You are ignoring the author of this comment. Click to temporarily show the comment.
John,
I did not interpret OP this way.

Because he said that 4 was forcing to 4, I interpreted the 4 bid as a “fit non-jump.”.
That means partner expects us to make (or at least have a good play for) 4 and is showing his suit to give me info relevent to our side's possibly bidding 5 in competition if necessary (i.e. if they compete to 5 and I have a fit with his s, I am invited to bid 5).

You seem to think 4 here is just a random forcing bid showing a suit with lots of HCPs but not necessarily a fit. If I played that way, then I would play that 4 in competition is NOT FORCING (not what the OP is playing).
May 22, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I generated 100 random deals where you have:
:KJxx and a total of 8 HCPs.

I made no assumptions about lengths of your other suits (even though I suspect you probably meant that you have no suit longer than four cards).

I gave your partner 7 HCPs.

I gave your LHO one of:
4333 with 9 HCPs
balanced hand with no 4 card major and 9 HCPs
2=2=5=4 or 2=2=4=5 with 9 HCPs

I gave RHO a balanced hand with 16 HCPs.

Partner held the A and/or the Q on 39 of these deals.

So I estimate the probability he holds one of these spade honors at 39% (with a fairly large margin of error due to the small number of deals in the simulation).
May 22, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
(a) I wouldn't open 2 with this hand (but who cares?)

(b) If we have agreements such that partner's 4 has
created a force on our side to at least 4, then
clearly we are in a force over (4) in the sense
that we cannot defend (4) undoubled.

© But there are different kinds of forcing competitive
situations.

The “usual” is when our side has shown strength,
we've already bid our game, and the opponent's bid
over it. In these situations, the assumption behind
“forcing pass” is that the opponents cannot make
their bid, so logically we cannot pass them out
undoubled. Hence PASS can be used to send some
special message (traditionally, willingness to hear
partner bid on but “doubt” between that vs. doubling).

But another kind is when one of us has made a strength
and fit showing call that is forcing to at least the
next level of our suit, and the opponent's have made
a bid that is below that level (as here).
This is usually below the game level, e.g. when
our side has bid Drury or made a fit-showing cue-bid
below 3 of our major (or below 4 of our minor),
and the opponents have bid over that but below the
next level of our suit (or they have doubled).
In these situations, our agreement is that bidding
“our” suit at the cheapest level immediately is our
weakest action and discourages our side from bidding
further, while a PASS would suggest a slightly better
hand such that partner could (with good values)
bid game with hopes of making.

But should this latter also apply when our side's
cheapest level (to which we are forced) is game?
I think so, since it must at least be conceivable
that we might sometimes want to bid even higher
if forced to by their competition.

(d) Thus, I conclude that I should bid 4 here to show
weakness/minimum values for my previous bidding and
to discourage partner from bidding higher should they
compete to (5). My (sub)minimum values combined
with extreme misfit for partner's s would seem to
make this the message I want to send.

I should emphasize, though, that my choice of 4 for OP's problem is based on the assumption that we have the agreement that bidding our suit at the minimum level immediately when the opponents act over our Drury or (lower level) fit showing cue-bids is our WEAKEST ACTION (i.e. Pass would be stronger).

Some pairs have the opposite agreement in these situations, i.e. that PASS is our weakest action.
With *that* agreement, I would PASS in the OP's problem situation.
May 22, 2018
Craig Zastera edited this comment May 22, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I consider it one of the better books of recent years. My opening leads vs. NT took an obvious and immediate turn for the better after I had read this book a few times and pretty much could choose the winning answers on all the hands in it.
May 22, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Poor choice of options (again).
I think it's a “bad” bid but I'm not sure I'd go so far as to say “extremely bad” as my partners give me ample opportunity to witness worse.
But since it is “bad” (IMO), I would not make the bid, hence option 2 is unsatisfactory also.
May 22, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Paul,
I do not know what you mean by “double dummy bias”. That almost sounds like an oxymoron to me.

If flat hands with a particular HCP strength play better in NT contracts than other shapes with the same HCPs (and I'm not saying they do–more research needed), then that is simply a fact (if true, that is).

There is no “bias” involved in double-dummy analysis, which simply accurately reveals the objective potential of a deal with perfect defense and declarer play.

It would be a reasonable research project to compare the performance of many randomly generated balanced hands of a particular shape and strength (say, 4333s with 15 HCPs) with
randomly generated balanced hands with the same HCP strength but other shapes (say 4432s and/or 5332s) in 3NT contracts opposite similarly specified random dummy hands.

