rates to be shorter in that
suit, and correspondingly longer in hearts, hence our fit is slightly
reduced.
Andrew
> To take an extreme example:
>
> if West has 6 diamonds, thus leaving 2 diamonds outstanding, then East
> and South are pulling cards from a deck that will have very few
> diamonds and consequently very many spades, hearts, and clubs. But if
> west is 4=4=3=2, then East and South will be pulling from a deck with
> relatively few hearts (5) and relatively more spades and clubs.
>
> So while I think that it could be done statistically for each red suit
> permutation (West is 1=4, 2=4, 3=4, 4=4, 0=5, 1=5, 2=6, 3=5, 4=5, 0=6,
> etc.) I think it would be a tedious calculation that might not give us
> better average information that, say, a 1000 hand simulation. I only
> chose 300 because DmP does not calcuate that particular statistic. I
> suppose Deal might, if someone were to program it to list the number
> of hearts held by South.
Bill's calculation is a good approximation of the reality. No other
calculation is needed.
Andrew
Player
Totally disagree with this comment. I have played this style of
overcall for a long time and have found it to be very effective.
Lawrence was certainly correct in his statement.
Eddie Grove
> You can't draw that conclusion from one sim. Run another one with the
> minor suit holdings reversed. If Bill is correct, then you will find a
> fit slightly more often when overcaller has short diamonds, If Mike is
> correct, you will find a fit more often with length in RHO's suit.
I ran these sims many years ago. The first thing you have to do is to decide
under which circumstances opener bids 1C vs 1D with 4=4 or 3=3 in the minors.
That makes a difference.
I ran 4 sims holding one of Lawrence's example hands. In one pair, I kept the
hand fixed and compared avg holding of partner comparing opening bid of 1C vs
1D. Then I changed the hand, swapping clubs and diamonds, and ran the pair of
sims vs 1C and 1D openers again. In one case length in opener's minor
increased expected fit by .1 and in the other it decreased the expected length
by .1, and the difference was all about the 3=3 and 4=4 hands.
Basically, the more diamonds you have when opener opens 1D, the more likely
that he opened with 4=4=3=2 distribution, in which case there are only 5
hearts to split between LHO and CHO. If the suit is clubs, opener could be
3=4=3=3 or 4=3=3=3 as well as 4=4=2=3 so there are fewer forced cards in the
3cm opener's hand in your major when opening 1C vs 1D. The 4=4 hands matter
too, but I no longer remember how.
In any event, a difference of about .1 out of about 3 expected support is no
big deal. I.e. about 1 hand in 30 partner holds 1 card more or less of
support. The other effects dealing with the placement of cards and underruffs
or overruffs will be more relevant.
I'll happily take a bet that partner's support does not change much, but I
would be unwilling to bet about changes in the double-dummy par contract.
Eddie
Eddie Grove
> In any event, a difference of about .1 out of about 3 expected support is no
> big deal. I.e. about 1 hand in 30 partner holds 1 card more or less of
Aargh. That should be 1 out of 10.
Bill Jacobs
856600...@d1g2000hsg.googlegroups.com:
> On Aug 18, 4:58 pm, John Blubaugh <jbluba...@yahoo.com> wrote:
>
>>
>> I happen to think Mike is the best bridge writer of all time. But that
>> is just my opinion which I value highly ;-)
>>
>> JB
>
> He's certainly one of the most re-readable writers of all time. I
> started on his material as a novice and got very useful lessons. When
> I re-read the same books a few years later, they were like new
> material to me. I could read another layer into the concepts.
> Obviously there's still stuff in there waiting for me to get, when I'm
> ready for it.
>
A great writer, I agree.
As for the 4-card-overcall topic, there's a discussion without a solution!
By the way, if you are talking about great writers, you must read Phillip
and Robert King's "The Kings Tales".
One of the chapters is a takeoff of Lawrence, advocating 3-card overcalls
and 2-card overcalls. Totally hilarious. It ends with "as for overcalling
on a singleton, the world is not ready for it ... yet".
Cheers ... Bill
Andrew
> Andrew <agump...@gmail.com> writes:
> > You can't draw that conclusion from one sim. Run another one with the
> > minor suit holdings reversed. If Bill is correct, then you will find a
> > fit slightly more often when overcaller has short diamonds, If Mike is
> > correct, you will find a fit more often with length in RHO's suit.
