Join Bridge Winners
All comments by Craig Zastera
You are ignoring the author of this comment. Click to temporarily show the comment.
The key is the negative double. After that, nothing responder bids is forcing except a cue-bid.

The essence of negative doubles is to show bad hands with some possibly important shape feature, so if you risk one with a good hand (particularly one with GF strength), it is up to *you* (the doubler) to later clarify your unexpected strength.

It might have been a somewhat different issue had responder started with 1 (over the (1) overcall) and then rebid 2 after opener bid 1NT. Still, nearly everyone plays that as non-forcing also without the interference, and I don't see why the (1) should change that agreement–the partnership's normal structure over opener's 1NT rebid (e.g. 2-way NMF) should remain intact.

Same if 1 response and opener rebids 2. If 2 now by responder would have been NF without the (1) overcall, I don't see why that agreement should be affected by the overcall. Same if it is played as forcing.
July 4, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
First, what makes you think I “blindly” did a simulation?

More likely I looked at the deal and observed that 3NT appears to be a much better contract than 4.

Then, to verify that observation, I did the simulation which demonstrates objectively what appeared to me to be the case–that 3NT is much better than 4.

BTW, double dummy *ONLY* a lead can ever beat the contract. And then only if s are 5-2 with at least one honor in the long hand. And then, only if declarer can't make 5 tricks in the majors without losing one.
July 4, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
John,
I said this was a tricky area.

I think your “4th best lead vs. NT” example is somewhat different from the AKx leads in the OP problem.

One minor issue is that 2 vs. NT might be from a 3 card suit–no law says lead must be 4th (and from longest and strongest).

But a more significant point, assuming “the good old days” where opponents can be relied upon to lead 4th best from their longest suit nearly all the time.

So the information from the lead is clearly “biased” in the lead vs. NT case. Opening leader selected his longest suit from which to lead. Therefore, RHO will necessarily be shorter in that suit (assuming they are not 4=4 or some such).

So is it valid to conclude that RHO is more likely to hold some card of interest in *our* long suit (i.e. a suit with majority length held by declarer/dummy)?

You could argue that had the hand been played from the other side with our RHO on lead, he would likely have selected a lead from *his* long suit, one in which LHO is shorter.

Do we now have valid reason to assume LHO is more likely to have the critical card in “our” suit?

You know, I'm not 100% sure.

I think my claim that probabilities are a function of our knowledge (and ignorance), hence will in general be different when we have different information is still valid.

Sometimes (perhaps in the NT case), we will be able to play the hand in such a way as to discover more possibly relevent information about the unseen hands before making our crucial decision. Maybe we can find out about the other suit where RHO is long and LHO is short which would change the probabilities about our missing key card.

But let's suppose we cannot. If all we can know is that LHO has 5 cards in suit “x” while RHO has only 2 cards, then I think it would still be statisically correct to play for RHO (the guy short in “x”) to hold our critical missing card, even though we know that LHO will selectively have chosen his longest suit to lead.

It is still true that LHO holds more cards in suit “x”, so, if we have no other knowledge, from the perspective of what we know he is less likely to hold a particular card in a different suit.
July 3, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Steve,
A nicely written and interesting post.

However, I don't think I quite agree with all of it.

First, except perhaps in the area of quantum mechanics, *probability* is really all about incomplete knowledge.

If we knew everything, then “probabilities” would not be relevent.

Therefore, there is no reason why two different people (say declarers) cannot *correctly* (i.e. correct according to *their* probabilities) choose different and incompatible actions based on their differing information.

That is, the “probability” of some specific occurrence (say the lay-out of the suit here) is *relative to the information declarer has* (or can obtain).

So if declarer “a” knows that s are 5=2, while declarer “b” knows that some other suit is 2=5, these two declarers have different information, hence can “correctly” come to different decisions about how best to play the suit.

Suppose both have to decide how to play the suit and it is not practical (or incurs additional risks) to try to obtain any additional information about the distributions of the unseen hands.

Then it is quite rational for declarer “a” to play his RHO (the one with only 2 s) to be longer in s, while declarer “b”, with different information about the opponents hands, can equally rationally decide to play his LHO to be longer in s.

In general, the “probability” of some unknown occurence changes constantly as the (relevent) information we have changes.

Now you bring up the issue of information obtained by force vs. information obtained by some voluntary plays made by the opponents.

That is a tricky issue. In the famous “Monty Hall” problem, Monty always deviously uses his knowledge about which door hides the prize to selectively open the one (remaining) door that doesn't have it.

So nothing useful can be inferred from his choice.
He is knowingly playing to try to fool us.

