Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Adjusting the Work Count by 3rds
153 views
Skip to first unread message
Fred.
Mar 8, 2012, 9:51:07 AM3/8/12
to
One adjustment to the Work count which maintains the 40
point total is to add 2/3 of a point for each ace and
take off 1/3 of a point for each queen or jack. This
makes it roughly propotional to a 6-4-2-1 count.
Do people feel that this is appropriate for no-trump
as well as suit bidding? That is, is
A32
K432
A432
K3
a strong no-trump (15-1/3) while
QJ6
QJ7
QJ8
KQJ5
ia a weak no-trump (12-1/3) at best?
Personally, when partner opens a weak
no-trump (12-14) in 1st or 2nd I expect
more controls than are in the latter hand.
Thoughts?
Fred.
point total is to add 2/3 of a point for each ace and
take off 1/3 of a point for each queen or jack. This
makes it roughly propotional to a 6-4-2-1 count.
Do people feel that this is appropriate for no-trump
as well as suit bidding? That is, is
A32
K432
A432
K3
a strong no-trump (15-1/3) while
QJ6
QJ7
QJ8
KQJ5
ia a weak no-trump (12-1/3) at best?
Personally, when partner opens a weak
no-trump (12-14) in 1st or 2nd I expect
more controls than are in the latter hand.
Thoughts?
Fred.
Bertil
Mar 8, 2012, 10:18:16 AM3/8/12
to
Thus the 1st hand is worth 20 pts and the 2nd would count as 16.
Bertil
Larry
Mar 8, 2012, 11:36:33 AM3/8/12
to
2/3 adjustment is too high IMO.
I prefer to use:
A = 4.5
K = 3
Q = 1.5
J = 0.5
T = 0.5
Still adds up to 10.
I prefer to use:
A = 4.5
K = 3
Q = 1.5
J = 0.5
T = 0.5
Still adds up to 10.
Carl
Mar 8, 2012, 12:17:34 PM3/8/12
to
On Mar 8, 9:51 am, "Fred." <ghrno-goo...@yahoo.com> wrote:
aces, but it has no long suit points and is totally devoid of
intermediates. It is probably worth what it started with, 14 HCPs.
The second hand appears to be de-valued properly. One could argue it
needs another something knocked off for being aceless. I would open it
because I can bid a four card suit. Move the cards around so I have to
open a 3-card minor and I very well might pass.
Opening the first hand 1NT is optimistic but not foolhardy.
Opening the second hand 1NT is not something I'd encourage.
Co Wiersma
Mar 8, 2012, 1:09:16 PM3/8/12
to
Op 8-3-2012 15:51, Fred. schreef:
But I would open it 1D
and certainly not 1NT
If partner answers with 1H, I would rebid 2H, but I have sympathy for 3H
>
> QJ6
> QJ7
> QJ8
> KQJ5
>
This I would not open a 15-17 NT
But I dont consider pass an option any close
Co Wiersma
> One adjustment to the Work count which maintains the 40
> point total is to add 2/3 of a point for each ace and
> take off 1/3 of a point for each queen or jack. This
> makes it roughly propotional to a 6-4-2-1 count.
>
> Do people feel that this is appropriate for no-trump
> as well as suit bidding? That is, is
>
> A32
> K432
> A432
> K3
>
> a strong no-trump (15-1/3) while
15-1/3 seems about right if you found a fit
> point total is to add 2/3 of a point for each ace and
> take off 1/3 of a point for each queen or jack. This
> makes it roughly propotional to a 6-4-2-1 count.
>
> Do people feel that this is appropriate for no-trump
> as well as suit bidding? That is, is
>
> A32
> K432
> A432
> K3
>
> a strong no-trump (15-1/3) while
But I would open it 1D
and certainly not 1NT
If partner answers with 1H, I would rebid 2H, but I have sympathy for 3H
>
> QJ6
> QJ7
> QJ8
> KQJ5
>
This I would not open a 15-17 NT
But I dont consider pass an option any close
Co Wiersma
OldPalooka
Mar 8, 2012, 1:53:24 PM3/8/12
to
try 13, 9, 5, 3, 1 which is another 40 point deck.
Bertil
Mar 8, 2012, 1:53:50 PM3/8/12
to
E.g. the best known is the Kaplan-Rubens 4C method,
which would count the 1st as 14.5 and the 2nd as 11.1.
Also, look at the Cowan-Banzai count 5-4-3-2-1, which
yields 18/1.5=12 and 24/1.5=16 resp. Thus it reverses the value of the
two,
so that the 1st is a weak NT while the 2nd is a strong NT.
In the good old days of Bridge Whist at the start of the last
century , the
Robertson Rule with 7-5-3-2-1 was the point count and 25/1.8=13.9
today
was the norm for a good NT bid.
You've revived an long standing debate about how to evaluate balanced
hands.
I plan to say more about this topic soon.
Bertil
Will in New Haven
Mar 8, 2012, 4:26:18 PM3/8/12
to
On Mar 8, 9:51 am, "Fred." <ghrno-goo...@yahoo.com> wrote:
> One adjustment to the Work count which maintains the 40
> point total is to add 2/3 of a point for each ace and
> take off 1/3 of a point for each queen or jack. This
> makes it roughly propotional to a 6-4-2-1 count.
>
> Do people feel that this is appropriate for no-trump
> as well as suit bidding? That is, is
>
> A32
> K432
> A432
> K3
>
> a strong no-trump (15-1/3) while
Well, I wouldn't open it a 12-14 NT (or a 10-13), so I guess it might
> point total is to add 2/3 of a point for each ace and
> take off 1/3 of a point for each queen or jack. This
> makes it roughly propotional to a 6-4-2-1 count.
>
> Do people feel that this is appropriate for no-trump
> as well as suit bidding? That is, is
>
> A32
> K432
> A432
> K3
>
> a strong no-trump (15-1/3) while
be a Strong NT. I would not rebid in NT, I would raise partner's
Major, so I wouldn't be showing a Strong NT. Well, if partner
responded 2C, I would bid 2NT.
>
> QJ6
> QJ7
> QJ8
> KQJ5
>
> ia a weak no-trump (12-1/3) at best?
seats, we play 10-13 and I just can't shoehorn that hand into that
range. Schenken liked "lots of honor cards," which goes against the
modern grain, and might have opened that 1NT and he played 16-18 some
of the time.
--
Will in New Haven
Paul Hightower
Mar 8, 2012, 4:47:15 PM3/8/12
to
"Fred." <ghrno-...@yahoo.com> wrote in message
news:4963514.361.1331218267742.JavaMail.geo-discussion-forums@ynee13...
Thomas Andrews did an extensive study of evaluation methods and found few
performed better than the Work count for determining when to bid 3NT.
Specifically, 6-4-2-1 was worse at notrump. He did document a Fifths Count,
deducting 1/5th point per King and Queen and adding 2/5ths point per Ten, as
outperforming the Work count. My method is simply to look at Aces, Tens and
Nines -- an average hand should have one of each. If you're short,
downgrade, if you have extras, upgrade. For a point count, Ace = 4.25, Ten =
0.5, Nine = 0.25. That gives 11 points per suit or average hand, so deduct
one per hand. to get back to an average of 10. A five card suit at notrump
is equal to an extra ten.