>
> I ran these sims many years ago. The first thing you have to do is to decide
> under which circumstances opener bids 1C vs 1D with 4=4 or 3=3 in the minors.
> That makes a difference.
>
> I ran 4 sims holding one of Lawrence's example hands. In one pair, I kept the
> hand fixed and compared avg holding of partner comparing opening bid of 1C vs
> 1D. Then I changed the hand, swapping clubs and diamonds, and ran the pair of
> sims vs 1C and 1D openers again. In one case length in opener's minor
> increased expected fit by .1 and in the other it decreased the expected length
> by .1, and the difference was all about the 3=3 and 4=4 hands.
>
> Basically, the more diamonds you have when opener opens 1D, the more likely
> that he opened with 4=4=3=2 distribution, in which case there are only 5
> hearts to split between LHO and CHO.
In other words, the driver for your fit is RHO's shape. The more
diamonds you can expect RHO to hold, the bigger fit you can expect to
have.
Your own diamond length affects the heart fit only indirectly. Our
diamond length reduces RHO's expected diamond length, and therefore
also decreases the fit we can expect to find.
> If the suit is clubs, opener could be
> 3=4=3=3 or 4=3=3=3 as well as 4=4=2=3 so there are fewer forced cards in the
> 3cm opener's hand in your major when opening 1C vs 1D. The 4=4 hands matter
> too, but I no longer remember how.
>
> In any event, a difference of about .1 out of about 3 expected support is no
> big deal. I.e. about 1 hand in 30 partner holds 1 card more or less of
> support. The other effects dealing with the placement of cards and underruffs
> or overruffs will be more relevant.
>
> I'll happily take a bet that partner's support does not change much, but I
> would be unwilling to bet about changes in the double-dummy par contract.
I would lay odds you are correct.
Andrew
raija d
"Andrew" <agum...@gmail.com> wrote in message
news:c6620b96-8bff-40d2...@w1g2000prk.googlegroups.com...
I don't know Mike. I don't know Andrew. I don't know Bill Campbell. I don't
know how to do a sim.
Given a choice what to believe, I place my bet on Mike Lawrence, fully
expecting that if a sim is performed, it will support his opinion.
If his theory is "wrong", surely he would have published later works
refuting it himself. Or, someone else would have refuted the theory; the
book "Overcalls" is after all from 28 years ago.
Raija
metobillc
> "Andrew" <agump...@gmail.com> wrote in message
This is why I asked you exactly which claim of Mr. Lawrence's you
wished me to refute. You gave me carte blanche, and I took the quote
directly from his book, and showed you the argument that disproves the
claim that length in openers suit enhances partner's support for your
suit. You have shown no flaw in my argument, just a refusal to be
willing to learn, or to try to understand an argument.
> If his theory is "wrong", surely he would have published later works
> refuting it himself. Or, someone else would have refuted the theory; the
> book "Overcalls" is after all from 28 years ago.
>
> Raija
You are treating this like religion. It's not a question of belief or
not, it's mathematics. If you cannot follow the argument given, then
you need to rely on your belief. But the argument is independent of
who's making it.
You are naive to believe that "surely he would have published later
works refuting it himself". He hasn't come out with "Overcalls 2",
and it isn't really discussed much in his "Competitive Bidding".
Other authors have refuted the theory, notably Larry Cohen in "To Bid
or Not to Bid" and Hugh Kelsey in several of his works. Citing
authority is about the weakest form of knowledge there is.
Bill Campbell
OldPalooka
>
> - Show quoted text -
No, what I am saying is if you know or assume the length of opener's
suit, the remaining cards of that suit need to take up spaces in the
other two hands. Otherwise you end up with the absurd conclusion on
the example deal that if RHO opened 1C on a 3 card suit, he will
average less than 3 hearts, because he only has 10 empty spaces and
the other two hands have 13 each. That will only be true if he is
known to hold *at least* 3 clubs. If he holds precisely 3 clubs,
eight of the 26 empty spaces in the other two hands must contain clubs
so there are only 18 empty spaces available for hearts, diamonds, and
spades.