If differences were found, that would not indicate any “bias”–merely that certain shapes perform better than others in 3NT contracts. This might be a justification

But this is no different then, say, finding that 15 HCP balanced hands with two or more tens perform better on average than 15 HCP balanced hands with zero or one ten.
This would not indicate any “double dummy” bias, but merely identify an objective factor other than pure HCPs that is relevent to evaluating hand strength (for 3NT).
May 22, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I tried some 10,000 deal simulations on this hand.

I put this hand opposite random balanced hands with the additional requirement of “no 8+ card major suit fit.”
This is not because it is always wrong to play in NT with an 8+ card major suit fit (it's not), but just for consistency with my database of simulation results for pairs of “near 3NT” balanced hands that also uses this constraint.

I put no constraints on the defender's hands. Of course, this results in some deals where a defender will have some highly distributional hand which would likely have entered the auction. But again, this was done for consistency with my previous simulation results.

Anyway, here is how “average balanced 14 HCP hands” and
“average balanced 15 HCP hands” perform:

1. Random balanced 14 HCP hands opposite:
balanced 10 HCP hands: 3NT makes: 39.53%
balanced 11 HCP hands: 3NT makes: 59.23%

2. Random balanced 15 HCP hands opposite:
balanced 9 HCP hands: 3NT makes: 39.85%
balanced 10 HCP hands: 3NT makes: 58.66%

How did *this* hand (QJ3-QJ5-Q986-AQJ) fare in similar
simulations?
results opposite:
balanced 9 HCP hands: 3NT made 33.09% (2NT: 75.25%)
balanced 10 HCP hands: 3NT made 45.37%
balanced 11 HCP hands: 3NT made 70.17%

My criteria for downgrading (upgrading) a borderline hand is if it performs closer to an “average” hand with one less (more) HCP than to the average hand with the same number of HCPs, then I downgrade (upgrade) the hand.

Look at this “test hand” opposite the key 10 HCP dummies
in 3NT:
* it performed 5.84% better than average 14 HCP hand
* it performed 13.29% worse than average 15 HCP hand

This “test hand” performed much closer to an average 14 HCP balanced hand than to an average 15 HCP balanced hand in 3NT contracts, hence this hand merits a downgrade to 14 HCP status.
Therefore, playing 15-17 point 1NT openers, this hand should be opened 1 (planning to rebid 1NT) rather than 1NT.
May 21, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I don't think we need worry about partner having a “doubleton or less” in a major for his double.
It is enough to be concerned about the very real possibility that he might be 4-3 either way in the majors, so that our having a “responsive” type double available so as to find our 4=4 fit instead of our 4=3 could prove crucial.
May 21, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I do agree that real life declarer will not make this hand nearly as often as the 69% double-dummy figure suggests.

Still, I believe the optimal single dummy line will bring in 6 noticeably more than 50% of the time, justifying bidding it. The fact that s are 6-2 is not necessarily a bad thing as it means LHO likely has length in the other suits. If they don't split too badly (LHO with, say, 3=2=4=4 would be nice), it may be possible to reach 12 tricks on a cross-ruff line.

For example, imagine LHO with Qxx-Tx-QJ8x-Jxxx. This will make 6 single-dummy without doing anything too risky:
1. Win A
2. AK
3. AKQ (pitching dummy's )
4. AK, ruff in hand
5. (or ) ruff in dummy
6. ruff in hand (with J)
7. now or from hand scores dummy's 9 en passant.

On this particular lay-out, many other lines will work also, I just wanted to pick one that seemed plausible without seeing the defensive hands.
May 21, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Why not?
6 looks reasonable to me.

A quick 1000 deal simulation of the actual N/S hands (all small spots for the “x”s) with East having 6 decent s and 4 - 10 HCPs gave:
6 making on 692 deals

This is double dummy play and defense–real world declarer generally underperforms double dummy at the slam level, but only by a percent or two, so looks to me like this slam will make well over 50% of the time.
May 21, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Surprised to say that I like your partner's bidding on this hand.
I too voted for passing 4, but this actual hand for partner persuades me that action was unthinking.

Usually with four card support, partner would just raise s to the appropriate level. Raising to 4 would show a pretty good hand as our negative double might be based on as few as 8-9 HCPs.
So his cue-bid then raise must be better than that, but not a hand that could have bid 5 to demand we bid slam with a control (like some of the super-strong earlier examples).
His actual hand fits the bill perfectly and 6 is good.

Probably our best choice is 5 and then over partner's likely 5, continue with 5. If he can't bid the slam after that, *then* we can pass his 5 with a clear conscience.
May 21, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Just play “system off” over our 1NT overcalls. This is better anyway as transfers put the weak opponent on lead–not what you want to accomplish.