Link: http://bridge.thomasoandrews.com/valuations/
Andrews did not IIRC discuss Nines, but somewhere mentions that each honor
from the Queen on down is worth about half the previous honor, which is what
I based my count on. And the French Bridge Federation determined that a five
card suit at notrump was worth, on average, 0.4 hcp in actual play -- well
below the commonly reported advice to "add 1." Double-dummy, 4333 is better
at notrump than 5332, but in real life they don't always attack your weak
point and the long suit has more of a chance to help.
news:4963514.361.1331218267742.JavaMail.geo-discussion-forums@ynee13...
performed better than the Work count for determining when to bid 3NT.
Specifically, 6-4-2-1 was worse at notrump. He did document a Fifths Count,
deducting 1/5th point per King and Queen and adding 2/5ths point per Ten, as
outperforming the Work count. My method is simply to look at Aces, Tens and
Nines -- an average hand should have one of each. If you're short,
downgrade, if you have extras, upgrade. For a point count, Ace = 4.25, Ten =
0.5, Nine = 0.25. That gives 11 points per suit or average hand, so deduct
one per hand. to get back to an average of 10. A five card suit at notrump
is equal to an extra ten.
Link: http://bridge.thomasoandrews.com/valuations/
Andrews did not IIRC discuss Nines, but somewhere mentions that each honor
from the Queen on down is worth about half the previous honor, which is what
I based my count on. And the French Bridge Federation determined that a five
card suit at notrump was worth, on average, 0.4 hcp in actual play -- well
below the commonly reported advice to "add 1." Double-dummy, 4333 is better
at notrump than 5332, but in real life they don't always attack your weak
point and the long suit has more of a chance to help.
Bertil
Mar 8, 2012, 10:08:22 PM3/8/12
to
>
> - Show quoted text -
My calculator says 13+9+5+3+1=31. How do you get 40 from this set?
Your set is strangly similar to Kleinman's 13-9-5-2-1=30, which
differs very
little from 3x(4.5-3-1.5-0.75-0.25) or 13.5-9-4.5-2.25-0.75. Please
explain.
Bertil
Chris xxxxx
Mar 8, 2012, 10:42:06 PM3/8/12
to
1NT regardless of whether I were playing 12-14 or 15-17. In fact,
reverse the majors and I can assure you I would open 1D playing either
NT range. Suit-oriented hands with four spades (and, to a lesser
extent with 3=4 in the majors) should avoid opening in NT if they are
so close to the edge of their range that opening a suit is a
reasonable alternative.
The second hand is a classic weak NT. No way would I open it with a
strong NT regardless of whether I had heard of alternative point
counts. And passing is a little too anti-field to contemplate. But
weaken the hand slightly and I might well pass. (I once got an 85%
board on the second day of a national event for chucking a balanced 14
count in 4th chair--and I chucked it largely because I was downgrading
quacks. I haven't chucked a balanced 15 yet.)
Christopher Monsour
Co Wiersma
Mar 9, 2012, 7:44:54 AM3/9/12
to
Op 9-3-2012 4:42, Chris xxxxx schreef:
you will have to rebid 2NT
And that will show any some range
> The first hand is an example of a hand that I would probably not open
> 1NT regardless of whether I were playing 12-14 or 15-17. In fact,
> reverse the majors and I can assure you I would open 1D playing either
> NT range. Suit-oriented hands with four spades (and, to a lesser
> extent with 3=4 in the majors) should avoid opening in NT if they are
> so close to the edge of their range that opening a suit is a
> reasonable alternative.
but after 1D- 2C
> 1NT regardless of whether I were playing 12-14 or 15-17. In fact,
> reverse the majors and I can assure you I would open 1D playing either
> NT range. Suit-oriented hands with four spades (and, to a lesser
> extent with 3=4 in the majors) should avoid opening in NT if they are
> so close to the edge of their range that opening a suit is a
> reasonable alternative.
you will have to rebid 2NT
And that will show any some range
dak...@aol.com
Mar 9, 2012, 8:10:47 AM3/9/12
to
Agree Paul Hightower.
Thomas Andrews published online decades ago.
His research answers these evaluation questions.
Do you find in error in his work?
Or have you not researched what IS done in evaluation?
This is a peeve of mine. Research has been done,
but stick-in-the-muds don't use it. Why?
Vested interest in dis-proven evaluator?
Do not know these can be tested?
Did not know these are/were tested?
Do not keep up with decade-old research?
Thomas Andrews published online decades ago.
His research answers these evaluation questions.
Do you find in error in his work?
Or have you not researched what IS done in evaluation?
This is a peeve of mine. Research has been done,
but stick-in-the-muds don't use it. Why?
Vested interest in dis-proven evaluator?
Do not know these can be tested?
Did not know these are/were tested?
Do not keep up with decade-old research?
Will in New Haven
Mar 9, 2012, 9:58:02 AM3/9/12
to
in a game-forcing auction need not be a tight range. And the hand _is_
better than a 12-14 NT, although more for a suit contract.
Paul Hightower
Mar 9, 2012, 1:50:23 PM3/9/12
to
<dak...@aol.com> wrote in message
news:27958957-be85-44e1...@y17g2000yqg.googlegroups.com...
Keep in mind that Andrews himself questions the applicability of
double-dummy results to real at-the-table bridge. The FBF's 0.4 hcp value
for a five card suit translates to aomething like an extra 7-8% chance of
making 3NT -- yet double-dummy rates 4333 better than 5332. But absent more
at-the-table results, I'm going to assume that 4-3-2-1 with an adjustment
(primarily) for tens is more accurate at notrump than BUM-RAP or other
suit-oriented valuation methods.
news:27958957-be85-44e1...@y17g2000yqg.googlegroups.com...
double-dummy results to real at-the-table bridge. The FBF's 0.4 hcp value
for a five card suit translates to aomething like an extra 7-8% chance of
making 3NT -- yet double-dummy rates 4333 better than 5332. But absent more
at-the-table results, I'm going to assume that 4-3-2-1 with an adjustment
(primarily) for tens is more accurate at notrump than BUM-RAP or other
suit-oriented valuation methods.
Carl
Mar 9, 2012, 3:35:39 PM3/9/12
to
hands. It coincidentally adjusts the Work Count by 3rds, more or less
fitting in with the subject title nicely. Marty suggests adding a
point for a "good" five card suit, and adding a point for any
"quality" suit, with is 4 or more cards in length and containing three
or more honors.
On rare occasions the additions/subtractions tend to over/under-value
the hand.
Eric Leong
Mar 9, 2012, 4:33:43 PM3/9/12
to
Eric Leong
OldPalooka
Mar 9, 2012, 6:09:32 PM3/9/12
to
Chris xxxxx
Mar 9, 2012, 8:32:28 PM3/9/12
to
On Mar 9, 7:44 am, Co Wiersma <W...@xs4all.nl> wrote:
then I likely have the edge of whatever range it shows, so I am still
fine.
However, I don't necessarily have to bid 2NT. If playing KS, I can
rebid 2D. Once in a long while partner will pass 2D with a 3=3=3=4
good 9 count or bad 10 count. At IMPs I'll take that in exchange for
right-siding 3NT when partner's rebid is 2NT or 3NT. After all, this
hand has essentially no lead value after the 2C response.
