Join Bridge Winners
All comments by Craig Zastera
1 2 3 4 ... 54 55 56 57 58 59 60 61 62 ... 94 95 96 97
You are ignoring the author of this comment. Click to temporarily show the comment.
Nice methods.
Hopefully, they included sufficient discussion with partner to agree on distinctions between all these choices.

My guess is 2NT “offensively oriented LR”, unless there is an agreement that says this promises 10 HCPs.

If so, perhaps this hand better matches agreed definition for “mixed raise” (3) of a 1 level overcall.

suit is not good enough for any “fit bid” in that suit.
May 16
You are ignoring the author of this comment. Click to temporarily show the comment.
I do not understand all the votes for “double.”

The OP states that their agreement is that double here is take-out.
It seems clear that you don't have that hand type, hence double is inappropriate with the stated agreements.

Are those who voted for double saying:
(a) I think double here should be penalties–long s
or
(b) I understand double is take-out, but I double anyway ??

If (a), the what call would you make with a “take-out” hand type?
May 15
You are ignoring the author of this comment. Click to temporarily show the comment.
As is so often the case, this poll conflates too many issues and none of the options offered expresses my views.

First, yes, I play that 2 here is natural.

Second, yes, I would sometimes overcall 2 even with four s.

Third, this particular hand is not strong enough for 2 in my view.

Fourth, if you strengthened the hand by replacing the 2 with a honor, by the time the hand was strong enough for a 2 overcall (surely the A would do it, perhaps the K), the suit is now too good to overlook by making a 2 overcall–that is, the chances that partner has 4+ s and short s whereby we belong in s but don't find that strain are too great to risk with an immediate 2 overcall.
May 15
You are ignoring the author of this comment. Click to temporarily show the comment.
Not an easy question to answer “scientifically” (i.e. with objective studies that demonstrate one approach is better).

Therefore, if partner is vehemently in favor of one approach, it's probably right to play his way. But keep track of every deal that comes up that is relevent to this choice and whether his way or yours would have worked better. Maybe if you can accumulate enough objective data supporting your view, you can persuade him to change.

That said, at *matchpoints* it seems to me that take-out doubles must be better. Responder will frequently have enough strength to want to compete (even 4 HCPs means we probably have half the deck or more) but no long suit in which to do so. So without negative doubles, he will have to pass all these, say, 4-7 point hands, which probably will more often than not produce an inferior result.
The pure “penalty double” responding hands, I think, will not occur frequently enough to compensate, and at matchpoints, frequency is everything.

At IMPs, I still play “negative doubles”, but I think the case for penalty double is stronger. Sure, they won't occur all that often, but when they do the result might be a big plus score (300 or more) on a partscore deal (or we might go plus when otherwise we'd fail in a thin game).
And when partner has to pass with a decent hand that would have like to have made a “negative” double, the loss (if any) often won't be large (maybe we're -110 instead of -50 or -100, or +50/+100 instead of +110 or +140).
May 15
Craig Zastera edited this comment May 15
You are ignoring the author of this comment. Click to temporarily show the comment.
I don't think you are summarizing the conclusions from these simulations quite correctly.

After a 15-17 HCP 1NT opener, when responder is in the borderline game range with a 5 card major:
1. When responder is 5332 and opener is 4333, even with
his 4 card suit opposite responder's 5 card major,
3NT is likely better than 4M.

2. If responder is *NOT* 5332, then 4M is likely
better when an 8+ card fit exists, even if
opener is 4333.

3. However, if opener is 4333 and responder is 5422,
then the advantage of 4M is slight, so either
contract is statisically satisfactory.

4. If opener has 3 card support for responder's major
and is NOT 4333 (e.g. 4432 or 5332), then 4M is
likely to be considerably better than 3NT.

Thus, a practical conclusion is that Jacoby transfer followed by a NT rebid (2N or 3N) should tend to suggest
5332 shape, or at worst 5422 (particularly if doubletons are strong), so that opener can reasonably play NT when he is flat.