Since you used a *typical length* and not a *minimum length* in your
diamond example [I think that is the right approach], you neglected to
account for the other 3 diamonds when computing expected heart length
around the table.
-- Bill Shutts
brsri...@gmail.com
I would rather stick with Mike regarding evidence and logic if I were
you:)
Boris
Eddie Grove
> In other words, the driver for your fit is RHO's shape. The more
> diamonds you can expect RHO to hold, the bigger fit you can expect to
> have.
That's at least mostly true. It might be completely true, I don't know, but the
correlations between hcp and shape and 1N openers and choice with equal minor
lengths may cause some extremely rare exceptions. E.g., even before you look at
your hand, the average number of hearts held by a 1C opener is *not* precisely
equal to the average number of hearts held by a 1D opener. Things are messy.
If we assume RHO dealt and opened and note the cards in our hand,
including a 4 card heart suit, one thing you can count on is
E[hearts in CHO] = 1/2 * (13 - 4 - E[hearts in RHO])
where the expectations are conditioned by the cards in our hand and what the
particular opening bid means in their system.
E[hearts in RHO] does vary with our length in his suit, but I remember being
surprised how little it was. The difference in heart length just between
opening 1C vs 1D may be a bigger difference than whether we hold 2 vs 5 cards
in opener's suit. I wish I could remember the details of my sims so long ago.
Eddie
Jari Boling
Standard 5 card major longer minor 1D opening with north and the predealt
hand S82, HAQT7, DT7654, CAQ with east. West had the following heart
lengths:
0: 0.8515%
1: 7.3301%
2: 22.4055%
3: 32.1373%
4: 24.3261%
5: 10.2196%
6: 2.3952%
7: 0.3137%
8: 0.0203%
9: 0.0007%
On average 3.18 cards, likelihood for 3+ card fit is 70.7%.
And then I changed east hand to S82, HAQT7, DAQ, CT7654, and the results are
0: 0.7979%
1: 6.8861%
2: 21.5744%
3: 31.9372%
4: 24.9546%
5: 10.8362%
6: 2.6568%
7: 0.3393%
8: 0.0169%
9: 0.0006%
On average 3.14 cards, likelihood for 3+ card fit is 69.4%.
So diamond length seems to actually decrease the likelihood for a heart fit,
but the difference is very small.
Andrew
Mike is not a scholar whose purpose in life is to produce absolutely
accurate scholarly works. He is a professional author who writes
bridge books for the masses to earn a living. He has not updated
Overcalls for a practical reason: it would not help him sell any more
copies. He told it to me himself.
He has not refuted his error probably because he is not convinced that
he is wrong on this point, and since he is not particularly
mathematically inclined, I doubt he'd have the patience to read
through the argument.
Andrew
Andrew
Jari Boling
average 3.18 hearts, and there is a 70.7% likelihood for a 3+ card support,
and with 5 diamonds the numbers are 3.14 and 69.4% respectively.
brsri...@gmail.com
Since we are all so mathematically inclined:), what are the chances of
both partnerships finding a fit (any fit) after you happen to find
length in the opponent's suit (diamonds in this case)?
Boris
Dave Flower
> � In another thread, I stated that Mike Lawrence's reasoning
> for clarification, so I wanted to present some arguments that show his
> claim is false. �I have a lot of respect for Mr. Lawrence's writing,
> and have read almost all of the bridge books he has written. �In
> almost all cases, I have found them to be well-reasoned, with the
> exception of the claim below from his book "The Complete Book on
> Overcalls in Contract Bridge".
>
> 1D: �S 8 2
> � � �H A Q T 7
> � � �D T 7 6 5 4
> � � �C A Q
>
> In the example, your RHO opens 1D, and you hold the hand above. �Mr.
> Lawrence recommends overcalling 1H on the good 4-card suit, reasoning
> as follows:
>
> � "If the possession of five cards in the suit opened bothers you,
> that your partner (and LHO) are also short. �This means your partner
> is likely to have heart support. �The length in diamonds therefore is
> not a minus but, rather, an asset."
>
> between the two unknown hands, that of your partner and your LHO.
> Anything that is true for one of them must be true for the other, as
What worries me about 4-card overcalls is either partner overbids,
believing you to have at least five, or that partner underbids when
you have a 5+ card overcall, suspecting that you may hold four!