Christopher Monsour
Steve Willner
Mar 9, 2012, 10:40:00 PM3/9/12
to
On 2012-03-08 9:51 AM, Fred. wrote:
> One adjustment to the Work count which maintains the 40
> point total is to add 2/3 of a point for each ace and
> take off 1/3 of a point for each queen or jack. This
> makes it roughly propotional to a 6-4-2-1 count.
>
> Do people feel that this is appropriate for no-trump
> as well as suit bidding?
The best I've seen -- based on double-dummy simulations, which may be
> One adjustment to the Work count which maintains the 40
> point total is to add 2/3 of a point for each ace and
> take off 1/3 of a point for each queen or jack. This
> makes it roughly propotional to a 6-4-2-1 count.
>
> Do people feel that this is appropriate for no-trump
> as well as suit bidding?
inaccurate -- is this:
For suit contracts, use (Alex Martelli's) BUM-RAP: add 1/2 for each A,
1/4 for each ten; subtract 1/2 for each Q, 1/4 for each J.
For notrump contracts, use (Thomas Andrews') fifths count: subtract 1/5
for each K or Q; add 2/5 for each ten.
The suggested "1/3 adjustment" is pretty good for suits, not so good for
notrump. I'd guess that for notrump, it's worse than no adjustment at all.
Of course any point count whatsoever is subject to large errors on
individual hands. Fit and controls are far more important than raw
point count.
--
Help keep our newsgroup healthy; please don't feed the trolls.
Steve Willner Phone 617-495-7123 swil...@nhcc.net
Cambridge, MA 02138 USA
Bertil
Mar 10, 2012, 8:45:43 AM3/10/12
to
>
> - Show quoted text -
Pray tell who is Brigitta? I don't recall having ever seen her name
> - Show quoted text -
and I could not find her
in the B.Crane 500 list. On what statistical study does she base her
opinion?
D.Kleinman used the Bennion count. On what study does Brigitta base
her scale?
You seem to think 4/124=1/31=3% is of great significance, why?
Construct all possible hands with 16 Whcp and no T and 4-3-3-3 pattern
and you'll get 28.
Sort them in order of probability and test any point count for mean
value of all 28 and you'll
get a good idea which point count best approximates 16.
Bertil.
Bertil
Nick France
Mar 10, 2012, 9:01:05 AM3/10/12
to
>
> - Show quoted text -
And all you great knowledge seems not to be wanted by those that
> - Show quoted text -
worship at her feet.
Nick France
Carl
Mar 10, 2012, 9:21:04 AM3/10/12
to
seemed to never actually play much bridge, developed this count in
conjunction with circular points, used for balanced hands with lots of
high card points but not much potential to take tricks. I have a
reference book here somewhere on circular points, but it has been
filed and I can't seem to round it up.
Bertil
Mar 10, 2012, 4:26:15 PM3/10/12
to
>
> - Show quoted text -
When searching Google I found nothing realted to Bridgitta nor to
> - Show quoted text -
circular point poin in bridge.
Are her ideas mentioned in the new ACBL Encyclopedia? If not, who
cares.
Bertil
Fred.
Mar 11, 2012, 11:37:06 AM3/11/12
to
and what to open. Opening decisions come at a time when visuallizing
partner's hand can be difficult for most of us.
However, I had come to the point of deciding that any gains over
the Work count were likely to be small for NT, and wanted to make
a sanity check.
Thanks to all for their responses.
Fred.
KWSchneider
Mar 22, 2012, 11:20:32 AM3/22/12
to schneid...@comcast.net
weeks. As most of you know, hand evaluation is very close to my heart.
And, I too subscribe to the Bridgitta concept but I've modified it
heavily over the years and supported it with my own work outlined
below.
A few general comments:
1) Assuming Work neglected to value the T but assumed that every hand
had one, the 40HCP count would statistically include one card of each
denomination from Ace through deuce. Ergo, a hand with the four top
honors but deficient in T's or 9's would yield a hypothetical value of
less than 10points. By corollary, the T [and 9, 8, etc] contribute
[albeit microscopically] to arriving at a 10point suit count.
2) TAndrew's work [ESPECIALLY BINKIE POINTS] has always been relevant
for notrump with the following proviso - DD is a questionable analysis
method in notrump contracts, specifically since the lead can make or
break the contract. And DD always finds the right lead.
3) Using a HCP count for unbalanced hand evaluation is nonsense, and I
strongly disagree with this approach. TA's work simply assigns an
aggregate value [value at BEST contract] in trump situations and
doesn't acknowledge the higher importance of both length in the trump
suit and trump honor cards.
For balanced hand evaluation only, if you look at TA's original work
carefully, we can develop the following equations for honors in
isolation [and in "average" length suits, ie 3-4 card suits]:
Total points in suit = 10
A=x, K=2/3A, Q=1/2K, J=1/2Q, T=1/2J, 9=1/2T...
Solving for x, we obtain [in points]
A=30/7=4.3pts
K=20/7=2.9pts
Q=10/7=1.4pts
J=5/7=0.7pts
T=5/14=0.4pts
9=5/28=0.2pts, etc
For simplicity's sake, we can adjust to A=4.25, K=3, Q=1.5, J=0.75,
T=0.5 [or T=9=0.25] which is what I use for balanced hand evaluation.
Kurt
Fred.
Mar 22, 2012, 6:31:42 PM3/22/12
to schneid...@comcast.net
NT do you round 14.5 up or down?
On the second, I don't really contest the valuation of the
hand at 13 points for no-trump. But, opening it a weak
NT in 1st position, who says we're going to play NT?
Fred.
KWSchneider
Mar 23, 2012, 9:34:42 AM3/23/12
to
On Mar 22, 6:31 pm, "Fred." <ghrno-goo...@yahoo.com> wrote:
<snipped>
A=4.25, K=3, Q=1.5, J=0.75, T=0.5 [or T=9=0.25]
A32
K432
A432
K3
I get 14.5 points for this hand...
QJ6
QJ7
QJ8
KQJ5
And I get 12 points for this hand.
> I guess the question on the 1st hand would be playing 15-17
> NT do you round 14.5 up or down?
I would play that 14.5 is not 15, hence not not open 1N. I need 14.75
to round up. Add a 9 and you're good to go.
> On the second, I don't really contest the valuation of the
> hand at 13 points for no-trump. But, opening it a weak
> NT in 1st position, who says we're going to play NT?
I'm not sure what your point is here. The hand values at 12 points as
a balanced hand - as a suited hand it maximum value is in clubs with
11 points, making notrump a preference unless responder has a
significant bias towards a different suit. Either way, the hand is
ugly and a marginal opener, certainly not a 1N hand. Playing strong
notrump, I would open it 1C.
Let's assume partner has 8 HCP and something like Axxxx xx Axx xxx.
This hand values at 8.5 points in notrump, 12 points in spades, and 6
points in clubs. As a dummy, opener values at 8points in spades. So
for the purposes of contract optimization, notrump yields 20.5 points
[7.5 tricks - 25points = 9 tricks], clubs yields 17 points [6.8 tricks
- 25points = 10 tricks] and spades 20 points [8 tricks]. The optimal
MP contract is probably 1N making 2, or perhaps 2S making 3...
