[tomconlon]

Is there guidance on when responder should rebid a 5c M after a minor open, and a 1NT rebid by opener? We play TWCB so with invitational or better would bid TWCB. So, the case i'm asking about is when responder has 6-10p and a 5c M which opener has denied 4c support for by rebidding 1NT. Is it best to rebid that M, or pass 1NT? Or if sometimes one, sometimes the other, how would you decide which? Thanks if you have thoughts!

[DavidKok]

I always rebid the 5cM. Similar to always giving a Jacoby transfer with a weak hand with 5 when partner opened 1NT, it is percentage to do so here. Knowing that partner doesn't have 4 makes the odds slightly worse but they are still favourable.

However, this is in part because I will never rebid 1NT with a singleton in partner's suit. If you systemically rebid 1NT with a suitable minimum with a singleton, e.g. a 1=4=3=5 minimum on 1♣-1♠, or a 1=4=4=4, or the likes, it becomes more interesting to pass.

There are other considerations that have a minor impact on the auction, such as whether you play Walsh or not and what meaning you assign to 1m-2M, but I don't think it changes the approach (merely the inferences).

[sfi]

You won't go too far wrong if you always rebid the major. You're likely to take an extra trick or two in the weak hand and at least break even in the long run.

The hands where you might consider passing 1NT is when you have a max weak hand (9-10 or so) and a poor suit. That means your side will have the balance of power, your scattered values will be useful in NT, and you don't risk running into a bad trump break.

[pescetom]

DavidKok, on 2023-December-28, 16:00, said:

I always rebid the 5cM. Similar to always giving a Jacoby transfer with a weak hand with 5 when partner opened 1NT, it is percentage to do so here. Knowing that partner doesn't have 4 makes the odds slightly worse but they are still favourable.

However, this is in part because I will never rebid 1NT with a singleton in partner's suit. If you systemically rebid 1NT with a suitable minimum with a singleton, e.g. a 1=4=3=5 minimum on 1♣-1♠, or a 1=4=4=4, or the likes, it becomes more interesting to pass.

There are other considerations that have a minor impact on the auction, such as whether you play Walsh or not and what meaning you assign to 1m-2M, but I don't think it changes the approach (merely the inferences).

I basically agree.
I think that in general rebidding the 5cM is more advantageous at MP than IMPs and more with spades than with hearts.
Another minor consideration is if you play XYZ/NMF/TWCB and you know partner would stretch 1NT to get you there.

[mycroft]

Rebid the major - if your partner guarantees 2 cards for their 1NT. Check if they do, first.

If they feel that "partner's covered spades, I can rebid 1NT with my 1=4=4=4 (or 1=3=4=5)", well then. Convince them not to do that?

The other thing that leans away from "always rebid the major" is the answer to "when will partner raise with 3-card support rather than bidding 1NT?" If it's "never", then yeah, always rebid. If it's "with honours and a weak doubleton", probably do it anyway. If it's "pretty much any 3244 or 3235", then you're likely in a 7-carder, or with no ruffs in the short hand, and maybe you only want to rebid the major when you think it will take more of your trumps.

Of course, if the field is already in 2M (because their auction was 1NT-transfer), that's another whole question. We've covered that, and all the arguments around it, many times before if you care.

[mikeh]

Test

Here’s what I think is the best approach:

With a weak hand, say 5-bad 8 always rebid your major. At best the points are balanced, so you will usually but not always, score one to two tricks better in the major…I’m not saying you will make your contract, just that more often than not you’ll do a little better than playing 1N when the opps often hold more hcp than your side

When you hold a good 8 up to about 10, look at your suit and the location of your hcp and your shape. With a weak suit….say Jxxxx or Qxxxx with no spots, it’s usually best to pass 1N. Your hcp will help in the side suits and you’ll avoid a disaster in trumps when partner has xx or Qx opposite Jxxxx or Jx opposite Qxxxx. If your suit is reasonable, bid 2M.