Dave Flower
Will in New Haven
Right. "Give up your mind; an authority-figure has spoken."
brsri...@gmail.com
You love to criticize because you think of yourself as an authority-
figure. At least I know I am not. As for the discussion raised upon
four card major overcalls if we don't have a fit (any fit) after
overcalling with our four card major and length in their minor then
how many times our opponents will have a fit when this happens? I am
asking you this because if we don't have a fit and they don't have a
fit too then there will be no competition and we have enough bidding
room and bidding gadgets, like the cue-raise, the 2N raise or jump-cue
raise showing 4 card support, for finding out whether partner is
raising with 3 or four card support. We can stop in 2 of a major or 2N
or in the worst case in 3 of the overcalled major or perhaps we will
find an excellent Moysian fit or 3N. I really think that Mr. Lawrence
is in fact right.
And yes he is an authority and I am not. Does it hurt that much.
Boris
Will in New Haven
The math, as shown by the original poster, is rather simple and it is
correct. If someone very important and authorative says "A is not A,"
you can believe him if you want. I will believe the math.
That doesn't mean that four-card overcalls are wrong. It means that
Lawrence's reasoning is incorrect. You are not more likely to find
partner with a fit. However, when he does have a fit the hand will
play very well. And you are starting out at the one-level, can often
escape unscathed when you don't have a fit. You are getting into the
auction, directing a lead, might find _another_ fit because you have
shown some sign of life. Except for the premise that you are more
likely to find a fit for your suit, which is incorrect, the case for
four-card overcalls at the one-level is strong.
Laughing at believers never hurts.
dak...@aol.com
likelihood? LHO/partner 4-4 is a fit 3-4 is a fit 2-4 is a fit, 4-3
is a fit. The fit IS likely which is ML's claim. The length in
opener's suit increases the fit playing well --a common meaning of
"fit" verses YOUR meaning as "number of pieces" That is what I took
"asset"(ML) to mean. Didn't you?
brsri...@gmail.com
Right, in that case the entire discussion about whether it is
theoretically right and when to overcall on a four card major is
completely out of place because there is nothing one can gain from
knowing that Lawrence's statement that you are more likely to find a
fit in the aforementioned case is wrong. The fact is that overcalling
when holding a solid 4 card major and length in the opener's minor is
probably the right thing to do.
Boris
Will in New Haven
Indeed. It is very likely a win or break-even play. And every time it
keeps one from making an offshape takeout double it has improved ones
bidding. It is a rather minor point that one shouldn't be so
optimistic about finding a fit.
Dave Flower's point, that the four-card overcall may lead to
difficulty when partner plays you for five and may lead to a problem
when you _have_ five and partner plays you for four, is a serious
argument against four-card suit overcalls. The math argument isn't an
argument against the overcalls. It is an argument against one of the
reasons Lawrence gives.
brsri...@gmail.com
In this case I have to apologize for the spoken words. One can reason
with reason :) The point was the discussion was pretty one-sided and I
didn't liked it very much.
Boris
Will in New Haven
No need to apologize. Nice talking with you.
Andrew
Thank you for your simulations. The simulation suggests that while
Bill is correct that length in RHO's suit decreases the chance of
catching a fit--the effect is so small it can be ignored. Therefore
the important thing about length in RHO's suit is its effect on
playing strength.
Andrew
raija d
"Jari Boling" <jboling AT abo DOT fi> wrote in message
news:48aa...@newsflash.abo.fi...
> Got avereges and likelihoods mixed up, sorry! With 2 diamonds partner has
> on average 3.18 hearts, and there is a 70.7% likelihood for a 3+ card
> support, and with 5 diamonds the numbers are 3.14 and 69.4% respectively.
>
>
Thank you for the sim. It appears that for all practical purposes the
difference is very small.
Raija
Player
<bill.re...@taylorandfrancis.com> wrote:
snipped
> Dave Flower's point, that the four-card overcall may lead to
> difficulty when partner plays you for five and may lead to a problem
> when you _have_ five and partner plays you for four, is a serious
> argument against four-card suit overcalls. The math argument isn't an
> argument against the overcalls. It is an argument against one of the
> reasons Lawrence gives.