Kurt
<snipped>
A=4.25, K=3, Q=1.5, J=0.75, T=0.5 [or T=9=0.25]
K432
A432
K3
I get 14.5 points for this hand...
QJ6
QJ7
QJ8
KQJ5
And I get 12 points for this hand.
> I guess the question on the 1st hand would be playing 15-17
> NT do you round 14.5 up or down?
to round up. Add a 9 and you're good to go.
> On the second, I don't really contest the valuation of the
> hand at 13 points for no-trump. But, opening it a weak
> NT in 1st position, who says we're going to play NT?
a balanced hand - as a suited hand it maximum value is in clubs with
11 points, making notrump a preference unless responder has a
significant bias towards a different suit. Either way, the hand is
ugly and a marginal opener, certainly not a 1N hand. Playing strong
notrump, I would open it 1C.
Let's assume partner has 8 HCP and something like Axxxx xx Axx xxx.
This hand values at 8.5 points in notrump, 12 points in spades, and 6
points in clubs. As a dummy, opener values at 8points in spades. So
for the purposes of contract optimization, notrump yields 20.5 points
[7.5 tricks - 25points = 9 tricks], clubs yields 17 points [6.8 tricks
- 25points = 10 tricks] and spades 20 points [8 tricks]. The optimal
MP contract is probably 1N making 2, or perhaps 2S making 3...
Kurt
Bertil
Mar 23, 2012, 10:45:57 AM3/23/12
to
by 5 and you'll get
6-4-2-1, which is twice the old 4A's count or Work + controls, as I
pointed out in my article
about Long Suit Evaluator.
The most common 16 HCP hand with the 4-3-3-3 pattern has 2A+1K+2Q+1J
which
counts as 12+4+4+1=21/1.3=16.15 or 8.5+3+3+0.75=15.25 while the 2nd
most
common has 2A+2K+1Q, which wuld count as 16.92 vs 16 by your formulas.
Wayne Burrows showed long ago that the mean number of T's is 1 at 10
HCP but
decreases linearly to 0.5 at 25 HCP. Thus at 16 HCP it would be 0.77
with a std.dev. of 0.77.
Thus a 16 HCP hand is as likely to have 0 as 2 T's. An alternative way
of counting points for T
is to disregard the 1st but count the 2nd and each additional but
deduct for no T, thus maintaining
a 40 count for the deck.
There are a total of 28 different combinations of A, K, Q, J that make
16 HCP.
Of these the top 17 account for 98% by frequency.The 17th has just
4Aces but is only about 1/40
as frequent as the 1st.
Have you submitted your point count to Bridge World, and if so, please
send me a copy.
Stig
>
> For simplicity's sake, we can adjust to A=4.25, K=3, Q=1.5, J=0.75,
> T=0.5 [or T=9=0.25] which is what I use for balanced hand evaluation.
>
KWSchneider
Mar 23, 2012, 11:24:14 AM3/23/12
to
On Mar 23, 10:45 am, Bertil <stigfjor...@hotmail.com> wrote:
<snipped>
> > Total points in suit = 10
> > A=x, K=2/3A, Q=1/2K, J=1/2Q, T=1/2J, 9=1/2T...
>
> > Solving for x, we obtain [in points]
> > A=30/7=4.3pts
> > K=20/7=2.9pts
> > Q=10/7=1.4pts
> > J=5/7=0.7pts
> > T=5/14=0.4pts
> > 9=5/28=0.2pts, etc
> You amze me. How did you convert 1/6 to 1/7? Divide the set 30-20-10-5
> by 5 and you'll get 6-4-2-1, which is twice the old 4A's count or Work
> + controls, as I pointed out in my article about Long Suit Evaluator.
Thank you - but you have no idea where I got the 7 divisor. Solving
for the geometric series and solving for "x" in the equation: x+2/3*x
+(1/2)*(2/3)*x*sum{(1/2)^n} = 10 [where n=1->10]. This yields x+2/3*x
+1/3*(1-(1/2)^9)/(1-(1/2))*x = 10 or x+(4/3)*x=10 or x=30/7.
The fact that the numerators are 30/20/10/5/3/1 but the denominator is
7 is pure mathematics.
> The most common 16 HCP hand with the 4-3-3-3 pattern has 2A+1K+2Q+1J
> which counts as 12+4+4+1=21/1.3=16.15 or 8.5+3+3+0.75=15.25 while the 2nd
> most common has 2A+2K+1Q, which would count as 16.92 vs 16 by your formulas.
Again by which method are you comparing my calcs?
> Wayne Burrows showed long ago that the mean number of T's is 1 at 10
> HCP but decreases linearly to 0.5 at 25 HCP. Thus at 16 HCP it would
> be 0.77 with a std.dev. of 0.77.
Makes sense...
> Thus a 16 HCP hand is as likely to have 0 as 2 T's. An alternative way
> of counting points for T is to disregard the 1st but count the 2nd and
> each additional but deduct for no T, thus maintaining a 40 count for
> the deck.
Actually, using the logic espoused here, a 16 HCP hand is MORE likely
to have 0 tens than 2 tens. But what has this to do with our
discussion?
> There are a total of 28 different combinations of A, K, Q, J that make
> 16 HCP. Of these the top 17 account for 98% by frequency.The 17th has just
> 4Aces but is only about 1/40 as frequent as the 1st.
> Have you submitted your point count to Bridge World, and if so, please
> send me a copy.
If you would like a copy of the "latest" revision, please email me and
I will send to you for your private review and comments.
Kurt
<snipped>
> > Total points in suit = 10
> > A=x, K=2/3A, Q=1/2K, J=1/2Q, T=1/2J, 9=1/2T...
>
> > Solving for x, we obtain [in points]
> > A=30/7=4.3pts
> > K=20/7=2.9pts
> > Q=10/7=1.4pts
> > J=5/7=0.7pts
> > T=5/14=0.4pts
> > 9=5/28=0.2pts, etc
> You amze me. How did you convert 1/6 to 1/7? Divide the set 30-20-10-5
> by 5 and you'll get 6-4-2-1, which is twice the old 4A's count or Work
> + controls, as I pointed out in my article about Long Suit Evaluator.
for the geometric series and solving for "x" in the equation: x+2/3*x
+(1/2)*(2/3)*x*sum{(1/2)^n} = 10 [where n=1->10]. This yields x+2/3*x
+1/3*(1-(1/2)^9)/(1-(1/2))*x = 10 or x+(4/3)*x=10 or x=30/7.
The fact that the numerators are 30/20/10/5/3/1 but the denominator is
7 is pure mathematics.
> The most common 16 HCP hand with the 4-3-3-3 pattern has 2A+1K+2Q+1J
> which counts as 12+4+4+1=21/1.3=16.15 or 8.5+3+3+0.75=15.25 while the 2nd
Again by which method are you comparing my calcs?
> Wayne Burrows showed long ago that the mean number of T's is 1 at 10
> HCP but decreases linearly to 0.5 at 25 HCP. Thus at 16 HCP it would
> be 0.77 with a std.dev. of 0.77.
> Thus a 16 HCP hand is as likely to have 0 as 2 T's. An alternative way
> of counting points for T is to disregard the 1st but count the 2nd and
> each additional but deduct for no T, thus maintaining a 40 count for
> the deck.
to have 0 tens than 2 tens. But what has this to do with our
discussion?
