Join Bridge Winners
All comments by Steve Willner
You are ignoring the author of this comment. Click to temporarily show the comment.
Looking at a deal from partner's or an opponent's point of view is a hugely important bridge skill. It may even be the most important skill that separates experts from the rest of us.
Aug. 15, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
In 6 and 7, the finesse fails _if there is a ruff_. If not, the finesse succeeds in these cases. That agrees with what Sriram wrote about the technical play if no ruff is possible.

In case 5, either play will succeed (though finesse scores an overtrick). That gives the edge to the drop if you are certain C-8 is stiff (perhaps from the tempo of the lead).

At matchpoints in a decent field, though, technical considerations are secondary to the matchpoint considerations Mike wrote about.
Aug. 14, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Pardon me if I'm missing the obvious, but why does “the hesitation before bidding 3 suggested that West was considering bidding on himself” (AC quote from Bulletin writeup)? It seems more likely to me that West was thinking of passing 3Hx or bidding 3NT, either of which would suggest 3 over 4.

Some comments on other legal issues but not expressing any opinion on this specific case:
1. it would have been improper to call the TD before East's hand was revealed, i.e., before play was complete.

2. what East himself would have done without the UI is legally irrelevant. The legal question is what other players with the same overall skill and same methods, but not the same personal proclivities, might have done. The effect is that sometimes UI makes illegal an action a player was certain he was going to take.

3. the old “25% rule” (which was really a 75% rule) for logical alternatives is gone. Nowadays, a LA is one that “would be given serious consideration by a significant proportion of such players, of whom it is judged some might select it.” In other words, any action that isn't ridiculous is probably a LA. This is a qualitative rule, not a numerical one, but I'd say it corresponds to something like 10-15% in the old terms. It is definitely more stringent than the former rule.

4. authorities could adopt a procedure for announcing in advance one's intentions to bid on as part of screen regulations. So far as I know, no authority has done so. Despite that, if one explains partner's action as “GF,” it's unlikely opponents will complain if you later bid game. (They may complain if you don't, of course.)

5. self-serving evidence should be given less weight than other evidence but rarely zero weight. Few players will directly lie, but many will “overlook” less favorable evidence that may exist. (But per item 2, “I was always going to …” gets zero weight, not because it's self-serving but because it's irrelevant.)
Aug. 12, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
If you are pitching twice, you can discourage both red suits (as John wrote). If you don't want a trump promotion, either encourage one of the red suits (even with nothing in it if you know it won't hurt) or discourage one and pitch in it again (implying you prefer the other red suit).

It's harder if you are pitching only once. You can look for a dramatic “alarm clock” discard, but failing that, you can either discourage the suit partner is likely to switch to or encourage an unattractive suit (sort of a low-grade alarm clock). Obviously neither of these is perfect, and partner may need to be a good guesser. However, he should at least be considering a trump promotion as a possible play.
Aug. 9, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Of the 32 players unable to get visas for Philadelphia, how many had qualified to represent their country on official teams?
Aug. 8, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Sorry for the confusion. We seem to be in violent agreement here. I might differ slightly in saying an artificial score is technically legal (what's obvious to you and me might be “not obvious” to a lazy or incompetent Director), but giving an artificial score is a truly horrid idea unless the board is wholly unplayable.

In most places, but not in the ACBL, adjusted scores can be assigned as weighted sums of as many possible outcomes as are needed. I don't know whether that makes a Director's job easier or harder, but they cope somehow.
July 28, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
The fact that something is legal does not make it a good idea. I'll stick with “horrifically bad directing” and “self-admission of bad directing.” In this case, the result of not illegally pulling the final double is pretty obvious.

Edit: it occurs to me that perhaps some explanation is in order. Under the Laws of bridge, the final 4H bid was illegal. Any knowledgeable and honest player would have passed the double, and the result would have been -730. Instead, we have a player who made an illegal bid, presumably unknowingly, but still illegal. Why would you give that player a better score than the one the knowledgeable, honest player would have received?
July 27, 2013
Steve Willner edited this comment July 27, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
The first paragraph is OK. The second one is a bit severe for a club game but legal. The third… well, not so much. It's the Director's job to work out what would have happened if the various infractions had not occurred. (Directors should always consult on judgment rulings, of course.) After 3S, P-P-x-AP is certainly “at all probable,” and it looks “likely” to me. (See Law 12C1e.) One's bridge judgment can differ, so give a different result if you like, but there's no reason to give an artificial score here. Doing so is a self-admission of bad directing.

After the slow penalty double was pulled, there's no problem at all. That easily leads to 3Sx= (unless your bridge judgment is that pass wasn't a logical alternative).
July 27, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
If the pause before the double was that long, and the hands were the ones shown by Aviv, passing the double is certainly a logical alternative, and pulling the double is suggested by the long pause (UI). That makes 3Sx= a simple and obvious ruling.

The Laws don't give any clear procedure when there are multiple infractions by one side, but generally the idea is to give the non-offenders the best of any combination of “had the irregularity not occurred.” There are other routes to 3Sx, but I don't see anything better for the non-offenders.
July 27, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Option 3 is technically legal (Law 12C1d), but it is horrifically bad directing.

In some circumstances 3S= (undoubled) would be possible, but if the hands are the ones Aviv Shahaf reported, I don't think that will apply here. West is never going to let 3S play undoubled at matchpoints.