>
> --
> Will in New Haven
Dave's argument is totally spurious. Presumably if you play this
style, partner is aware that 4 card overcalls are possible. Presumably
then you also have the mwherewithall to find out if it is a 5 or 4
card overcall. I mean really!
Wayne
Don't the numbers you give suggest the opposite of your conclusion.
With five diamonds we had a 3.18 average (3+ 70%) with two diamonds
we had a 3.14 average (3+ 69%).
More diamonds in east produced on average slightly more hearts in west
and a higher percentage chance of a fit.
Not "diamond length seems to actually decrease the likelihood for a
heart fit"
diamond length seems to increase the likelihood for a heart fit.
Am I missing something?
I am not convinced about this data - I might run a simulation of my
own.
Wayne
Ivan Popivanov
wrote:
> hand and West the following:
>
> 12-20 hcps, 0-4 spades, 0-4 hearts, 3-13 diamonds, 0-13 clubs,
> diamonds as long or longer than clubs, excluding 15-17 balanced.
>
> I wanted to see how many hearts South would have in this sim, and here
> are the results;
>
> 0-1 hearts = 21 deals, or 7%
> 2 hearts = 75 hands, or 25%
> 3 hearts = 90 hands, or 30%
> 4+ hearts = 114 hands, or 38%
>
> Perhaps a bigger sim would come up with results substantially
> different than this one, but from where I sit, the 1h overcall is
> going to hit 3+ heart support about 60-ish percent of the time.
>
> That would appear to support Lawrence's view, but one sim of 300 hands
> is of course far from conclusive. Perhaps someone can run a longer
> sim as a double check against these results.
>
> Henrysun909
The results of a bigger simulation confirm what has been already
suggested in this thread.
The goal of my first simulation was to verify the claim that partner
is more likely or not to hold a H fit. For that, I run two one million
simulations under the following two conditions:
1. West had the original hand: 82 AQT7 T7654 AQ
2. South has opened 1m - 1C for the first simulation and 1D for the
second
1D Opening 1C Opening
0 8,637 0.86% 8,094 0.81%
1 73,717 7.37% 70,436 7.04%
2 223,941 22.39% 218,607 21.86%
3 321,400 32.14% 321,655 32.17%
4 243,481 24.35% 248,195 24.82%
5 101,606 10.16% 104,832 10.48%
6+ 27,218 2.72% 28,181 2.82%
Looking at the results, to me there is no significant difference
whether the opening was 1C or 1D. The stability of the results
confirms what was suggested earlier - it's the number of the "non-
heart" cards and not the length of a particular suit that matters.
With my next experiment I tried to estimate how well the partnership
fairs in hearts on average under the same conditions. For the first
simulation I added one more condition to the above:
3. East has exactly 4 hearts, thus EW have an 8-card fit.
If South has opened 1D, EW had a slightly better trick average: 8.55
vs 8.42 over 10,000 hands.
The results were similar if I restricted East to 7 HCP. Although
insignificant, this 0.13 of a trick difference surprised me - I would
expect that having length in the opener's suit will reduce the trick
taking potential of EW because the diamond suit is badly broken more
often than not. The only reasonable explanation I can come up with is
that it's easy to ruff the diamonds in dummy and this over compensates
for the possible uneven breaks in the diamond suit. This was also
already suggested in this thread.
Finally, to estimate the pure trick taking potential in hearts, I came
back to the original two conditions:
1. West had the original hand: 82 AQT7 T7654 AQ
2. South has opened 1m - 1C for the first simulation and 1D for the
second
but this time generating 10,000 hands for the 1C and 1D opening only
and counting the average number of heart tricks. The results were 7.45
tricks for the 1D opening, and 7.58 for the 1C opening. Comparing this
with the previous results, when EW did have a fit, shows that EW fair
better in hearts if South has opened 1C than 1D.
One might argue that the last difference is also insignificant.
However, it makes more sense when other bridge factors are added -
bidding 1H on this hand wins mostly when we can play in H - we have no
second suit and not enough strength to be too happy in NT contract.
Thus, we are hoping that partner has a fit for us. Now, considering
the fact that the long diamonds do not really improve the success of a
H contract, I think the conclusion is that it is better to PASS than
to overcall a 1D opening.