But the case where LHO plays A, K, and a for his partner to ruff is NOT LIKE the Monty Hall situation.

LHO (and RHO) did not play s in order to show you how the s are distributed as an attempt to fool you about the lay-out of the suit. No, they are not in a position to be so devious. LHO led s because that appeared to be his best lead from :AKxxx, and RHO (presumably) signaled encouragement because he thought that might be the best defense.

So although it is POSSIBLE that some other suit is distributed 2=5 (or whatever is needed to offset the inferences available from knowledge of the lay-out), it is also possible (and more likely actually) that the other suits are distributed normally so that the information about distribution does indeed lead to correct conclusions about the likely lay-out.

If all declarer knows (and can conveniently know) at the point where he has to decide how to play the suit is the lay-out of the suit, then it is completely correct for him to use this knowledge of the lay-out to draw probabilistic conclusions about the liklihood of various lay-outs of the suit (i.e. that RHO with more “vacant spaces” is more likely to have length and/or Q then LHO with fewer vacant spaces).

The need to assume that RHO has the K in order to have any chance of making 4 is relevent, I think, in that it reduces RHO's “vacant spaces” by one.

Also, any s that have already been played by either opponent at the point declarer has to make the critical decision should also be considered in reducing the number of “vacant spaces” in each opponent's hand.

Thus, if, unlike the actual declarer, one decided to start s by leading to dummy's A, so as to retain the option of finessing RHO out of :Qxxx or :Jxxx in the event that LHO should play the other honor on the first round, and LHO does in fact play a honor on the first round while RHO follows low under dummy's A and low again when a is led from dummy, the vacant space analysis is:

LHO:
5 s + 1 + 1 = 7 ==> LHO VSs = 6
RHO:
2 s + 3 s + 1 + 1 (K) = 7 ==> RHO VSs = 6

At this point, only one is missing (the other honor), and it appears to me the each opponent has 6 “vacant spaces” which might hold it.
Does this mean that the two plays are mathematically equally good?
I don't think so because of “restricted choice.”
If LHO started with H:QJ, he might equally well have played either on the 1st round of the suit, whereas if he held the one he played singleton, he would have had no choice.
So, I believe if playing the suit A first from dummy, if LHO plays an honor, it is best to play for it to have been stiff.

The only relevence of the “vacant spaces” analysis here is that since each opponent has exactly *6*, the odds that the restricted choice play (finesse) will win are actually slightly better than they would be in the “normal” restricted choice case.
That is because in the “normal” case, RHO's having one more “known” card (the one he plays to the second round of the suit) than LHO gives RHO one fewer “vacant space”, hence slightly reducing the chances of the finesse winning.

Notice that there is no need to consider possibilities other than stiff honor or :QJ doubleton with LHO.
Sure, he might have dropped honor from :Hx, but in that case the other honor would pop from East.
Or if he dropped honor from :QJx, RHO would show out.

Now, how about the analysis of the actual case where declarer started the s with K from hand (RHO dropping the J) and T, LHO following low twice.

Should declarer now play dummy's K (playing RHO for an original :QJx) or run the T (playing RHO for an original :Jx) ?

Restricted choice would suggest playing RHO for an original : Jx. There are *6* ways he could have started with :Hx but only *3* ways he could have started with :QJx – a powerful argument for the finesse.

How about vacant spaces? Including the necessary assumption that RHO must hold the K, at the point where declarer must make the critical decision:
LHO: 5s + 2s + 1 = 8 ==> 5 vacant spaces.
RHO: 2s + 2s + 1 + 1 = 6 ==> 7 vacant spaces.

Therefore, missing will be with RHO 7/12 of the time
vs. 5/12 with LHO.
BUT, this vacant space ratio is not enough to overcome the 2 to 1 odds that RHO started with :Hx rather than :QJx, so finessing LHO is correct.
July 3, 2018
Craig Zastera edited this comment July 3, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
A 2NT rebid over a 1NT response shows 18-19 HCPs, so this hand is not even close to enough for that.

The usual way to show 16-17 is to rebid 2m (often a 3 card minor), and then if responder doesn't pass (he usually won't), bid 2NT the 3rd time (say over responder's 2 preference).

It is true that the honor dispersion of this particular hand makes the system mandated 2 rebid (after 1NT) somewhat unappealing on : 532.
That would still be my choice though.

I think some “BART” aficionados might advocate a 2 rebid with this hand (and others with a similar shape, even with better 3 card holding), so as to pave the way for their favorite convention as often as possible.
July 3, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
To me, the most interesting point about this deal is the *ENORMOUS* superiority of 3NT (from either side) vs. 4.