So here, 2 would be natural. 2 (cue-bid) is Stayman.
Over Stayman cue-bid, overcaller must show both min vs. max
and whether he has 4 cards in major or not (by using appropriate jumps).

Playing systems off, 3m jumps are natural, invitational, while 3M (unbid major) jumps are natural, choice of games.

With other (balancedish) invitational hands, bid 2NT.
If partner is minimum, he passes and you *may* miss playing
in a 5-3 major suit fit (when responder has invitational strength with a 5 card major).
But if overcaller is strong enough to accept 2NT invite, he bids any 3+ card (unbid) major he might have on the way to 3NT in case advancer has 5 in that major.
May 20, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I believe that ACBL rules require the void to be A, K, or Q.
May 20, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
No question about the 10's–valuable cards, particularly at NT and uncounted by standard HCPs.

But *FOR NT CONTRACTS* I doubt that adding bonus points for aces (as opposed to same number of points comprised of lower honors) is justified.

I have done many simulations in 3NT comparing hands with same number of HCPs with one being A & K heavy and the other laden with Qs and Js and found that the secondary honor heavy hands hold their own (sometimes outperforming) the hands with similar shapes and heavy on As and Ks with few or no Qs and Js.
May 20, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
In practice, partner can infer that you don't have a Yarborough or close to it from the flow of the auction. Not guaranteed, perhaps, but likely.

I don't want to punish partner for not giving up over their 3, so will not bid 4 with partner's strength and motivations ill defined.
May 20, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Well, I find your auction strange in multiple ways, not to mention your methods. I certainly would not have thought 4NT was keycard for s, among other things.
But you say it is, so it is.

I certainly see no reason why I should expect partner to have more than 5 s on this auction. I mean, sure–he *could* have more, but why should he?
Your 3 asked him (mainly) to bid 3NT if he could. When he couldn't, he made the cheapest plausible rebid, 4. I don't think that promises any more s than before.

So I'm reluctant to commit to playing in s when we might have two losers in that suit with 6NT making.
Give partner something like xxx-xx-QJx-AKxxx
May 20, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
But Paul–I said nothing about specific hands.
I am talking about large, random simulations–thousands of deals.
So the 4% advantage for declarer in 3NT vs. double dummy declarer will be a valid adjustment across such large random simulations, even though that adjustment could be very wrong (in either direction) on any specific deal.

BTW, you are quite right that the opening lead choice is critical. In fact, if you compare double dummy results vs. real world declarer play in 3NT *after the opening lead*, you will find that double-dummy now has around a 4%-5% advantage over real world declarer-a swing of around 9% from the same comparison made starting at trick 1.

Of course, studying simulation results (as in David Bird's books) can lead to better opening leads in the real world, too.
I'd love to see a study that required the statistically best opening lead (ala Bird's studies), but then compared full double dummy play vs. real world play thereafter.

It should be possible for real world defenders to approach always finding the “best percentage lead” as determined by Bird type simulations by carefully studying such results.
May 20, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I've done a lot of simulations of balanced hands opposite balanced hands with the combined in the borderline 3NT range (around 25 HCPs combined).

My experience is that with 10,000 deal samples, the margin of error for the estimate of making 3NT is around 0.2%-0.3%.

When the simulation is only 1000 deals, the margin of error is probably at least 1%, perhaps more.

So it is nice to do 10,000 deal simulations, but they take *so* long to run that the temptation is to make do with only 1,000 deal ones.

There have been comparisons of how 3NT with two balanced hands does with double dummy analysis vs. real at the table results. Statistically over a large sample, real life declarer does around 3% - 4% better than double dummy.
So, when doing NT simulation studies, just keep this figure in mind, and useful results can be obtained.
May 20, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Hah!–funny, John.

After reading your vehement defense of “there is only one option here”, I had to rush to push the “see all public votes” button to find out what it was.

I'll admit that I was somewhat surprised to learn that the one option was “penalties” (the choice I happened to vote for BTW, although some of the discussion here has made me wonder if it is really best).

I would suppose someone (else) might be inclined towards making almost as strong a statement about “take-out” being the “only option” needed.
May 20, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I'm not sure I agree with “completely unreasonable.”

My guess is that if I were playing with a good player as partner, but we were a new partnership and had not yet discussed this double, the odds would be that he would interpret it as “take-out” (2 places to play) and not as penalties. That is, I believe that modern expert practice is to play that most low level doubles in competitive auctions where we have not already agreed on a suit are “take-out” oriented. That is the default, IMO.

Thus, one could just as easily argue that making a “penalty” double in a competitive auction without prior discussion is completely unreasonable.
May 20, 2018
.

Bottom Home Top