I like Steve Bloom's point 4, but a “5-10 second pause” is an eternity. More like 2-3 s is normal, at least where I play. The basic point, though, is that an in-tempo double has to be neither too fast nor too slow, and tempo that would be normal in a non-competitive auction will be too fast in this one.
July 27, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Thanks! I'm a bit slow but have put this into a spreadsheet. I thought the VPs never quite reached 20, no matter how big the margin, but now I see they do, e.g., at 30 VPs for a 4-board match.

My spreadsheet is generally consistent with the USBF scale at page 31 of http://usbf.org/docs/COC/General%20CoC.pdf , but there are a few anomalies. For example, for four boards and a margin of 6 IMPs, I get 13.28 VPs but the USBF gives 13.29. That doesn't seem needed to fix the “decreasing margin” problem. Anyway, thanks again for the information.
July 27, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
I think “separate or combine” is the wrong question, at least in competition. In that situation, it seems to me more important to indicate whether forcing passes are on or off. (This seems to be what Robson and Segal are saying, too.) A simple version is a) cue bid shows balanced raise, not necessarily to game, but forcing if opponents bid above wherever we stop, and b) 2NT shows an unbalanced raise, possibly but not necessarily with game values, but not forcing even if opponents bid 4S over our 4H. (I think it has to be forcing at the five level, but I like more forcing passes than most people.)
July 24, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
To be fair to Mr. Stewart, I was wrong about exactly what he wrote. He didn't say Line 2 is better, only that most humans would choose it and that it would work on the actual layout. No surprise, the column showed AJx with West.
July 24, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Rainer is referring to Thomas Andrews' “fifths count,” which is quite good (double-dummy) for balanced hands considering 3NT.
July 21, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
“If a bid is conventional, that fact does indeed make the bid alertable, unless it falls on a list of specific exceptions”

The quoted statement is correct, but a double is not a bid. In the ACBL, only “highly unusual and unexpected” doubles require an alert. As I wrote before, support doubles used to be in that category but arguably no longer are.

Negative and takeout doubles are conventions under the 1997 definition, which as far as I can tell is what the ACBL Alert Procedure is based on. So are support doubles, of course. Penalty is the natural meaning of double, even though hardly anyone uses it any more. :-)

Ed and Michael are right about negative doubles being alertable for many years. The decision to change and make penalty doubles alertable was controversial at the time, but now everyone is used to it.
July 21, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Neither takeout nor penalty is alertable. The traditional meaning was penalty, but I think takeout has become more common, perhaps depending on the exact sequence.
July 21, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
Kit's point count is almost identical to Alex Martelli's “BUM-RAP.” That one is 4.5, 3.0, 1.5, 0.75, 0.25 for ace through ten. It is intended for suit contracts and (as I'm sure Kit would agree) needs adjustment for honors in long versus short suits.
July 18, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
This sort of work has a long history – at least back to 1950 or so when Charles Goren paid a bridge-playing actuary to derive an accurate point-count method. It's no surprise that 4321 count plus distributional adjustments is pretty good.

Double-dummy simulations have been used for more than 15 years. Key names are Thomas Andrews
http://bridge.thomasoandrews.com/valuations/
and Alex Martelli
https://groups.google.com/forum/#!topic/rec.games.bridge/M6ImupkkWA0
Some of Martelli's work was published in _The Bridge World_, 2000 Jan and Feb, but there may be more articles that my quick search didn't find.

Tysen Streib has evaluated the evaluators and, IIRC, incorporated adjustments for distribution. (I may be confusing Streib's work with some of Martelli's.) Unfortunately I can't find any reference in a quick search. Zar Petkov has published yet another point count, but in the last version of his work I saw, the method he used was wrong, and I don't trust his results. (I think Streib's work also deprecates Petkov's.)

The overall concept of these studies is simple. One wants to find the values of coefficients a, b, c… in the equation
Tricks = a*(#aces)+b*(#kings)+c*…
that minimizes the difference over some set of deals between Tricks calculated from the formula and tricks actually made double dummy. This simple statement hides a lot of detail, so see references (and look for others) if you are interested. In particular, the statement about “some set of deals” is important: the coefficients differ depending on whether one is talking about balanced or unbalanced hands and on the potential level of the contract (part score, game, slam).

In my opinion – and I have certainly been wrong before – there is little of practical interest to be learned in further double dummy simulations. The new frontier is examining single-dummy simulations to allow for the well known – and perhaps still unknown! – biases in the double dummy results. Kurt Schneider has pioneered that approach, but I have some doubts about what I've seen of his methods and assumptions. I thought he was going to publish an article in _TBW_, but I don't see it indexed yet.
July 18, 2013
Steve Willner edited this comment July 18, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
I think Charles has it right, as usual. My one worry, as I think someone else mentioned, is that if West is a super player, the heart discard may be a “Greek gift” intended to put declarer off the working simple squeeze.

The argument for the double squeeze would be even stronger if East had had a weak 2D bid available.
July 17, 2013
You are ignoring the author of this comment. Click to temporarily show the comment.
There's no doubt that support double is a convention. The question is whether it's alertable. I've always alerted it, and everyone I know does likewise, but I don't find that actually stated in the document. (I could, of course, be missing it, but if so, I have lots of company.)

I think what may have happened is that support doubles (and redoubles) were “highly unusual and unexpected” when the document was created many years ago, and nobody has thought to update the document to make it say one way or the other.

One thing that definitely has changed is that pass when support double is available – therefore denying 3c support – used to be alertable but no longer is. I'm not sure exactly what changed to make it so.
July 17, 2013
.

Bottom Home Top