> There are a total of 28 different combinations of A, K, Q, J that make
> 16 HCP. Of these the top 17 account for 98% by frequency.The 17th has just
> 4Aces but is only about 1/40 as frequent as the 1st.
> Have you submitted your point count to Bridge World, and if so, please
> send me a copy.
I will send to you for your private review and comments.
Kurt
KWSchneider
Mar 23, 2012, 11:39:27 AM3/23/12
to
On Mar 23, 11:24 am, KWSchneider <questionofbala...@yahoo.com> wrote:
> > There are a total of 28 different combinations of A, K, Q, J that make
> > 16 HCP. Of these the top 17 account for 98% by frequency.The 17th has just
> > 4Aces but is only about 1/40 as frequent as the 1st.
> > Have you submitted your point count to Bridge World, and if so, please
> > send me a copy.
Somehow, I deleted a portion of my response. Jeff and I have been
> > There are a total of 28 different combinations of A, K, Q, J that make
> > 16 HCP. Of these the top 17 account for 98% by frequency.The 17th has just
> > 4Aces but is only about 1/40 as frequent as the 1st.
> > Have you submitted your point count to Bridge World, and if so, please
> > send me a copy.
revising for 4 months now. We have changed the name of the count to
"Unified Point Count" and it should be ready for publication as soon
as I'm comfortable. It is difficult to walk through an entire auction
and show how UPC works without overly complicating the concept. I'm
still struggling with making the concepts easy to understand in a 2
page article and I'm having a number of bridge acquaintances [BLM/SLM/
GLM] review for content and comprehension.
> If you would like a copy of the "latest" revision, please email me and
> I will send to you for your private review and comments.
Kurt
Fred.
Mar 23, 2012, 2:46:49 PM3/23/12
to
= 4*(2.25) + 3 = 10 + 3 = 13 in your methods and don't see
where I went wrong.
The point about not necessarilly playing the hand in NT is
just that. I don't like opening a weak NT becasue partner
is going to expext a better hand for suit play even with the
no-trump opener, and playing a weak NT the hand is far too
bad for 1C. I'm not sure we disagree.
Fred.
Bertil
Mar 23, 2012, 5:43:17 PM3/23/12
to
On Mar 23, 11:24 am, KWSchneider <questionofbala...@yahoo.com> wrote:
> On Mar 23, 10:45 am, Bertil <stigfjor...@hotmail.com> wrote:
>
> <snipped>
>
> > > Total points in suit = 10
> > > A=x, K=2/3A, Q=1/2K, J=1/2Q, T=1/2J, 9=1/2T...
>
> > > Solving for x, we obtain [in points]
> > > A=30/7=4.3pts
> > > K=20/7=2.9pts
> > > Q=10/7=1.4pts
> > > J=5/7=0.7pts
> > > T=5/14=0.4pts
> > > 9=5/28=0.2pts, etc
> > You amze me. How did you convert 1/6 to 1/7? Divide the set 30-20-10-5
> > by 5 and you'll get 6-4-2-1, which is twice the old 4A's count or Work
> > + controls, as I pointed out in my article about Long Suit Evaluator.
>
> Thank you - but you have no idea where I got the 7 divisor. Solving
> for the geometric series and solving for "x" in the equation: x+2/3*x
> +(1/2)*(2/3)*x*sum{(1/2)^n} = 10 [where n=1->10]. This yields x+2/3*x
> +1/3*(1-(1/2)^9)/(1-(1/2))*x = 10 or x+(4/3)*x=10 or x=30/7.
>
This calculation is unrealistic by including all cards from 10 to 2..
>
> <snipped>
>
> > > Total points in suit = 10
> > > A=x, K=2/3A, Q=1/2K, J=1/2Q, T=1/2J, 9=1/2T...
>
> > > Solving for x, we obtain [in points]
> > > A=30/7=4.3pts
> > > K=20/7=2.9pts
> > > Q=10/7=1.4pts
> > > J=5/7=0.7pts
> > > T=5/14=0.4pts
> > > 9=5/28=0.2pts, etc
> > You amze me. How did you convert 1/6 to 1/7? Divide the set 30-20-10-5
> > by 5 and you'll get 6-4-2-1, which is twice the old 4A's count or Work
> > + controls, as I pointed out in my article about Long Suit Evaluator.
>
> Thank you - but you have no idea where I got the 7 divisor. Solving
> for the geometric series and solving for "x" in the equation: x+2/3*x
> +(1/2)*(2/3)*x*sum{(1/2)^n} = 10 [where n=1->10]. This yields x+2/3*x
> +1/3*(1-(1/2)^9)/(1-(1/2))*x = 10 or x+(4/3)*x=10 or x=30/7.
>
An average 16 HCP flat hand will have 6-7 honor cards and thus only
7-6 spot cards.
Therefore, the only part of your formula that applies here is
x(1+2/3+1/3+1/6+1/12)=10.
Thus x=120/27=40/9=4.44, so that A=4.44,K=2.96,Q=1.48,J=0.74, T=0.37.
Using this count I can show that the frequency weighted mean will be
15.8
while your simplified count 4.25-3-1.5-0.75,T=0.25 yields 15.54.
I'll be happy to trade your write up for my planed article about
"Alternative counts for flat hands".
Stig
KWSchneider
Mar 23, 2012, 6:08:03 PM3/23/12
to
On Friday, March 23, 2012 5:43:17 PM UTC-4, Bertil wrote:
<snipped>
> This calculation is unrealistic by including all cards from 10 to 2..
> An average 16 HCP flat hand will have 6-7 honor cards and thus only
> 7-6 spot cards.
What does this have to do with anything? The value of a SUIT is 10 points. Not just the top 4 or 5 honors.
> Therefore, the only part of your formula that applies here is
> x(1+2/3+1/3+1/6+1/12)=10.
> Thus x=120/27=40/9=4.44, so that A=4.44,K=2.96,Q=1.48,J=0.74, T=0.37.
> Using this count I can show that the frequency weighted mean will be
> 15.8
> while your simplified count 4.25-3-1.5-0.75,T=0.25 yields 15.54.
Wrong! My count is 4.25, 3, 1.5, 0.75, 0.375, 0.125. You can either raise the T to 0.5 or average the 9/T at 0.25 each. "Frequency weighted mean" of what?
> I'll be happy to trade your write up for my planed article about
> "Alternative counts for flat hands".
I'm not looking for a trade. You asked to see my article and I said that you could have it, if you agreed to review and comment.
Kurt
<snipped>
> This calculation is unrealistic by including all cards from 10 to 2..
> An average 16 HCP flat hand will have 6-7 honor cards and thus only
> 7-6 spot cards.
> Therefore, the only part of your formula that applies here is
> x(1+2/3+1/3+1/6+1/12)=10.
> Thus x=120/27=40/9=4.44, so that A=4.44,K=2.96,Q=1.48,J=0.74, T=0.37.
> Using this count I can show that the frequency weighted mean will be
> 15.8
> while your simplified count 4.25-3-1.5-0.75,T=0.25 yields 15.54.
> I'll be happy to trade your write up for my planed article about
> "Alternative counts for flat hands".