A 5000 deal simulation gives the following results:
3NN: makes on 4579 deals (91.58%)
Choice of Opening lead:
# suits that can defeat 3NN:
*1* on all 421 deals

3NS: makes on 4578 deals (91.56%)
Choice of Opening lead:
# suits that can defeat 3NS:
*1* on all 422 deals

4HN: makes on 2361 deals (47.22%)
Choice of Opening lead
# suits that can defeat 4HN:
1: 412; 2: 136; 3: 1388; 4: 703

4HS: makes on 2367 deals (47.34%)
Choice of Opening lead
# suits that can defeat 4HS:
1: 433; 2: 704; 3: 929; 4: 567

So we see that 3NT (from either side) can rarely be defeated by any defense.
And when it is beatable, opening leader must lead *one specific suit* or 3NT will make even on those deals.

Meanwhile, 4 (from either side) can be beaten more than half the time.
And when it can be beaten, usually the opening lead is not so critical.

So perhaps more energy should be devoted to how to reach 3NT on these cards rather than to the very close (I suspect nearly 50-50) decision about which way to play the trump suit in 4 after the ruff.
July 3, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Methods which allow advancer to escape to two of any suit with a poor hand and a 5 card suit make 1NT overcalls considerably less dangerous then when 2 (and often 2) are usurped as artificial bids.

We use the cue-bid as Stayman, and overcaller needs to show both strength and his major suit length in reply, hence sometimes needing to jump.

So here, 2 would be Stayman, 2NT reply would show a minimum without four s, 2 a minimum with four s, etc.

3m advances are natural and invitational, while 3M (unbid major) advances are natural and GF (choice of game).

The one loss with this method is that 2NT advances must show invitational strength hands that don't fall into another category.
This includes invites with 5 cards (but not 4–those use Stayman cue) in an unbid major.

Over such an invitational 2NT, overcaller must pass with all minimums, thus possibly missing a superior 3M in a fit.

But if he has enough to accept an invite, he bids any four *or three* card major at the 3 level (forcing), so 5=3 and 5=4 major suit fits will be discovered.
July 3, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Besides general dislike for opening 1NT with 5 good s (missing superior contract too often), this particular hand also suffers from having two essentially unstopped suits–a significant negative for 1NT openings. Such hands fail in 3NT when partner has adequate strength to bid it significantly more often then balanced hands with more spread honors.
July 3, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Simple rule:
In competitive auctions, the only suit(s) you can make a splinter in is their suit. Jumps in other suits are fit-showing.
July 2, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
And in my methods, partner can bid 2any (except 2) for play.
July 2, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Wouldn't I have a similar problem without the (1)?
I would bid 1 in either case.

If partner rebids 1NT, I can rebid 2 (pass or correct).
If partner rebids 2, I can also bid a NF 2 if I deem that best.

BTW, some play that a 2 response shows 5 s & 4-5 s in a weak (say 5-8 or some such) hand just to try to improve chances with this hand type.
July 2, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Well, the first problem is that the range for the balancing 1NT is too wide–a 6 point range cannot be handled when advancer has a reasonable hand.

Assuming he wants to be in game (say 3NT) when his side has 25, that would suggest an invitational raise with 10.

But if balancer has only 10 HCPs, 2NT will be way too high (for safety, it is best to have 23 HCPs for 2NT).

So what can advancer do? I guess try to split the difference and invite game if 13 opposite will be enough.
That would suggest about 12 HCPs with a balanced hand.
July 2, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
We play 3 tiered splinters:
4 is a GF splinter, roughly 16-18 HCPs

3 is a split range splinter:
either just game invitational (about 14 HCPs), the usual
or
super-strong (like 19+ HCPs plus the shortness)

You could adjust the exact ranges slightly to taste, but the idea is that 4 is “enough for game, but not more than that”
while 3 covers splinters that are either weaker (game invites) or stronger.
July 2, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I too would open 1 with this in 3rd chair.

If it weren't for the vulnerability, I might choose 2 instead, so 1 is a bit of a compromise in recognition of the vulnerability.

I also believe that many (probably most) would open 1 in 3rd seat with this hand.

The problem (if there is one) is with partner's double of 1NT. He knows I may have opened light (the 1NT overcall is a bit of a clue, don't you think?), so a sound penalty double as a passed hand seems unlikely, although I suppose he could have something like :KQJTxx and a side ace.

But partner supposedly knows all this and has made a penalty double anyway. I simply have to assume he knows what he doing and PASS as I can't imagine what else I can do.

