You are ignoring the author of this comment. Click to temporarily show the comment.

Was 6♥ your value bid, or was it a compromise choice to overbid triggered by the concern that 5♥ might befuddle partner?

As a related question - was partner full-valued for 7♥? (My view is that 7♥ needs more than a bare trick; 6♥ doesn't promise the goods. A reasonable gamble is normal.)

You are ignoring the author of this comment. Click to temporarily show the comment.

Right - equal teams is an implicit condition underlying my argument that there's a diminishing return for each additional imp won.

If we are the weaker team, I believe there can be increasing return, up to a point, for incremental imps that we collect on a swing board - more so the longer the match or the greater the disparity between the teams. It's only “up to a point” because of course a huge swing in our favor is overkill unless the match is longish or we are big underdogs.

You are ignoring the author of this comment. Click to temporarily show the comment.

First, apologies for my oxygen-deprived post of yesterday. I usually know better than to trust simulation results that I don't understand clearly.

Under the simple, obvious, and obviously simplified model that, excluding the “swing” hand under discussion, all imp deficits for the match from -11 imp to 0 imp are equally likely (don't care about other differences), then the 11 imp swing is exactly 11 times more likely than the 1 imp swing to win the match. One just has to account correctly for the ties. In general winning X imps, X>0, has X-1 chances to win outright (wins if we're down 1, 2, …, or X-1 imps), plus the 2 chances for half a win (we're tied or down X) are worth another win. Hence a +X imp swing has a winning expectation proportional to X.

If the match is short, clearly in the extreme of a 1-board match the 11 imp swing is not 11x more valuable (in expectation) than a 1 imp swing, as you said.

As the match length grows, we asymptotically approach my simple model. Gradually the 11 imp gain approaches, from below, 11x better than 1 imp. I don't see how 11 imps can ever be *more* than 11x better in expectation than 1 imp.

I assume whatever Jeff Rubens writes is correct. But I don't see, based on your explanation of his view, any contradiction with my conclusion.

You are ignoring the author of this comment. Click to temporarily show the comment.

I'm ready to believe that an 11 imp gain is more than 11 times as likely to swing a match as a 1 imp swing based on some simple match simulations. So empirically, yeah, and I appreciate being surprised. Is the phenomenon general, or something special about 1-imp swings? The phenomenon that X imps is more than X/Y better than Y imps seems to me almost completely due to the effect of ties. Under a range of assumptions, a 12-imp gain is very close to twice as likely to win a match as a 6-imp gain.

Unfortunately though the hopeful intuitive reductio ad absurdum argument that “Obviously a 200-IMP swing would decide most matches with probability 1, thus being ‘infinitely’ more important than a 1-IMP swing” doesn't make the case. It fails on two fronts.

One, its “obvious” truth depends on the assumption 1 imp wins a match with only infinitesimal probability. But if we buy that to begin with then there's nothing to prove re the original 11-imp claim. Actually the 200-vs-1 assertion fails for a 32 board match provided that the distribution of imp swings even slightly favors a small swings, i.e. if 1 imp is more likely than 20.

Two, the idea of extrapolating a 200-vs-1 imp claim to an 11-vs-1 conclusion seems to rest on an implicit assumption that one of two extremes holds: An X-imp swing is more than X/Y better than a Y-imp swing either always (i.e. for all X & Y with X>Y) or never. If we buy that, then the observation that a 400 imp swing isn't twice as useful as a 200 imp swing implies “never”, hence implies that an 11 imp swing isn't 11-fold better than 1 imp.

You are ignoring the author of this comment. Click to temporarily show the comment.

It's not clear to me what the differences compare. Are columns C, D, E information about a point in time (as the column headings "36 months ago" etc. seem to say) or about a period of time (in which case meaningful titles could be 2019, 2018, 2017)? (And why are C, D, E in backwards order?)

For example the totals line is Tot 160,767 .76- 1.46- 1.21-

Does .76- mean the membership was 0.76% higher at the 3 years ago point in time than today, and 1.46% higher 2 years ago than today? In other words, was there an 0.70% increase in 2017, an 0.25% decrease in 2018, and a 1.21% decrease in 2019? At least that's a trend!

