On Tuesday, January 23, 2001 at 7:12:35 AM UTC-8, Alex Martelli wrote:

...and now, 20 years later, I've found yet another major author "reading from the same book" wrt hand evaluation:
Kevin Cadmus, author of the book "BFUN" (now in the 6th edition, with a 7th edition soon to come). Whatever you
think about the system (weak NT, canape, strong club, relays, ...), the "TTS" (trick-taking score) adjustments to
4-3-2-1 for balanced hands are pretty good (and proven through an even larger number of simulations than either I
or Andrews ran). A and 10 are a +, Q and J a -; some -'s for short suits headed by Q or J; for honor combinations,
AQx, AJx, and KJx (and longer) are a +, any suit headed by AK gets a MINUS -- the only evaluation system I know
of that considers Axx in a suit plus Kxx in another BETTER than AKx in a suit and xxx in another...

Alex

"Harry Rich" <harr...@donet.com> wrote in message news:t6ji2qr...@corp.supernews.com...
[snip]

and I tend to ignore jacks and tens unless there're together with

something

else, either in my hand or partner's. I'm not saying I'm right and BUM

RAP
[snip]

suit, etc. When a hand has better than average concentrations of high

cards

as indicated by QT it has a better than average potential to support a

high
In other words, you're prey to the most *typical* misconception
shared by most players AND writes on bridge: that honor concentration,
honors being "together with something else", is a GOOD thing,
something which makes a hand stronger.
The only writers who (to my knowledge) have argued to the contrary
are myself and Jean-Rene Vernes; not surprising, because what we
have in common is that we *LOOK AT THE DATA* -- in his case, tens
of thousands of hands played in the Bermuda Bowl; in mine, hundreds
of thousands of hands 'played' by computers. Our results are
strikingly similar (just as mine are to those of Thomas Andrews,
who also uses computer-played-hands statistics, but in a way that
is completely different from my approach -- I don't think he's
published his results, except on his website, though). http://www.deja.com/getdoc.xp?AN=598724841 is an article of mine
(from March 2000) where I spell things out in more detail. But,
summarizing, AJx is basically the only honor combination that is
a plus (to both offensive AND defensive strength) wrt A and J in
two separate suits; others are indifferent *OR MINUSES* -- a J
*loses* about 0.1 tricks' worth when joined to a Q, a K, or both,
for example!
And *THAT* is what makes 'quick tricks' just about worthless:
they depend crucially on honor-togetherness, which *DOES NOT
MATTER* to hand-strength (AJx combination excepted).
An Ace and a King have the *SAME TOTAL WORTH*, both in offense
and in defense, whether they are in the same, or separate, suits.
Their worth is correlated to the *length* of the suit[s] they
are in, but NOT to their 'togetherness'. This is what THE CARDS
say -- and it's thus not surprising that Jean Rene Vernes, Alex
Martelli, and Thomas Andrews, all read the same truth, if they
are "reading from the same book" (the reality of card-play), even
though we may be reading from different pages.
Players' intuition is just wrong about this -- it SEEMS that
having honors togethers should be a plus; in this case, what it
seems, and what IS, are not correlated (when one considers a
large-enough number of hands). I suspect that, out of all
results that have come from statistical studies, this is the
single most significant one *because it contradicts intuition
and tradition*; others, such as the fact that the Four Aces'
3-2-1 count is a good match for reality and Work's 4-3-2-1 is
not, or that a 7-2-2-2 is a substantially weaker shape than
7-3-2-1, etc, etc, basically tend to _confirm_ the intuition
of most good players.

Alex

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)