Any other interpretation of the these data? Are there any other
simulations that might shed more light on this situation?
Cheers,
Ivan
Dave Flower
I do not recall any published method of determining whether partner
has a 4-card overcall. Especially as a decision may have top be made
at a high level.
Dave Flower
Jari Boling
correction and the original conclusion should be right.
Probably so that longer diamond increases the likelihood for opener to have
short diamonds, which increaseses openers likelihood for having more hearts
(and reduces partners likelihood for having hearts). This was the claim of
the first post as well. It is easy to check the average number of diamonds
for opener in both cases, it is 4.5 if we have 5 diamonds and 4.7 if we have
2 diamonds. Compared with the numbers 5 and 6 respectively used in the
original analysis. Estimating openers expected heart length is equally
simple, it is 2.6 with 2 diamonds and 2.7 with 5 diamonds. Using vacant
places on this does not give the same expected heart lenght, but that
probabaly is due to not taking into account the constraints from the
bidding. For example a 1D opener rarely has 5 hearts, for the simulation I
have assumed that opener always opens 1H with 5H and 5+D. Checked that one
as well, opening 1D with 5M-6D and 15+ hcp did not change anything
significantly.
Well all these numbers were obtained with the same program I used for the
first analysis, everything naturally depends on that I got it right
specified. So please feel free to doublecheck my numbers! I assumed by the
way that opener opens 1C with 3-3 and 4-4 in the minors, and the higher suit
with 5-5 or better. For example assuming 4+D for a 1D opener increases the
likelihoods for a heart fit, but the differences are about the same
(3.23-3.19 and 72.1-71.1).
"Wayne" <wjbu...@gmail.com> wrote in message
news:d5758c48-8d38-4222...@n33g2000pri.googlegroups.com...
OldPalooka
>
> - Show quoted text -
Random thoughts:
For each deal generated swap North and East hands and replay.
Getting a lead in opener's suit is supposed to be part of declarer's
advantage. Specify a lead in opener's suit as well as the double
dummy opening lead. Averaging the result between the two leads might
be illuminating.
Change West to AQ AQTx Txxxx xx and compare results with 1C opener.
-- Bill Shutts
Eddie Grove
> the first post as well. It is easy to check the average number of diamonds
> for opener in both cases, it is 4.5 if we have 5 diamonds and 4.7 if we have
> 2 diamonds.
> way that opener opens 1C with 3-3 and 4-4 in the minors, and the higher suit
If you assume opener bids 1D with 4-4 minors that means your constraints force
opener to have 5 diamonds unless 4=4=3=2 or 4=4=4=1. I think you would see a
somewhat bigger difference if opening 1D on 4-4 minors.
Are you assuming a strong or weak 1N opener? That also makes a difference.
Eddie
Jari Boling
news:87myj8d...@hotmail.com...
> "Jari Boling" <jboling AT abo DOT fi> writes:
>
>> the first post as well. It is easy to check the average number of
>> diamonds
>> for opener in both cases, it is 4.5 if we have 5 diamonds and 4.7 if we
>> have
>> 2 diamonds.
>
>> way that opener opens 1C with 3-3 and 4-4 in the minors, and the higher
>> suit
>
> If you assume opener bids 1D with 4-4 minors that means your constraints
> force
> opener to have 5 diamonds unless 4=4=3=2 or 4=4=4=1. I think you would
> see a
> somewhat bigger difference if opening 1D on 4-4 minors.
Didn't try that one, here are the results:
With 5 diamonds:
partner: 3.24 hearts, 72.5% likelihood for 3+H
opener: 2.51 hearts, 4.52 diamonds
With 2 diamonds:
partner: 3.26 hearts, 72.9% likelihood for 3+H
opener: 2.48 hearts, 4.72 diamonds
Didn't give as many digits earlier, so lets repeat for clarity. With longer
minor, meaning 1C with minors 3-3 and 4-4:
With 5 diamonds:
partner: 3.14 hearts, 69.5% likelihood for 3+H
opener: 2.72 hearts, 4.52 diamonds
With 2 diamonds:
partner: 3.18 hearts, 70.8% likelihood for 3+H
opener: 2.63 hearts, 4.73 diamonds
And I also did with 4+D but 1C with minors 4-4:
With 5 diamonds:
partner: 3.19 hearts, 72.2% likelihood for 3+H
opener: 2.61 hearts, 4.66 diamonds
With 2 diamonds:
partner: 3.23 hearts, 73.2% likelihood for 3+H
opener: 2.54 hearts, 4.83 diamonds
In all cases opener can have 5+H if longer D and 15+hcp, I think there was
some minor changes in some numbers. And of course it was a different run.