You are ignoring the author of this comment. Click to temporarily show the comment.

For what it's worth, dealing probabilities – Assuming spades 5=4 and ♣A,K split, 1. 80/429 (19%) that both red suits split. 2. 42/715 (6%) that East has the equivalent of ♥Jxxx, of which it the ♣J is 5:4 to be with West.

You are ignoring the author of this comment. Click to temporarily show the comment.

If non-vul West might have QJxxx, Qxx, xx, Axx. I suspect the challenge is to find something legitimate and more likely than meeting up with an East who would find a preemptive raise on 10xxx, Jxx, xx, KJxx.

You are ignoring the author of this comment. Click to temporarily show the comment.

Comparing when winning “works” with when ducking “gains” isn't apples to apples.

Since at matchpoints the criterion for “works” isn't clear, the reasonable comparison is some version of gains vs gains. Why “version”? Because to be strictly accurate, the relevant comparison isn't between two hypothetical trick 1 plays at our table, it's what we can gain against other tables. As an extreme example to make the point clear, if you play this hand in 7NT winning and ducking trick 1 have the same matchpoint result, zero, so neither play gains over the other.

The practical significance for the present hand is that, we'd prefer a line which optimizes our trick expectation when the heart lead is from K109xx hence likely to be replicated at other tables, but tends to accomplish the somewhat different goal of maintaining an advantage or recovering from a deficit as the case may be, if we are facing a non-obvious lead from fewer hearts.

You are ignoring the author of this comment. Click to temporarily show the comment.

The bidding might be important. Why did East cover at trick 1? Did East know declarer has a heart suit? It looks to me that if East has more than two clubs and West has at least ♦QJxx, the cover handed us a chance we didn't otherwise have.

Therefore, playing Pavlicek's line, if for example East discards three or even four diamonds on the trumps there's an argument to make that it's clubs, not diamonds, that East has unguarded.

You are ignoring the author of this comment. Click to temporarily show the comment.

Suit-play's analysis is at most a starting point and not an answer for two reasons. It 100% does not take context (i.e. Jim's reasons for guessing the K location) into account, and about as certainly (reason omitted) it assumes that the defense has complete information.

Once you lead the Q, subsequent questions are in the domain of psychology, not computed probability. Why? Note that double-dummy neither defender can gain by withholding the K on the first trick. (And LHO ducking Kx can lose.) Ducking is a matter of whim, mind game, confusion, or incomplete information on the defender's part.

Therefore when the Q holds and RHO plays low again on the next trick, the dealing odds for KJxx vs K8xx (and if you like J8xx) with RHO have only limited relevance. More important is to imagine why and whether RHO might choose to duck from each combination.

And of course if instead RHO played the K first time, one might overlook that the fact of covering is a clue as to where the J is, but it is a clue. So declarer has a similar psychological problem either way.

You are ignoring the author of this comment. Click to temporarily show the comment.

That's what the widely scoffed label-maker is good for.

However, eventually I gave up on the BW binders. Their spines are way too greedy of shelf space, and I found a printer who did a nice job of tape binding (6 issues, $3, per book, about 5/8" (16mm) wide) with embossed spines. But they stopped offering the service. As a result I now, via diligent procrastination, have accumulated five or ten years of loose magazines.

But as several folks have pointed out, Bridge Worlds now aren't just stapled. They have a spine. Binding them now is probably a mistake. Just shelve them with vertical support at intervals. I'm about to experiment with magazine shelf holders. It looks like the cardboard Bankers Box - 4 inches wide and $1.31 per - can be cut down to BW height.

You are ignoring the author of this comment. Click to temporarily show the comment.

It's an open question whether Hugh Ross's wit was more quick or more wry. He dropped a trick when, holding Q1043, he failed to cover the lead of J from dummy's AJ986. “How can that ever be right?” his exasperated partner exclaimed. Seeing an opening Hugh explained “I thought he had stiff king.”