Kurt
KWSchneider
Mar 23, 2012, 6:17:32 PM3/23/12
to
On Friday, March 9, 2012 10:40:00 PM UTC-5, Steve Willner wrote:
<snipped>
> The best I've seen -- based on double-dummy simulations, which may be
> inaccurate -- is this:
On what basis do you say "best I've seen?" Just curious...
> For suit contracts, use (Alex Martelli's) BUM-RAP: add 1/2 for each A,
> 1/4 for each ten; subtract 1/2 for each Q, 1/4 for each J.
I will continue to mock suit evaluations that discount trump length and the relative importance of trump honors. Even Thomas Andrew's work in a suit context [which started me on this journey in 1999], uses an arbitrary measurement called "best trump suit", which is essentially meaningless.
> For notrump contracts, use (Thomas Andrews') fifths count: subtract 1/5
> for each K or Q; add 2/5 for each ten.
> The suggested "1/3 adjustment" is pretty good for suits, not so good for
> notrump. I'd guess that for notrump, it's worse than no adjustment at all.
> Of course any point count whatsoever is subject to large errors on
> individual hands. Fit and controls are far more important than raw
> point count.
Agreed - but please give me some supporting documentation for your choices.
Kurt
<snipped>
> The best I've seen -- based on double-dummy simulations, which may be
> inaccurate -- is this:
> For suit contracts, use (Alex Martelli's) BUM-RAP: add 1/2 for each A,
> 1/4 for each ten; subtract 1/2 for each Q, 1/4 for each J.
> For notrump contracts, use (Thomas Andrews') fifths count: subtract 1/5
> for each K or Q; add 2/5 for each ten.
> The suggested "1/3 adjustment" is pretty good for suits, not so good for
> notrump. I'd guess that for notrump, it's worse than no adjustment at all.
> Of course any point count whatsoever is subject to large errors on
> individual hands. Fit and controls are far more important than raw
> point count.
Kurt
derek
Mar 23, 2012, 7:12:55 PM3/23/12
to
On Mar 23, 12:24 pm, KWSchneider <questionofbala...@yahoo.com> wrote:
> On Mar 23, 10:45 am, Bertil <stigfjor...@hotmail.com> wrote:
>
> > You amze me. How did you convert 1/6 to 1/7? Divide the set 30-20-10-5
> > by 5 and you'll get 6-4-2-1, which is twice the old 4A's count or Work
> > + controls, as I pointed out in my article about Long Suit Evaluator.
>
> Thank you - but you have no idea where I got the 7 divisor
You amze me. You actually expect Stig to have any idea about
> On Mar 23, 10:45 am, Bertil <stigfjor...@hotmail.com> wrote:
>
> > You amze me. How did you convert 1/6 to 1/7? Divide the set 30-20-10-5
> > by 5 and you'll get 6-4-2-1, which is twice the old 4A's count or Work
> > + controls, as I pointed out in my article about Long Suit Evaluator.
>
> Thank you - but you have no idea where I got the 7 divisor
mathematics when he can't even handle simple arithmetic?
Bertil
Mar 24, 2012, 7:28:58 AM3/24/12
to
On Mar 23, 6:08 pm, KWSchneider <questionofbala...@yahoo.com> wrote:
> On Friday, March 23, 2012 5:43:17 PM UTC-4, Bertil wrote:
>
> <snipped>
>
> > This calculation is unrealistic by including all cards from 10 to 2..
> > An average 16 HCP flat hand will have 6-7 honor cards and thus only
> > 7-6 spot cards.
>
> What does this have to do with anything? The value of a SUIT is 10 points. Not just the top 4 or 5 honors.
>
> > Therefore, the only part of your formula that applies here is
> > x(1+2/3+1/3+1/6+1/12)=10.
> > Thus x=120/27=40/9=4.44, so that A=4.44,K=2.96,Q=1.48,J=0.74, T=0.37.
> > Using this count I can show that the frequency weighted mean will be
> > 15.8
> > while your simplified count 4.25-3-1.5-0.75,T=0.25 yields 15.54.
>
> Wrong! My count is 4.25, 3, 1.5, 0.75, 0.375, 0.125. You can either raise the T to 0.5 or average the 9/T at 0.25 each. "Frequency weighted mean" of what?
If you construct all possible 16 HCP hands with A=4,K=3,Q=2,J=1 and
> On Friday, March 23, 2012 5:43:17 PM UTC-4, Bertil wrote:
>
> <snipped>
>
> > This calculation is unrealistic by including all cards from 10 to 2..
> > An average 16 HCP flat hand will have 6-7 honor cards and thus only
> > 7-6 spot cards.
>
> What does this have to do with anything? The value of a SUIT is 10 points. Not just the top 4 or 5 honors.
>
> > Therefore, the only part of your formula that applies here is
> > x(1+2/3+1/3+1/6+1/12)=10.
> > Thus x=120/27=40/9=4.44, so that A=4.44,K=2.96,Q=1.48,J=0.74, T=0.37.
> > Using this count I can show that the frequency weighted mean will be
> > 15.8
> > while your simplified count 4.25-3-1.5-0.75,T=0.25 yields 15.54.
>
> Wrong! My count is 4.25, 3, 1.5, 0.75, 0.375, 0.125. You can either raise the T to 0.5 or average the 9/T at 0.25 each. "Frequency weighted mean" of what?
T=0.5 all having the 4-3-3-3 pattern
you'll find a total of 77 combos having a total of 642 honor cards.
The unweighted mean is thus 8.34, which
comes from two extremes, viz. 4 for hands with only 4 Aces and 3 hands
with 12 honor cards. But the
frequencies for ech varies greatly, ranging from a relative value of
about 30 300 to less than 1. Thus one must
take into consideration the frequency weighted means when comparing
various point counts.
My calculations refer to hands with 16 W-hcp without any T, of which
there are only 28. Of these two have 9
honor cards and five have 8 honor cards. The unweighted mean is 6.71.
Keep in mind that the 4 Aces hands
rank very low in frequency. Counting point values for an occasional 9
and 8 in flat hands is a waste of time
You have my e-mail address if you care to send me a copy of your
latest write up. As in the past
I'll do my best to point out shortcommings. You are well aware that
I've been helpful in the past.
Stig
Bertil
Mar 24, 2012, 11:23:30 AM3/24/12
to
constructive vale.
Just your usual crap. Why do you persist in documenting being the
prime shithead at RGB?
Bertil
KWSchneider
Mar 24, 2012, 11:17:38 AM3/24/12
to
On Friday, March 23, 2012 2:46:49 PM UTC-4, Fred. wrote:
<snipped>
> I counted the second hand as 4*(Q+J) + 1*K = 4*(1.5 + .75) + 3
> = 4*(2.25) + 3 = 10 + 3 = 13 in your methods and don't see
> where I went wrong.
>
> The point about not necessarilly playing the hand in NT is
> just that. I don't like opening a weak NT becasue partner
> is going to expext a better hand for suit play even with the
> no-trump opener, and playing a weak NT the hand is far too
> bad for 1C. I'm not sure we disagree.
>
> Fred.
Fred - 4*2.25 = 9 not 10...