The suggestion that responder, with 5332, should bid Stayman rather than Jacoby is not supported by these simulations.
However, I have done other simulations that suggest that when responder is 5332 *AND* has significantly more than minimum GF strength (e.g. 12-14 HCPs), *AT MATCHPOINTS* he will on average do better to just Stayman the hand and play 3NT unless a 5=4 major fit exists. I do not believe this strategy is best at IMPs, however.
May 15
You are ignoring the author of this comment. Click to temporarily show the comment.
Yes, this one is very close.
An “average” 18 HCP balanced hand opposite average 7 HCP balanced hand, with no 8+ card major suit fit, will make 3NT 54.9% of the time.

So if you want to draw the line between calling a hand “17” vs. calling it “18” at the midpoint between average 17 and average 18 performance opposite 7 HCP balanced hands, then that would be a 46% make percentage.
So this one falls a little below on that test, so it would
not be wrong to evaluate it as just a good 17.

So maybe I stretched a tiny bit to “follow the field” and open 1. It will make a good dummy for s, though, if partner has 5, but that could be handled by “super-accepting” if partner transfers to that suit over a 1NT opening.

It is interesting, though, that if partner happens to have 4 or 5 s (bolstering our obvious weakness in that suit), then the 3NT performance of this hand goes way up, easily earning it full 18 HCP status.
May 15
You are ignoring the author of this comment. Click to temporarily show the comment.
1000 deal simulations w/ random N/S cards:

1. partner has 7HCPs, balanced, < 5 s
3NTW makes: 437 deals
(avg bal 17 HCP opp avg bal 7 (no 8+ maj fit): 3NT: 37.1%)

2. partner has 7 HCPs, balanced, 4-5 s
3NTW makes: 566 deals

3. partner has 7HCPs, balanced, 5 s
4E makes: 553 deals
May 15
You are ignoring the author of this comment. Click to temporarily show the comment.
John,
If the responder is not 5332, then the advantage of 3NT over 4M when opener is 4333 pretty much disappears.

Even when responder is 5422 (which I have checked recently),
the make percentage for 4M and 3NT are about equal when opener's 4 card suit is opposite one of responder's doubletons, while 4M makes more often than 3NT when both of responder's doubletons are opposite 3 card holdings. This was the point of the two simulations I cited in my later comment.

I don't have any recent simulations for 3N vs. 4M when responder is 5431, but I suspect in that case 3NT will be definitely inferior statistically.

And, yes, it seems to me that in order for a partnership to take advantage of this rather striking result (i.e. that 3NT will make significantly more often than 4M when one hand is 4333 and the other 5332), it would be ideal to have sequences for responder that suggest 5332 with a high degree of probability, e.g. 1N-Jac. xfer-2M-3N (and, ideally, 2NT).
May 15
Craig Zastera edited this comment May 15
You are ignoring the author of this comment. Click to temporarily show the comment.
I answered “depends on which suit they opened” because
(a) if (2), then double is TO, 3 is majors
(b) if (2M), then double is TO, 3M cue is stopper ask
(with 4m = m + OM)
May 15
You are ignoring the author of this comment. Click to temporarily show the comment.
Remarkably, it appears that 3NT fares better than 4M even when opener's 4 card suit is the same as responder's 5 card major, although in that case the advantage of 3NT vs. 4M (in 5-4 fit) is not as great as when opener's 4 card suit
(in 4-3-3-3 hand) is a different suit from responder's 5 card suit.
May 15
You are ignoring the author of this comment. Click to temporarily show the comment.
Showing the 4 card minor (when reasonably good suit) without having slam going strength can win in several ways. Sometimes, we can discover a playable minor suit fit with a weak suit for NT and play 5m successfully.

But as simulations like the ones I mentioned above show, there is a crucial difference between responder's 5332 (5M) and 5422 and 5431 shapes opposite NT opener in that when opener is 4333, 3NT is often significantly superior to 4M whereas when responder is 5422 and especially 5431 it is generally better to play 4M when responder is 4333.

Therefore, my view is not only that it is a good idea for responder to show 5M4m shapes when he is GF, even when he is only invitational it is useful (and there are methods to facilitate this).