Note the small inconsistency between first two and last two cases regarding
partners heart holding. It is possible if partner has more often exactly 3
hearts in the last two cases, can't figure out why it would be so.
> Are you assuming a strong or weak 1N opener? That also makes a
> difference.
15-17 1NT opener.
OldPalooka
Actually the club length with opener 4-4 or 4-5 (depending on that
assumption) drives the expected heart length for opener down along
with the expected diamond length. I looked at combination frequencies
(which is also an incomplete assumption). Where I assumed a clubs
first strategy in common situations, opening clubs on 4-4 yields ~4.8
clubs vs. 4.5 diamonds. Opening diamonds on 4-4 drops the mean
diamond length but increases the mean minor length enough to drop the
expected heart holding by almost 0.2. Also if 1D promises 4, expected
heart length drops by another .1.
-- Bill Shutts
vsp...@hotmail.com
Diamond length should decrease the likelihood of three card support.
The length increases the chances of opener starting with 4=4=3=2.
Jari Boling
news:bf31365d-40f5-4915...@j1g2000prb.googlegroups.com...
You are right about that, if East has 5 diamonds the likelihood for opener
(North) having 4=4=3=2 is 6.5%, and if East has 2 diamonds the same
likelihood is 5.2%. But it seems as it is not enough, the likelihood for
having short hearts in the remaining 93.5% hands is probably increased, as
East's second longest suit is hearts.
The likelihood for 4=4=3=2 when opening 1D, without any constraints on the
other hands, is 4.5%. The reason why it is higher than that with the hand
with 2 diamonds is probably because it has 5 clubs. The two different hands
compared were S82, HAQT7, DT7654, CAQ and S82, HAQT7, DAQ, CT7654.
John R. Mayne
Good thread! FWIW, I've been making four-card overcalls with
substantial length in opener's suiit for years now, and I've found the
playability issues to work out well. This means I play more 4-3 fits
than others, but I play a fair number of other 4-3 fits, too.
I don't remember playing a 4-2 fit on this auction type, but it surely
could happen. The good things happen far more often, and I don't
recall ever getting shot where I stand in doing this (though that may
be failure of recollection.)
I think the math arguments made by the OP and backed by various other
posters are absolutely compelling, but this won't deter me from
continuing my four-card felonies.
--JRM
>
> Andrew
Steve Willner
> Mike {Lawrence] is no mathematician and he would be the first one to admit that.
Is that true? I was once told he is something of a "math geek," but the
person may have been mistaken. Of course even a math geek can make a
small error, especially about something that is only a side issue to the
main point he is writing about.
rubberduck
metobillc wrote:
> In another thread, I stated that Mike Lawrence's reasoning
> supporting 4-card overcalls was suspect. I received some email asking
> for clarification, so I wanted to present some arguments that show his
> claim is false. I have a lot of respect for Mr. Lawrence's writing,
> and have read almost all of the bridge books he has written. In
> almost all cases, I have found them to be well-reasoned, with the
> exception of the claim below from his book "The Complete Book on
> Overcalls in Contract Bridge".
>
> 1D: S 8 2
> H A Q T 7
> D T 7 6 5 4
> C A Q
>
> In the example, your RHO opens 1D, and you hold the hand above. Mr.
> Lawrence recommends overcalling 1H on the good 4-card suit, reasoning
> as follows:
>
> "If the possession of five cards in the suit opened bothers you,
> forget it. Your length in diamonds plus opener's length assures you
> that your partner (and LHO) are also short. This means your partner
> is likely to have heart support. The length in diamonds therefore is
> not a minus but, rather, an asset."
>
<snip>
A merit of overcalling the four card heart sut is that hearts are
likely to be our best strain especially when partner is a passed hand
(no 6 card spade suit or club pre-empt). Flesh and blood opponents
often do not find the optimum defence on these hands.