You are ignoring the author of this comment. Click to temporarily show the comment.

No, this is no ordinary guard squeeze. If solely East could beat dummy's heart spot the hand would be over without a second squeeze (against West). So what we have here is special.

On the other hand if there exists a layout in which first one defender then the other is guard squeezed, then only such a position would fully merit the “double guard” moniker. If not, then the hand here is as “double” as we can expect.

You are ignoring the author of this comment. Click to temporarily show the comment.

Unless declarer is playing a very deep game partner has ♦AJ109(8?). With an outside entry, ♦A and another would have been natural. Instead, partner suggested a spade card so, maybe why I lose at bridge, I'd lead a spade either as squeeze defense against xx, KQxxx, xx, AKQx, or perhaps to burn a dummy entry prematurely in some obscure layout.

You are ignoring the author of this comment. Click to temporarily show the comment.

I agree it's weird rules that allow your hand to be a secret from the opponents.

By the way, if declarer concedes the remaining tricks is it currently considered a claim - hence allowing defender to insist on seeing the hand (and to hear a plan of play)?

I do understand the justification in part - if time is short, asking an opponent to pull out their hand from the last board and show it to you will be annoying in the same way discussing the last hand is annoying. And in practice even if there were a rule entitling one to see the cards, I can imagine an impatient opponent brusquely replying unhelpfully “The diamond king was onside” or “nobody revoked.” Then enforcing the new rule would have an emotional toll.

Charles Brenner

As a related question - was partner full-valued for 7♥? (My view is that 7♥ needs more than a bare trick; 6♥ doesn't promise the goods. A reasonable gamble is normal.)

Charles Brenner

If we are the weaker team, I believe there can be

increasingreturn, up to a point, for incremental imps that we collect on a swing board - more so the longer the match or the greater the disparity between the teams. It's only “up to a point” because of course a huge swing in our favor is overkill unless the match is longish or we are big underdogs.Charles Brenner

Under the simple, obvious, and obviously simplified model that, excluding the “swing” hand under discussion, all imp deficits for the match from -11 imp to 0 imp are equally likely (don't care about other differences), then the 11 imp swing is exactly 11 times more likely than the 1 imp swing to win the match. One just has to account correctly for the ties. In general winning X imps, X>0, has X-1 chances to win outright (wins if we're down 1, 2, …, or X-1 imps), plus the 2 chances for half a win (we're tied or down X) are worth another win. Hence a +X imp swing has a winning expectation proportional to X.

If the match is short, clearly in the extreme of a 1-board match the 11 imp swing is not 11x more valuable (in expectation) than a 1 imp swing, as you said.

As the match length grows, we asymptotically approach my simple model. Gradually the 11 imp gain approaches, from below, 11x better than 1 imp. I don't see how 11 imps can ever be *more* than 11x better in expectation than 1 imp.

I assume whatever Jeff Rubens writes is correct. But I don't see, based on your explanation of his view, any contradiction with my conclusion.

Charles Brenner

Unfortunately though the hopeful intuitive reductio ad absurdum argument that “Obviously a 200-IMP swing would decide most matches with probability 1, thus being ‘infinitely’ more important than a 1-IMP swing” doesn't make the case. It fails on two fronts.

One, its “obvious” truth depends on the assumption 1 imp wins a match with only infinitesimal probability. But if we buy that to begin with then there's nothing to prove re the original 11-imp claim. Actually the 200-vs-1 assertion fails for a 32 board match provided that the distribution of imp swings even slightly favors a small swings, i.e. if 1 imp is more likely than 20.

Two, the idea of extrapolating a 200-vs-1 imp claim to an 11-vs-1 conclusion seems to rest on an implicit assumption that one of two extremes holds: An X-imp swing is more than X/Y better than a Y-imp swing either always (i.e. for all X & Y with X>Y) or never. If we buy that, then the observation that a 400 imp swing isn't twice as useful as a 200 imp swing implies “never”, hence implies that an 11 imp swing

isn't11-fold better than 1 imp.Charles Brenner

Charles Brenner

ago" etc. seem to say) or about a period of time (in which case meaningful titles could be 2019, 2018, 2017)? (And why are C, D, E in backwards order?)For example the totals line is

Tot 160,767 .76- 1.46- 1.21-

Does .76- mean the membership was 0.76% higher at the 3 years ago point in time than today, and 1.46% higher 2 years ago than today? In other words, was there an 0.70% increase in 2017, an 0.25% decrease in 2018, and a 1.21% decrease in 2019? At least that's a trend!