When I transfer with Jacoby and then bid 2NT or 3NT, I do not expect partner to “mechanically” play NT with 2 in my major and 3M or 4M with 3. No–he is supposed to treat my sequence as asking him to choose strain (NT vs. M). He is free to choose NT with 3 (even 4) in my major or choose the major even with only a doubleton when he has reason to do so.
May 14
You are ignoring the author of this comment. Click to temporarily show the comment.
Rainer,
With 5M4m22 and GF values, I would usually bid the four card minor after transferring to the major.

Exceptions would be with very strong doubletons and weak 4 card minor. Showing the minor would be strongly indicated with most points concentrated in the two long suits.

Furthermore, when responder is 5M4m22 with invitational or minimum GF values and opener is 4333, playing 3NT is generally as good as 4M.

I recently completed a couple of 1000 deal simulations
where responder has 9 HCPs with 5 s, a 4 card minor, and 2 and 2 in the Om.
I gave opener 15-17 HCPs with 4333 shape (only 3 s).
Here are the two results:
(a) when opener's 4 card suit was :
3NT : made on 608 deals
4 : made on 608 deals
An exact tie (for IMP purposes anyway). Remarkable.

(b) when opener's 4 card suit was a minor:
3NT : made on 589 deals
4 : made on 609 deals
So a slight win for 4 in this case, but this
will be reduced because when responder chooses
not to show 4 card minor, his hand will be more
NT suitable (e.g. strong doubletons).
May 14
You are ignoring the author of this comment. Click to temporarily show the comment.
John,
(a) minor suit transfers show 6+ card suits, not 5

(b) A new suit by responder after a minor suit transfer
showing shortness is standard (not universal, but
certainly the common treatment).
In support of this, I quote from Bridge World Standard
2017 summary of structure over our 1NT openings:

“(e) modified four-suit transfers with
2 = either:
* a range inquiry
(opener bids 2NT with a minimum)
or:
* s
(simple new-suit rebid by responder
shows shortness)

2NT = s
(simple new-suit rebid by responder
shows shortness)

3 = both minors weak (nonforcing)

3 of another suit = both minors strong
(3 = no major-suit shortness;
3M = at most one card in the suit bid).”
May 14
You are ignoring the author of this comment. Click to temporarily show the comment.
John,
I'm not sure if most people even have a minor suit transfer available over 2NT openings.
I know there are many structures over 2NT, but common (not mine) is to use 3 as minor suit Stayman and 4 and 4 as Texas transfers.

I suppose 4 is commonly Gerber, although I find that a poor use of the bid (I like 5=5 majors, no slam interest).

Anyway, even over 1NT openings, if a pair has minor suit transfers (e.g. 2 for s and 2N or 3 for s), I think usual use of 3M continuation is splinter.
May 14
You are ignoring the author of this comment. Click to temporarily show the comment.
John,
The simulations I did for this post use only the OP's specific East hand, so you are correct that one cannot conclude from these results that West should, in general, prefer 3NT with 4333 shapes.

However, other more general simulatons I have done *do* suggest that 4333 hands opposite 5332 shape hands on average do play better in NT than in a 5-3 or even 5-4 fit.

Thus, if East transfers and then bids *3NT*, I believe opener should leave it there with 4333 shape hands as that responder sequence strongly suggests 5332 (because with a 4+ card side suit and GF, he will usually bid it).

However when East transfers and then bids only *2NT*, the inference that he is likely 5332 is much weaker unless the partnership has some gadget (e.g. 2 stage transfers) available to show a second suit with only invitational values. In that case, it may be much riskier for opener to prefer NT with 3 card support just because he is 4333.
May 14
You are ignoring the author of this comment. Click to temporarily show the comment.
Nothing “gratuitous” or “overstatement” about it.
Just look up the topic in standard texts and this is what you will find.
Only rank beginners would think that 4NT would be ace (or key-card) ask after a major suit reply to Stayman.
May 14
You are ignoring the author of this comment. Click to temporarily show the comment.
I just completed some simulations which strongly suggest that the East hand is in fact too strong for a mere game invitational sequence, hence should in fact force to game.