The point that I was making is that you open 1N when your hand is NOT an appropriate suit play hand - ie a "generalist hand". This hand would qualifies perfectly. Using your logic, the hand cannot be opened 1N but using my logic, it cannot be opened 1C [if playing weak notrump].
Kurt
<snipped>
> I counted the second hand as 4*(Q+J) + 1*K = 4*(1.5 + .75) + 3
> = 4*(2.25) + 3 = 10 + 3 = 13 in your methods and don't see
> where I went wrong.
>
> The point about not necessarilly playing the hand in NT is
> just that. I don't like opening a weak NT becasue partner
> is going to expext a better hand for suit play even with the
> no-trump opener, and playing a weak NT the hand is far too
> bad for 1C. I'm not sure we disagree.
>
> Fred.
The point that I was making is that you open 1N when your hand is NOT an appropriate suit play hand - ie a "generalist hand". This hand would qualifies perfectly. Using your logic, the hand cannot be opened 1N but using my logic, it cannot be opened 1C [if playing weak notrump].
Kurt
derek
Mar 24, 2012, 1:00:53 PM3/24/12
to
On Mar 24, 12:23 pm, Bertil <stigfjor...@hotmail.com> wrote:
> You don't amaze me with another post typically void of any positive or
> constructive vale.
> Just your usual crap. Why do you persist in documenting being the
> prime shithead at RGB?
Aw, Stiggy, I'd be hurt if you didn't call all the intelligent people
> You don't amaze me with another post typically void of any positive or
> constructive vale.
> Just your usual crap. Why do you persist in documenting being the
> prime shithead at RGB?
on this group by such names. You get posts devoid of positive value
from me (and others) because you demean us when we _do_ give a
constructive response. Try showing a little respect _and_ thought, and
you could see a real change.
Bertil
Mar 24, 2012, 4:44:04 PM3/24/12
to
respects my opinion,
which proves how absurd you are.I've no problem dealing with
intelligent and well
manered people. It's the louses who give me a pain.
You are the bridge equivalent of the f-n scumbag Zimmerman in
Florida,
who appointed himself block captain, so he could abuse and kill a
person
of a kind he detests. You are just a sick mongrel barfing on your own
behalf.
Your only bedmate, crony or sidekick is the Mutt. You two are like a
pair of peas
in a pod. If you have nothing positive or constructive to say, STFU.
Bertil
Mar 25, 2012, 11:56:03 AM3/25/12
to
On Mar 24, 1:00 pm, derek <de...@pointerstop.ca> wrote:
the very few like you,
the Mutt and your ilks. I intend to apply the Flodrida 'rule' of
"Stand you ground"
and use whatever expression to defend myself against any f-n verbal
vigilante.
You and they are a disgrace and I'll refer to you as Derek the Jerk.
Note the similarity
in spelling. For short I might use DJ.
Never getting another reply or comment from you would be too soon.
Recall what became of HS 409 and join him.
My name is Bertil
vsp...@hotmail.com
Mar 25, 2012, 7:54:51 PM3/25/12
to
On Friday, March 9, 2012 7:40:00 PM UTC-8, Steve Willner wrote:
> On 2012-03-08 9:51 AM, Fred. wrote:
> > One adjustment to the Work count which maintains the 40
> > point total is to add 2/3 of a point for each ace and
> > take off 1/3 of a point for each queen or jack. This
> > makes it roughly propotional to a 6-4-2-1 count.
> >
> > Do people feel that this is appropriate for no-trump
> > as well as suit bidding?
>
> The best I've seen -- based on double-dummy simulations, which may be
> inaccurate -- is this:
>
> For suit contracts, use (Alex Martelli's) BUM-RAP: add 1/2 for each A,
> 1/4 for each ten; subtract 1/2 for each Q, 1/4 for each J.
>
I misplaced my copies of Martelli's articles
> On 2012-03-08 9:51 AM, Fred. wrote:
> > One adjustment to the Work count which maintains the 40
> > point total is to add 2/3 of a point for each ace and
> > take off 1/3 of a point for each queen or jack. This
> > makes it roughly propotional to a 6-4-2-1 count.
> >
> > Do people feel that this is appropriate for no-trump
> > as well as suit bidding?
>
> The best I've seen -- based on double-dummy simulations, which may be
> inaccurate -- is this:
>
> For suit contracts, use (Alex Martelli's) BUM-RAP: add 1/2 for each A,
> 1/4 for each ten; subtract 1/2 for each Q, 1/4 for each J.
>
in TBW. Therefore I can't read his remarks
on the subject first hand.
Queens and jacks are undervalued in suits taking
tricks and overvalued in side suits.
X facing QJX
The queen and jack is likely to take no tricks
on offense and one on defense.
AKXXX facing QJX
If this suit is trumps, the queen and jack is
likely to be worth two full tricks.
Carl
Mar 25, 2012, 8:04:55 PM3/25/12
to
the primary thrust of this entire thread seems to be figuring out a
fair assignment to QJx a priori. What you demonstrate (at least to me)
is the folly of attempting to assign a precise value (with decimals,
no less) to a card combination that carries a huge standard deviation.
Essentially, QJx is worth 3 points, plus or minus three points.
Bertil
Mar 25, 2012, 9:06:23 PM3/25/12
to
On Mar 8, 10:51 am, "Fred." <ghrno-goo...@yahoo.com> wrote:
> One adjustment to the Work count which maintains the 40
> point total is to add 2/3 of a point for each ace and
> take off 1/3 of a point for each queen or jack. This
> makes it roughly propotional to a 6-4-2-1 count.
>
> Do people feel that this is appropriate for no-trump
> One adjustment to the Work count which maintains the 40
> point total is to add 2/3 of a point for each ace and
> take off 1/3 of a point for each queen or jack. This
> makes it roughly propotional to a 6-4-2-1 count.
>
> Do people feel that this is appropriate for no-trump
> as well as suit bidding? That is, is
>
> A32
> K432
> A432
> K3
>
> a strong no-trump (15-1/3) while
>
> QJ6
> QJ7
> QJ8
> KQJ5
>
> ia a weak no-trump (12-1/3) at best?
>
> Personally, when partner opens a weak
> no-trump (12-14) in 1st or 2nd I expect
> more controls than are in the latter hand.
>
> Thoughts?
>
> Fred.
Maybe you should get hold of the Jackson-Klinger book " Better
>
> A32
> K432
> A432
> K3
>
> a strong no-trump (15-1/3) while
>
> QJ6
> QJ7
> QJ8
> KQJ5
>
> ia a weak no-trump (12-1/3) at best?
>
> Personally, when partner opens a weak
> no-trump (12-14) in 1st or 2nd I expect
> more controls than are in the latter hand.
>
> Thoughts?
>
> Fred.
Balanced Bidding:
The Banzai Method". It is based on the 5-4-3-2-1 point scale and has
been praised
as well as ridiculed by experts. Read it anyway. It might give you new
ideas.
vsp...@hotmail.com
Mar 26, 2012, 9:00:38 AM3/26/12
to
Prior to the start of the bidding
it is not worth the mental energy
to be more precise. One should be
constantly adjusting as new facts
warrant.
Martelli's study demonstrates that
on average an ace is worth 1.5 kings
rather than 1.3333 kings. But we
should all know holding
S A65432 H A432 D A2 C A
The CA is worth only one trick.
The SA is worth much more.
