Join Bridge Winners
All comments by Craig Zastera
You are ignoring the author of this comment. Click to temporarily show the comment.
Inviting will not “usually get you a bad score when partner declines.”

It will sometimes get you a bad score when partner declines *and* he can't make 2NT.

But inviting will often get you a good score vs. blasting 3NT when partner declines and he can't make 3NT.

And it will also get you a good score vs. passing 1NT when he accepts and makes 3NT.

Also, declining the invite need not occur “more often than not.” How often partner accepts an invite vs. declines is determined by your partnership agreements about how aggressively to invite.

A good agreement is to set this up so that partner will accept approximately 50% of the time. That way, the invitational strategy has maximal utility. If he accepted most of the time or declined most of the time, inviting would be less useful because you could just decide yourself and “usually” wind up in the same place.

In the case of 15-17 1NT openers, this works out roughly to declining with 15 and accepting otherwise, because the 15s occur roughly 50% of the time (exact percentage depends on policies about opening 1NT with 5 card majors and/or with 2=2=(45) shapes).

Whether inviting is the percentage strategy with a specific hand (as in this OP) depends on the probabilities of gaining vs. either of the other two strategies (i.e. passing 1NT or blasting 3NT in this case).

Simulations with this particular hand clearly showed that inviting gained overall at matchpoints relative to either passing 1NT or blasting 3NT.

VUL at IMPs, just bidding 3NT turns out to be the best strategy with this hand.

NV at IMPs it is quite close. If you accept the double dummy results, inviting is slightly still the winner. But it is close enough that real-world declarer advantage might be enough to push blasting 3NT into the lead.
Aug. 30, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
A set of 5000 deal simulations show that:
(a) when opener has a random, balanced 15 HCP hand,
3NT makes on about 39% of the deals
(2NT makes on about 73% of the deals)

(b) when opener has a random, balanced 16 HCP hand,
3NT makes on about 59% of the deals

© when opener has a random, balanced 17 HCP hand,
3NT makes on over 73% of the deals.

Inviting will tell you if partner has 15 HCPs (he passes) or 16-17 HCP (he bids 3NT). And that differences has quite a good chance of reflecting whether 3NT is a good bet or not.
Aug. 30, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
But in Walsh style (which I play), since the 1 response denies a four card major unless opening bid strength, a 2 rebid by responder to show specifically that hand type (5+ s, 4 s, opening values) seems like a reasonable and economical way to show that specific hand. Seems like this leaves plenty of room for any further investigation that might be needed.

With “xyz” added, one can extend Walsh to allow 1 response with 4M & 5+ s with only game invitational values.
With that hand type, after 1-1-1, use 2 ==> 2 then 2 to show 4 s, 5+ s, and specifically invite strength.

With this style, responder's *1* rebid (1-1-1-1) can be reserved for problem hands with fewer than four s.
The meaning we have found useful is that it shows a NT type hand (of various strengths) without a stopper.
Opener can convert to NT (usually 1NT) with a stopper, then responder can pass or raise to 2NT or 3NT depending on how strong he is. So this 1 covers a variety of hands and gets NT right-sided.
Whether this 1 could be extended to include yet other hand types is an issue we haven't explored, probably because the “xyz” structure already gives us so many options that we haven't encountered a need to try to leverage 1 further.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
If this hand is a 2 bid in their methods, then partner definitely shouldn't think 4 now shows a high card control in s (rather than an offer to play 4).
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
No reason to duck.
Most people will be playing 4.

It appears that they will be making at least 11 tricks (when both and finesses lose).

You must try to make at least as many tricks as the people in 4. If you go down in 3NT in the effort–who cares. Few matchpoints at risk in that case.

You could make as many as 13 tricks in NT (picture :Kx onside and the finesse working).

If the finesse is working, you can likely make 12 tricks in NT if s are 3-2 even with the finesse offside (or :Kxx or :K stiff onside).

So you win in dummy and lead a low to your J.
Later, you will take finesses. Chances of making as many tricks as people in 4 (likely 12) are not too bad.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
In addition to honor distribution, shape, point count and form of scoring, VULNERABILITY is also relevent.

“Winning” the race to 1NT is more beneficial (especially at matchpoints) when your side is not vulnerable.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Seems to me this hand has exceptional support and orientation for previous bidding.

Partner has gone out of his way to alert us (and the opponents) to potential danger for NT.

Thus, 4 seems strongly indicated to me vs. 3NT.

Give partner some hand like:
KQJT-KQTxxx-xx-x
A simulation indicated 3NTE to be 44% while 4W was just short of 80%.

Of course, in light of our strong rebid, partner does not need to be as strong as this example hand to continue over 3. That would, it seems to me, argue even more strongly for 4 rather than 3NT.

For 3NT to be right, in addition to the s running, we need probably two *fast* tricks in the majors. I see no particular reason why partner has to have those.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
It's a 15 count, but I agree that 3 rebid is less than obvious.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
As I don't play these methods, I'm just guessing what seems reasonable to me. I would think that those who do play this way would have discussed policy for a common-looking situation like this.

Anyway, my judgment opinion is that if not playng transfers, a natural 1NT would be reasonable here, so I chose 1 transfer to 1NT in your methods.