I did a series of 1000 deal simulations putting this East hand opposite various *15* HCP balanced hands with random distribution of the N/S cards.

I looked only at 15 HCP West hands since 16/17 hands would presumably go to game opposite an invite, whereas 15 HCP hands in general would not.

Here are my results.
Notice:
1. 3NT far outperforms 4 when opener is 4333,
even when his four card suit is s.

2. for every balanced shape 15 HCP opener,
at least one of 3NT or 4 is well above 50%

a. West: 15 HCPs with *2* s (30.0% of 15 HCP bal hands)
results: 3NT made 60.0%; 4 made 36.8%

b. West: 15 HCPs with *3* s (46.7% of 15 HCP bal hands)
results: 3NT made 53.0%; 4 made 60.5%

b1. West: 15 HCPs, *3* s, some 4-3-3-3 shape (18.2%)
results: 3NT made 67.3%; 4 made 48.0%

b2. West: 15 HCPs, *3* s, some 4-4-3-2 shape (12.4%)
results: 3NT made 46.8%; 4 made 65.1%

b3. West: 15 HCPs, *3* s, some 5-3-3-2 shape (15.9%)
results: 3NT made 44.6%; 4 made 67.0%

c. West: 15 HCPs with *4* s (19.9% of 15 HCP bal hands)
results: 3NT made 46.2%; 4 made 58.6%

c1. West: 15 HCPs with 3=4=3=3 shape (4.0%)
results: 3NT made 57.5%; 4 made 35.5%

c2. West: 15 HCPs, *4* s, some 4-4-3-2 shape (17.0%)
results: 3NT made 45.2%; 4 made 65.4%

d. West: 15 HCPs, *5* s, some 5-3-3-2 shape (3.4%)
results: 3NT made 34.0%; 4 made 55.7%

Conclusions:
1. East must force to game
2. West should prefer 3NT when 4333, even s is his
four card suit.
3. Actual OP deal was unusual in that the poor mesh
of cards actually makes game poor (but the favorable
placement of N/S cards allows it to make anyway).
4. Those two hard-working ten spots in the East hand
make it much better than an average 2=5=3=3 9 count.
May 14
You are ignoring the author of this comment. Click to temporarily show the comment.
I said nothing about 4–I agree that 4 (or 4) following Stayman are length showing (and forcing).

But *4* following Stayman cannot be length showing, hence is available as a slam try in s after 2N-3-3.
May 14
You are ignoring the author of this comment. Click to temporarily show the comment.
This is pretty standard stuff–Stayman or a Jacoby transfer followed by are 4NT are natural.

After 2N-3-3, if responder wants to agree s and show slam interest, he bids 4.

One could have the agreement that over this 4, opener either rebids 4 to show a “bad” hand for slam (responder can still continue with 4NT RKCB s over this), or else responds keycards above 4 (4N/5/5/5) with a hand he deems “good” for slam.
May 14
Craig Zastera edited this comment May 14
You are ignoring the author of this comment. Click to temporarily show the comment.
In general, it is better to avoid bidding the short suit.

In fact, I once developed a method (I called it “SBS” for “Strain Below Shortness”) which allowed shortness (stiff or void) to be shown in all situations without ever bidding the suit of the shortness.
It applied in all splinter cases and for Exclusion asks too.
It also worked whether there were three possible short suits or only two.
I'm sure this was theoretically superior to other schemes of encoding shortness.

Two alternative schemes for shortness encoding are
(1) “LMH” (Low, Middle, High) collapsing to “LH” when there are only two possible short suits.
Here, 1st step shows shortness in the lowest
possible suit, etc.
This scheme sometimes results in bidding the short suit.

(2)“natural with substitution”.
Here, you bid the strain of the short suit. When
not possible, use NT or trump suit to substitute when
bid in strain of the shortness is not available.

Haven't been playing SBS lately as it is an extra memory
strain and easily forgotten (usually in favor of “natural”) in the heat of battle.
May 13
1 2 3 4 ... 54 55 56 57 58 59 60 61 62 ... 94 95 96 97
.

Bottom Home Top