Join Bridge Winners
All comments by David Levin
You are ignoring the author of this comment. Click to temporarily show the comment.
I'm curious as to how West became declarer in spades (assuming that the auction wouldn't violate this website's Community Guidelines).
July 30
You are ignoring the author of this comment. Click to temporarily show the comment.
6 by East doesn't seem so hot. Let's say spades 3-2 or jack-stiff is 2/3. If the opening lead is a heart, 1/4 you can't go wrong, 1/4 you can't go right, and 1/2 you'll get it right say half the time, which comes out to 1/2. So for heart opening leads, making comes to about 1/2 x 2/3 = 1/3.

If anything but a heart is led, then making comes to about 1 x 2/3 = 2/3. So, 6 by East seems to be 50% if South would lead a heart half the time, and below 50% if South would lead a heart most of the time.

Others have suggested auctions that might land in a strain where West is declarer.
July 30
You are ignoring the author of this comment. Click to temporarily show the comment.
“All I know is declarer has the 9.”

Did I miss our being informed what heart was played from Dummy?
July 29
You are ignoring the author of this comment. Click to temporarily show the comment.
Does your web browser fail to display an elevator bar to the right of the auction?
July 29
You are ignoring the author of this comment. Click to temporarily show the comment.
Jan, I guess an intuitive sense of economics must run in your family. 8^)
July 28
You are ignoring the author of this comment. Click to temporarily show the comment.
Maybe they weren't looking for a spade fit so much as planning to convey information that might help Partner judge.
July 28
You are ignoring the author of this comment. Click to temporarily show the comment.
Assuming that a 1N intervention by South would have shown the pointed suits, Partner could have the strongest hand at the table but been stuck over 1 (with say, 2=4=4=3 and 17 HCP), especially in this age of light openers and responses.
July 26
You are ignoring the author of this comment. Click to temporarily show the comment.
I agree it's possible, although I would think that a higher spot is less likely to be misread.

Since posting the earlier comment, I've tried in vain to construct a layout where even if East reads Partner for a singleton, East would think it necessary to continue with a low one rather than cashing another. So maybe the potential misread shouldn't be much of a factor.
July 24
You are ignoring the author of this comment. Click to temporarily show the comment.
How does West know that East won't read the 7 as a singleton rather than discouraging (in light of West's failure to raise to 3) and consequently underlead the A for a presumed ruff?
July 24
You are ignoring the author of this comment. Click to temporarily show the comment.
Having not made clear the purpose behind questions #2 and #3, I've constructed a deal to try to address it.

AJx
Ax
QJxx
98xx
109xx Qx
Q109 Jxxxx
1098x x
xx AKJ10x
Kxxx
Kxx
AKxx
Qx
In an uncontested auction with E/W vulnerable, South deals and opens 1N, raised to 3N.

If E/W are wired, West can ensure a one-trick set by leading a club. But if E/W aren't wired, West might well lead the 10 and likely end up with two undertricks after Declarer loses the spade finesse.

In this case, it is misleading to use only the board result as a measure of a lead's effectiveness and thus as a proxy for dishonesty of the defenders. This effect might become negligible as the number of boards played increases, but can this be shown?

Compounding the confounding is that a wired human might opt for a diamond lead on the grounds that it will probably lead to a set and that a club lead would be perceived as suspicious. This suggests that an algorithm that's good at detecting suspicious actions by humans might be less so for bots, or the other way around.

(Maybe these concerns have already been addressed. I've read all of the comments to this article but am not knowledgeable in this area.)
July 23
You are ignoring the author of this comment. Click to temporarily show the comment.
I had a few questions.

1. In evaluating sample layouts, would the bots in question aim for highest expected number of tricks taken, best chance of a set, or something else? (Perhaps this question is more about how bots work in general than about how the bots in your proposal would work.)
2. How would the actual effectiveness of an opening-lead be defined? (For example, would it assume double-dummy play by both sides, starting with Dummy's play to Trick 1?)
3. For a given auction and opening-leader's hand, should record be kept of which leads would be “optimal” (as defined in the answer to #2), to help assess the significance of a bot's choice of lead?
July 22
You are ignoring the author of this comment. Click to temporarily show the comment.
“You can survive an unlikely 1-6-6-0 if East is queen empty. The defense is a bit tricky. South has to ruff the first diamond low. South pitch spades on subsequent diamonds to promote J as trumps are frozen.”

If the previous play is, opening lead to the Q, exit to the Q, black-suit ruff, crossing to the A, black-suit ruff, diamond exit ruffed by South, and black-suit exit ruffed by Declarer, then I'm not seeing why (1) promoting the J wouldn't occur only after Declarer's sixth trick, (2) Declarer couldn't take six tricks by exiting with a low trump to catch the defenders in a winkle, and (3) South's returning the Jx instead wouldn't allow Partner to overtake, drive out the Q, ruff in, draw trump, and cash winners.
July 22
David Levin edited this comment July 22
You are ignoring the author of this comment. Click to temporarily show the comment.
If Declarer's 1=6=6=0, you exit to the Q, and Declarer goes black-card ruff, cross to the A, black-card ruff, and exit in diamonds, how do the defenders prevent Declarer's getting in with a trump, exiting with another diamond, and taking another trump? (I'm assuming we're crediting Declarer with the Q and J.)
July 21
You are ignoring the author of this comment. Click to temporarily show the comment.
A and J unless Partner discourages on the former. After taking the anticipated K, Partner can cash the hypothetical J. I want to exhaust as many of Declarer's side-suit exit cards as possible.
July 21
You are ignoring the author of this comment. Click to temporarily show the comment.
Top diagram on page 9, delete the J and 4.
July 21
You are ignoring the author of this comment. Click to temporarily show the comment.
In my editing practice, I often arrange a deposit for 50% of the projected fee and the balance (of the actual fee) to be paid after I've delivered all of my work. That way, the client accepts the “risk” at the start, and I accept it at the end. This has worked extremely well, but I don't claim it would necessarily work for other professionals or other professions.
July 19
You are ignoring the author of this comment. Click to temporarily show the comment.
As I recall, Larry Cohen's original formulation of LOTT called for adjustments based on honor purity and how a suit that could contain losers is distributed between partners.
July 18
You are ignoring the author of this comment. Click to temporarily show the comment.
“…commit stayman”

Would that be a misdemeanor?
July 17
You are ignoring the author of this comment. Click to temporarily show the comment.
On the open team from District 2, it would have been neat if Stephen Mackay were named Mark Colbert.
July 17
You are ignoring the author of this comment. Click to temporarily show the comment.
If the robot were modified to add samples if it finds that each of the present samples would allow Declarer to take the rest of the tricks, then these added samples might be based on relaxing the constraints, such as giving Declarer 1 HCP less than the lower limit or a shape that's slightly outside the distributions shown in the bidding.

I'm not saying it would be trivial to implement, but it seems fairly consistent with I surmise the robot currently does.
July 16
.

Bottom Home Top