[BLML] Eric's position on disclosure when unsure

Jerry Fusselman jfusselman at gmail.com
Tue Nov 6 01:09:28 CET 2012


I'll try something slightly new to discuss the question of recommended ways
to *disclose when unsure*, which is currently at issue in BLML in the
threads "Nasty refutation [...]."

Rather than citing everything Eric Landau and others have argued on this
subject lately, I'll try to paraphrase the two main positions that have
come up.  I am only trying to summarize our positions in this post.  Please
don't assume that I have gotten Eric's position at all right until he says
so.

The subject may be a mine field due to our *conflict of visions* (his is
the unconstrained vision, and mine is the constrained vision), but I'll try
anyway, okay?  Hopefully he will correct my errors, and I intend to restate
the corrections in my own words.  Eric writes so well that there is some
hope for me getting it right.  If I need to revise, I'll revise here on
BLML.

The situation, suggested by Eric in response to Robert Frick's example, is
simple:

   - North has opened 2D.  He assumed that South almost surely knew what it
   meant.
   - South has to decide whether to describe it is a weak-two bid or
   Flannery or some combination of the two.  South is not sure which is
   intended, but he thinks that no other meaning is possible.  Eric would like
   me to add that this is not a matter of South forgetting what the
   understanding is.  I am not sure why Eric wants this, but we can assume it.
   - In the club where North and South play the vast majority of the
   time---call it "our club"---2D is usually Flannery.
   - Other than South's hand, and his observation that North opened 2D,
   there is no other fact to help him determine which meaning of 2D is the
   partnership understanding.
   - One assumption (p) will vary by case, as we'll see.

Marvin and I write that the following is good (and possibly the best)
advice to South: "Pick one agreement, perhaps the one that is most likely
according to your understanding, and state that without equivocation."

Eric disagrees quite strongly.  His position is more nuanced.  To
illustrate my understanding of his position, I need one extra assumption:
 Let

p = South's subjective probability that North intends 2D as Flannery.


Here are the cases:

   - p = 99.9%.  "It's Flannery."
   - p = 99.0%.  "It's Flannery."
   - p = 98.0%.  "It's either Flannery or a weak two in diamonds.  In our
   club, 2D is usually Flannery."
   - p = 90.0%.  "It's either Flannery or a weak two in diamonds.  In our
   club, 2D is usually Flannery."
   - p = 80.0%.  "It's either Flannery or a weak two in diamonds.  In our
   club, 2D is usually Flannery."
   - p = 60.0%.  "It's either Flannery or a weak two in diamonds.  In our
   club, 2D is usually Flannery."
   - p = 40.0%.  "It's either Flannery or a weak two in diamonds.  In our
   club, 2D is usually Flannery."
   - p = 20.0%.  "It's either Flannery or a weak two in diamonds.  In our
   club, 2D is usually Flannery."
   - p = 10.0%.  "It's either Flannery or a weak two in diamonds.  In our
   club, 2D is usually Flannery."
   - p = 2.0%.  "It's either Flannery or a weak two in diamonds.  In our
   club, 2D is usually Flannery."
   - p = 1.0%.  "It's a weak two in diamonds."
   - p = 0.1%.  "It's a weak two in diamonds."

Eric has been very clear that, "One *should not* provide probabilities or
subjective assessments."  Therefore, according to Eric, the p values should
never be disclosed.  (By the way, I wish I knew why.)  The sentences on the
right in each case are what I am believe Eric recommends South to say.  Now
in fairness, Eric has never explicitly stated which probabilities are "
significant" according to his advice to players, so I chose cutoff points
somewhere between 1 and 2 percent and between 98 and 99 percent. Eric can
update the cutoff point, but I think it is clear that if you are 99.9% sure
that it is Flannery, Eric's advice matches Marvin's---avoid any
equivocation.  Similarly, it is clear that if you are 80% sure that it is
Flannery, Eric's explicit advice under the assumptions I gave is almost
exactly the sentence I gave in quotes for that case.

If Eric claims that the cutoff points vary by player, then I would
appreciate hearing his personal cutoff points, or the cutoff points he
recommends to players following his advice.

I kept the p values out of the quoted sentences for each p value, because
Eric is very clear that the p value must never be stated.

So far so good, Eric?  Please try to be kind and patient with me.  :)  I am
happy to update the sentences and numbers in this table as needed.

Jerry Fusselman
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.rtflb.org/pipermail/blml/attachments/20121106/ca098a14/attachment.html 


More information about the Blml mailing list