SA > HA > DA > CA.
Work count is a sufficiently good
estimate of relative points. Just
be aware that mental adjustments
are necessary. While other point
counts may be more accurate, they're
too taxing for most human players.
The small gain is not more the
effort.
Fred.
Mar 26, 2012, 2:37:19 PM3/26/12
to
column instead of the 2.25. Sorry.
I agree that playing a 12-14 no-trump, I open this
hand 1NT if I open at all. With my regular partner
I'd pass this hand in a flash 1st or 2nd at IMPs because
in our partnerhsip a weak 1NT opening tends to promise
some versatility.
At MPs I may open in 1st or 2nd to keep from going too
counter field, but with an appology ready. Even 1NT-3NT
may be based on a good suit and the expectation of some
controls.
Fred.
Martin Ambuhl
Apr 1, 2012, 2:07:04 AM4/1/12
to
Bertil wrote on Friday, March 23, 2012 10:45 AM (rec.games.bridge):
>> 1) Assuming Work neglected to value the T but assumed that every hand
>> had one, the 40HCP count would statistically include one card of each
>> denomination from Ace through deuce. Ergo, a hand wit
>> 1) Assuming Work neglected to value the T but assumed that every hand
>> had one, the 40HCP count would statistically include one card of each
>> denomination from Ace through deuce. Ergo, a hand wit
> You amze me. How did you convert 1/6 to 1/7? Divide the set 30-20-10-5
> by 5 and you'll get
> 6-4-2-1, which is twice the old 4A's count or Work + controls, as I
> pointed out in my article
> about Long Suit Evaluator.
Thank you for not mentioning who pointed this out to you.
> by 5 and you'll get
> 6-4-2-1, which is twice the old 4A's count or Work + controls, as I
> pointed out in my article
> about Long Suit Evaluator.
> Wayne Burrows showed long ago that the mean number of T's is 1 at 10
> HCP but
> decreases linearly to 0.5 at 25 HCP.
monotonically. Or else you and Wayne Burrows are simply wrong.
Bertil
Apr 1, 2012, 6:51:14 AM4/1/12
to
On Apr 1, 2:07 am, Martin Ambuhl <mamb...@earthlink.net> wrote:
> Bertil wrote on Friday, March 23, 2012 10:45 AM (rec.games.bridge):
>
> >> 1) Assuming Work neglected to value the T but assumed that every hand
> >> had one, the 40HCP count would statistically include one card of each
> >> denomination from Ace through deuce. Ergo, a hand wit
> > You amze me. How did you convert 1/6 to 1/7? Divide the set 30-20-10-5
> > by 5 and you'll get
> > 6-4-2-1, which is twice the old 4A's count or Work + controls, as I
> > pointed out in my article
> > about Long Suit Evaluator.
>
> Thank you for not mentioning who pointed this out to you.
My memory is suffering from old age syndrome, so please refresh it
> Bertil wrote on Friday, March 23, 2012 10:45 AM (rec.games.bridge):
>
> >> 1) Assuming Work neglected to value the T but assumed that every hand
> >> had one, the 40HCP count would statistically include one card of each
> >> denomination from Ace through deuce. Ergo, a hand wit
> > You amze me. How did you convert 1/6 to 1/7? Divide the set 30-20-10-5
> > by 5 and you'll get
> > 6-4-2-1, which is twice the old 4A's count or Work + controls, as I
> > pointed out in my article
> > about Long Suit Evaluator.
>
> Thank you for not mentioning who pointed this out to you.
with reference
Bertil
Steve Willner
Apr 3, 2012, 9:54:55 PM4/3/12
to
SW> The best I've seen -- based on double-dummy simulations, which may be
SW> inaccurate -- is this:
On 2012-03-23 6:17 PM, KWSchneider wrote:
> On what basis do you say "best I've seen?" Just curious...
"Best" is a personal opinion based on a combination of accuracy and ease
of use. Accuracy is based on what I've seen written plus personal
experience, but I was particularly influenced by Tysen Strieb's
evaluation of other people's count methods and (I think it was his)
evaluation of suit patterns. And bear in mind the explicit
qualification of double-dummy evaluators, which we know are biased in
specific ways.
> I will continue to mock suit evaluations that discount trump length and the relative importance of trump honors. Even Thomas Andrew's work in a suit context [which started me on this journey in 1999], uses an arbitrary measurement called "best trump suit", which is essentially meaningless.
No one thinks distribution is irrelevant, but I should have made that
explicit. My comments Kurt quoted were limited to honor count. For
distribution, Goren count is as good as anything I've seen except that
it overvalues 4441 shape by about half a point. Honor count in trumps
is slightly tricky because minor honors are worth _less_ as length
increases, but at that level, no point count is going to be adequate.
For the record, I think Kurt's work with single-dummy evaluators is
important and innovative, but I'm not yet convinced that it's accurate.
Part of the reason is that I simply haven't seen enough data to make a
judgment. My main worry, though, is that Kurt's count seems to be based
on knowing the trump suit, whereas when opening the bidding (for
example), we really don't know what suit will be trumps. What's needed,
I think, is some kind of weighted average over all denominations. Minor
worries are that GIB apparently makes weird opening leads, and I'm not
sure Kurt's representation of his data is as simple as possible (but see
above about lack of data). That said, I'm sure single-dummy evaluation
has potential to be a major step forward.
--
Help keep our newsgroup healthy; please don't feed the trolls.
Steve Willner Phone 617-495-7123 swil...@nhcc.net
Cambridge, MA 02138 USA
SW> inaccurate -- is this:
On 2012-03-23 6:17 PM, KWSchneider wrote:
> On what basis do you say "best I've seen?" Just curious...
of use. Accuracy is based on what I've seen written plus personal
experience, but I was particularly influenced by Tysen Strieb's
evaluation of other people's count methods and (I think it was his)
evaluation of suit patterns. And bear in mind the explicit
qualification of double-dummy evaluators, which we know are biased in
specific ways.
> I will continue to mock suit evaluations that discount trump length and the relative importance of trump honors. Even Thomas Andrew's work in a suit context [which started me on this journey in 1999], uses an arbitrary measurement called "best trump suit", which is essentially meaningless.
explicit. My comments Kurt quoted were limited to honor count. For
distribution, Goren count is as good as anything I've seen except that
it overvalues 4441 shape by about half a point. Honor count in trumps
is slightly tricky because minor honors are worth _less_ as length
increases, but at that level, no point count is going to be adequate.
For the record, I think Kurt's work with single-dummy evaluators is
important and innovative, but I'm not yet convinced that it's accurate.
Part of the reason is that I simply haven't seen enough data to make a
judgment. My main worry, though, is that Kurt's count seems to be based
on knowing the trump suit, whereas when opening the bidding (for
example), we really don't know what suit will be trumps. What's needed,
I think, is some kind of weighted average over all denominations. Minor
worries are that GIB apparently makes weird opening leads, and I'm not
sure Kurt's representation of his data is as simple as possible (but see
above about lack of data). That said, I'm sure single-dummy evaluation
has potential to be a major step forward.
--
Help keep our newsgroup healthy; please don't feed the trolls.
Steve Willner Phone 617-495-7123 swil...@nhcc.net
Cambridge, MA 02138 USA
0 new messages