Charles Brenner

Assuming spades 5=4 and ♣A,K split,

1. 80/429 (19%) that both red suits split.

2. 42/715 (6%) that East has the equivalent of ♥Jxxx,

of which it the ♣J is 5:4 to be with West.

Charles Brenner

Charles Brenner

Since at matchpoints the criterion for “works” isn't clear, the reasonable comparison is some version of gains vs gains. Why “version”? Because to be strictly accurate, the relevant comparison isn't between two hypothetical trick 1 plays at our table, it's what we can gain against other tables. As an extreme example to make the point clear, if you play this hand in 7NT winning and ducking trick 1 have the same matchpoint result, zero, so neither play gains over the other.

The practical significance for the present hand is that, we'd prefer a line which optimizes our trick expectation when the heart lead is from K109xx hence likely to be replicated at other tables, but tends to accomplish the somewhat different goal of maintaining an advantage or recovering from a deficit as the case may be, if we are facing a non-obvious lead from fewer hearts.

Charles Brenner

Therefore, playing Pavlicek's line, if for example East discards three or even four diamonds on the trumps there's an argument to make that it's clubs, not diamonds, that East has unguarded.

Charles Brenner

Charles Brenner

Once you lead the Q, subsequent questions are in the domain of psychology, not computed probability. Why? Note that double-dummy neither defender can gain by withholding the K on the first trick. (And LHO ducking Kx can lose.) Ducking is a matter of whim, mind game, confusion, or incomplete information on the defender's part.

Therefore when the Q holds and RHO plays low again on the next trick, the dealing odds for KJxx vs K8xx (and if you like J8xx) with RHO have only limited relevance. More important is to imagine why and whether RHO might choose to duck from each combination.

And of course if instead RHO played the K first time, one might overlook that the fact of covering is a clue as to where the J is, but it is a clue. So declarer has a similar psychological problem either way.

Charles Brenner

However, eventually I gave up on the BW binders. Their spines are way too greedy of shelf space, and I found a printer who did a nice job of tape binding (6 issues, $3, per book, about 5/8" (16mm) wide) with embossed spines. But they stopped offering the service. As a result I now, via diligent procrastination, have accumulated five or ten years of loose magazines.

But as several folks have pointed out, Bridge Worlds now aren't just stapled. They have a spine. Binding them now is probably a mistake. Just shelve them with vertical support at intervals. I'm about to experiment with magazine shelf holders. It looks like the cardboard Bankers Box - 4 inches wide and $1.31 per - can be cut down to BW height.

Charles Brenner

Charles Brenner

On the other hand if there exists a layout in which first one defender then the other is

guardsqueezed, then only such a position would fully merit the “double guard” moniker. If not, then the hand here is as “double” as we can expect.Edit: https://en.wikipedia.org/wiki/Guard_squeeze shows a layout in which each opponent does get guard-squeezed. Interesting.

Charles Brenner

Charles Brenner

Charles Brenner

By the way, if declarer concedes the remaining tricks is it currently considered a claim - hence allowing defender to insist on seeing the hand (and to hear a plan of play)?

I do understand the justification in part - if time is short, asking an opponent to pull out their hand from the last board and show it to you will be annoying in the same way discussing the last hand is annoying. And in practice even if there were a rule entitling one to see the cards, I can imagine an impatient opponent brusquely replying unhelpfully “The diamond king was onside” or “nobody revoked.” Then enforcing the new rule would have an emotional toll.

Charles Brenner

Per “ethics” as defined in the laws, I'm unaware of any ethical problem. But if by “ethics” you mean personal code, I wouldn't do it.

Charles Brenner