If we get a bad result, it certainly won't be my fault.
July 2, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I too think this is a stupid auction–why not just start with a forcing 1 response and then continue appropriately next round?

However, here is a very useful rule about negative doubles:
After starting with a negative double, NOTHING you bid
later is forcing except for a cue-bid.
July 2, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I still don't get what is so trivial about this hand.
I chose “pass” also, but I don't pretend that this call has to be the winner.

Suppose partner holds, for example:
Qxxx-Kxx-x-xxxxx
Now that is a pretty unspectacular 2 raise, yet 4 is cold.

Conversely, partner might hold:
Qxxx-KJ-Jxxx-KJx
Too much for a mere 2 raise, yet 4 has no play.

It seems to me that whatever call opener chooses could easily be wrong (or right), and will be at best a guess about the relative liklihood of responder's holding the right hand or the wrong one.

You might also notice that opening 1NT (yuck!) with this hand does nothing to solve this problem.
With the first hand, you will play 1NT with 4 cold.
With the second, you will most likely play 4 -1.
July 2, 2018
Craig Zastera edited this comment July 2, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Ah, I think I understand your point about “P/D inversion” possibly being beneficial if the two members of the partnership are not on the same wavelength as to whether PASS is forcing or not.

I have attempted to make this point myself in defending the benefits of “P/D inversion”, but have always been laughed at.
The prevailing feeling is “first, every partnership must have clear rules for when a PASS is forcing and when it is not”, so talk about the “benefits” of a convention occuring when we don't know whether FP is “on” or not is just silly.

So let's assume that a serious partnership *KNOWS* whether FP is “on” or not in every situation. Then, is there a benefit in playing “P/D inversion” (when FP is “on”) vs. using the traditional FP semantics???

The only one I've been able to think of is the one about P/D inversion making the two meanings of “PASS” (weakest call, or else a slam try if PASSer later pulls partner's double) as far apart (in terms of strength) as possible so as to minimize adverse rulings in case of hesitation situation.
July 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Well, your spot cards for East in the first diagram are inconsistent with what the text says was played.

But anyway, it seems to me that with the 2nd hand, East should pitch the K in order to get West to shift to a after he finishes the s.

In fact, in the original problem, even if he has only the K he should signal for a shift (low playing USDCA, or high if no easy to read low available).

Only some combination of hard-to-read (or too valuable to discard) spot cards in both majors should cause East any signaling problems if he holds the K or the KQ.

If he doesn't have or can't afford a “clarifying” discard in a major, then I guess all he's left with is to select his 2nd pitch in a way that attempts to give suit preference between the majors. Maybe partner won't be able to figure it out, but that might be the best East can do.
July 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Michal,
I don't understand your “bidding after a simple raise should be trivial with this hand” comment.

Do you mean that we should be playing some methods such that this hand fits well with one of them? If so, what methods?

Or are you saying that you think the correct rebid with this hand is “trivial” with standard methods? If so, what is that trivially correct rebid?
July 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Steven,
Perhaps you should re-read the “Revision Club” documentation on responding to 1NT. Here is the relevent excerpt for your benefit:

“1NT- 2C:
Forcing Stayman.
With 5-4 either way or 5-5 in the majors, and a hand
that just wants to sign off at the two level, pick a major and transfer to it.

Never Stayman with a weak balanced hand and 4-4 in the majors.
Our auction 1NT-2C, 2D-2H is not choice of majors at the two level–it is Smolen, and forcing.

WE DO NOT PLAY ‘GARBAGE STAYMAN.’
I have been monitoring occurrences of this treatment for a long time, and do not see it accomplishing much.
I see GS players responding 2C with 5-4 either way in the majors, and ending in 4-3 fits (when opener is 3-2 the wrong way in the majors) that play much worse than the 5-2 fit that could have been reached by a simple transfer. And what is opener supposed to do when he is 2-2 in the majors, which our 1NT openings can be?

Whatever merits there may be in playing GS (I am convinced there are not many), it has no place in Revision.
We prefer to concentrate on bidding games and slams.
If you have a five-card major (or two) and a weak hand, transfer out and be done with it.

However, it is acceptable to bid 2C on less than normal strength, with hands that are planning to pass partner’s response at the two level.
The usual holdings for these would be 4=4=4=1,
(4-3)=5=1, and 4=4=5=0.
A hand likeKxx=Qxxx=xxxxx-x would bid 2C.
These hands do not pose a problem as long as at least one four-card major is held; if partner rebids 2NT, showing both majors, the worst that can happen is to play three of a major in a known fit.”
July 1, 2018
.

Bottom Home Top