But a partnership might have a policy about never suppressing a 4 card major. If so, then obviously 1 (==> ).

Given the limited nature of the 1 opening (as well as the length ambiguity), I would think PASS with these very minimum responding values (for a “free” or NT bid) would not be unreasonable either, but presumably the partnership has some policy as to values promised by a “free” bid here.

Playing somewhat “sounder” minimum opening bid requirements in a standard system, I do not think I would Pass this responding hand.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
The meaning of South's 2 does not affect the support double interpretation of West's double of that call.

Support doubles are useful when responder's suit is s. You could take a poll, but my impression is that most who play support doubles use them for s as well as when responder has shown a major. I know we do.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Funny–Patrick Laborde upthread suggested that it was “obvious” that 4 would be a splinter raise of s with short s. I didn't agree with that either.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Wouldn't X here be a “support double” (3 card support)?
I am sure that is what it would mean in my partnerships.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
And why can't XX show 1st round control of s?
Will partner know that XX shows specifically 2nd round control of s?
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
You must be kidding.
I doubt if my partners would be confident that 4 here would be a splinter in support of s.
It might be an “auto-splinter” in support of s.
It might be just some strong, forcing “cue-bid” to show a great hand without clear direction, or a direction to be clarified later.

Splinters are usually jumps, but 4 here would not be, so that alone is enough to cast doubt on that interpretation.

I'm not saying that defining 4 here as a splinter in support of s is a bad idea. But I am saying that without some explicit agreement to that effect, it is far from clear that partner would so interpret that call.

BTW, in my partnerships, we use 3 here as specifically a game invitational strength hand with s. GF or stronger responding hands with s bid 3. I have found that explicitly distinguishing invitational from GF strength immediatley can be very valuable when the opponents compete further.
Aug. 29, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Don't understand why you find these figures “suspiciously low”.
Remember, the *defense* is double dummy too.
So for some contracts (typically high level, e.g. slams), real world declarers do *worse* than DD declarers.
For low level contracts (partscores), real world declarers do better than DD.
The turn-over point is around 4M/5m.

These comparisons of real world results vs. double dummy results are from Richard Pavlicek's excellent website.
They are based on a large database of deals from top level tournaments.

Here is a pointer–you can peruse the data yourself:
http://www.rpbridge.net/8j45.htm
Aug. 28, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
This hand is too good for a mere jump to 4.

4 might be OK if clearly a splinter in support of s, but there is no indication of this agreement, and it is not clear that is the ideal description anyway.

I think passing now will give me a good chance of learning more (i.e. partner's choice of rebid) before making a choice next round. I'm willing to bid (probably 5) on my own if LHO jams with 5 and partner passes.
Aug. 28, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
That is not really a valid criticism of “inviting.”

The “mesh” of honors between the two hands is always important in borderline 3NT deals and, in general, unknowable (I'm thinking of 1N-3N vs. 1N-2N-? auction types).

The benefit of the invite is to ask if partner is minimum or maximum for his 1NT opening.

Computer simulations of two relatively balanced hands with combined strength in the “borderline 3NT” range (around 25 HCPs) shows that each additional HCP raises the statistical probability of 3NT making by about 20%.

So if 3NT is statistically 40% opposite a wide range of random 15 HCP balanced hands, it will be 60% opposite random 16 HCP hands (and in the upper 70%s opposite 17).

The 40% will not be good enough (except VUL at IMPs), but the 60% will be good at all forms of the game.

I will add that specifically VUL at IMPs, it is doubtful that a NT invite is ever the percentage action. Since a below 40% make percentage is sufficient under those conditions, it is probably best for responder to simply make a “go/no go” decision himself (i.e. pass 1NT or bash 3NT.)
This is because there is a significant “penalty” incurred by inviting–when opener declines (with 15), if he can't make 8 tricks we lose to passing 1NT. When he has 16-17, he will accept, and we will be in the same place as the bashers.

But at other vulnerabilities and at matchpoints, there are responding hands that will show some net gain (statisically speaking of course) from inviting 3NT rather than responder's simply deciding himself.
Aug. 28, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
The computer simulations don't tell you too much about what would happen at the table on any one specific deal (because of double dummy deviations from at the table play).

But statistically, over *thousands* of deals, all with constraints such that they satisfy the actual given hand and bidding, they tell you all you need to know about whether this hand is worth a pass of 1NT, an invite to 3N, or a bash of 3NT.

You say that DD bias will persist even over thousands of deals? True. But the amount of this bias is pretty well known (around 3% - 6% for 3NT contracts–closer to the lower number when hands are balanced).

So you can just adjust your DD thousands of deals simulation results for this bias.

On this particular hand (OP), you will find from large simulations that the correct answer is quite clear.

On some other similar hand, DD simulations *might* yield sufficiently borderline results that a clear “answer” might not be available. That's OK too–it just means that that hand is a very close call (not this one though).
Aug. 28, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
This sort of question is very easy to answer using computer simulations.
Aug. 28, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I don't pre-empt at unfavorable unless I have a near perfecto Rule of 2/3/4 (i.e. within 2 tricks of my contract) so that my bid is highly descriptive for partner.

So this hand is an easy pass for me.
Aug. 27, 2018
.

Bottom Home Top