Rubber
Andrew
> Andrew wrote:
> > Mike {Lawrence] is no mathematician and he would be the first one to admit that.
>
> Is that true? I was once told he is something of a "math geek," but the
> person may have been mistaken.
That person must have been thinking of someone else. In 20 years, I
have never seen Mike show an interest in math.
Andrew
rubberduck
In another thread, there is a discussion on ML's theory on 4 card
overcalls. Much discussion follows based on the following hand and
ML's statement
S 8 2
H A Q T 7
D T 7 6 5 4
C A Q
In the example, your RHO opens 1D, and you hold the hand above. Mr.
Lawrence recommends overcalling 1H on the good 4-card suit, reasoning
as follows:
"If the possession of five cards in the suit opened bothers you,
forget it. Your length in diamonds plus opener's length assures you
that your partner (and LHO) are also short. This means your partner
is likely to have heart support. The length in diamonds therefore is
not a minus but, rather, an asset."
From the beginning...
You pick up the above hand (2425 shape) and ask yourself two
questions
1a) how many hearts does partner have ? and b) what is your best fit ?
a1) There are 9 missing hearts distributed between 3 hands so partner
is likely to have 3.
b1) Your best fit is likely to be in diamonds.
2) Assume RHO opens 1st in hand, does this change the answer to
either question.
Yes, RHO bid removes a number of non-hearts, at the extreme if RHO
opens with a bid promising 10 non-hearts (e.g a 2NT bid showing 55
minors) this increases partners expected heart holding. If RHO
promises 5 non-hearts (e.g 1Spade) this also increases partners
expected heart holding - but by less than the previous 2NT example).
In the case of some nebulous bids (precision 1D, some 1C openings) it
is hard to guage the change in partners heart number - but it will
change even for 1NT (promising 6 non-hearts and 2 hearts)
so question 2a) how many hearts does partner have ? and b) what is
your best fit ?
a2) Depends on the bid, the more non-hearts it shows then the greater
partners expected number of hearts
b2) Also depends on the bid, but will usually be diamonds - unless RHO
has bid diamonds - in which case it will be hearts!
So ML is right and the sims in the thread are right., RHO bid of 1D
increases your partners expected heart holding - but so would most
bids by RHO. More importantly, RHO bid of 1D increases the chances
that hearts are your combined longest suit.
3) Assume partner passes 1st in hand, does that change the answers to
either question ?
Yes, assume partner would pre-empt with suitable hands this decreases
the number of shapely hands he is likely to hold and as he started off
holding more shapely hands with black suits than with red suits then
this will slightly increase the expected number of hearts.
Question 3a) how many hearts does partner have ? and b) what is your
best fit ?
a3) Partner likely to hold 3 and a fraction hearts
b3) Diamonds still best fit.
Thus, if you are third in hand after Pass-(1D) then you have good
chances of hitting 3+ hearts opposite and good chances that hearts
represent your best contract. Finally, you will have hit your sides
best fit before the opponents have found theirs in either clubs or
spades.
However, after Pass-(1C)- you have just as much chance of hitting 3+
hearts as over 1D, but now you may be missing your best contract in
diamonds and the opppoenents may already have found their club fit.
Rubber
Rubber
vsp...@hotmail.com
This Sunday
dlr: N, Vul E-W
South holds
S AKQT
H T
D T432
C J843
N E S W
p - 1D - 1S - 4H
4S- 6H - all pass
6H +7
N-S got 4 out of 50.
Andrew
You are right that the overcall and raise made it easier for the
opponents to bid 6H. Opener held: Jxxx, A, AKQJx, Axx. But I don't
think thats a slam on 4-card overcalls--its a general weakness of an
aggressive overcalling style that occasionally your bids allow the
opponents to diagnose a great fit. N-S has an 1100 sac in 6S, which
did not help their score.
Kieran Dyke
"Andrew" <agum...@gmail.com> wrote in message
news:803c429e-77f3-4a46...@t1g2000pra.googlegroups.com...
And if he had 3 small spades, he might have diagnosed shortage opposite, and
watched the opponents cash two (or three!) spades.
Tiggrr