[DavidKok]

Not to put too fine a point on it, but that's poor advice. It matches the sort of advice I was given a lot when I first started out. I think it makes responder feel more advanced or sophisticated, taking different actions with slightly different hands and claiming to be able to tell the difference. Always running with 5 is just significantly better.

A friend also claimed that it's actually better to sit with a good suit (as it'll be a source of tricks in NT) and run with a poor suit (as you won't have the values to establish a weak long suit in 1NT, but you are always making small trumps in 2M). This discussion sparked my curiousity, so I whipped up a quick dealer script to check double dummy simulations. I gave North a 12-14 balanced hand without a 5cM and at most three spades. I gave South a 5=3=3=2 with either 6-8 HCP or 9-10 HCP (split below) and either ♠J9542 (weak suit) or ♠AQT63 (strong suit), also split below. I then compared the double dummy scores of 2♠ by South with 1NT by North. All the usual double dummy caveats exist, but it's still better than nothing.

Here is the average double dummy bridge score of 2♠ minus 1NT (i.e. on this scale you get 1 IMP for scoring 20-40) not vulnerable, averaged over 5000 hands per example. A positive score indicates that 2♠ is better on average.
Weak suit, 6-8 HCP: +23.1
Weak suit, 9-10 HCP: +11.3
Strong suit, 6-8 HCP: +21.0
Strong suit, 9-10 HCP: +15.6

If people are interested I'd be happy to share the dealer script, seeds for the simulations, and/or discuss this in more detail.
To me the inferences are clear: pulling to 2M is always better, suit quality and HCP don't matter enough to influence that decision.

If there is a particular example of a hand where you think it is better to sit I'd be happy to discuss or simulate that as well. Simulations like this are a bit crude, and double dummy defence against 1NT in particular is very difficult. Still I don't really see a reason to pass 1NT.

[helene_t]

In French style, the 1NT rebid shows 2-3 cards in responder's suit, so rebidding the 5cM is usually right.

Playing natural responses to 1♣, the auction
1♣-1♠
1NT
could quite easily have a singleton spades in my preferred style (which is not French style), as a 2♣ rebid is almost always six cards. A consequence of this is that opener would raise with many hands with 3-card support - the 1NT rebid basically says "I am happy to play 1NT if you have five spades".

On the other hand,
1♦ - 1♥
1NT
will rarely have a singleton hearts and it's quite playable to let it guarantee 2-3 hearts. So here, responder should usually rebid a 5-card hearts.

1♦-1♠
1NT
is somewhere in between the two scenarios. I think most would rebid spades so you just accept that when opener has [1453], any rebid by opener can lead to a silly contract. So you just hope that opener won't get too many hands with that exact distribution and 11-15 points. But if you open 1♦ with 1444 (as is standard in North America), I think it's better to rebid 1NT with those hands, and as a consequence, responder should not usually rebid a 5-card spades.

Finally, if you play weak NT and agree that the 1NT rebid is unambiguously 15-17 balanced, you should usually rebid the 5cM.

If you play something more modern like T-Walsh or coded minors, the issue is obviously different.

[mikeh]

DavidKok, on 2023-December-29, 02:48, said:

Not to put too fine a point on it, but that's poor advice. It matches the sort of advice I was given a lot when I first started out. I think it makes responder feel more advanced or sophisticated, taking different actions with slightly different hands and claiming to be able to tell the difference. Always running with 5 is just significantly better.

A friend also claimed that it's actually better to sit with a good suit (as it'll be a source of tricks in NT) and run with a poor suit (as you won't have the values to establish a weak long suit in 1NT, but you are always making small trumps in 2M). This discussion sparked my curiousity, so I whipped up a quick dealer script to check double dummy simulations. I gave North a 12-14 balanced hand without a 5cM and at most three spades. I gave South a 5=3=3=2 or a 5=1=3=4 (split below) with either 6-8 HCP or 9-10 HCP (split below) and either ♠J9542 (weak suit) or ♠AQT63 (strong suit), also split below. I then compared the double dummy scores of 2♠ by South with 1NT by North. All the usual double dummy caveats exist, but it's still better than nothing.

Here is the average double dummy bridge score of 2♠ minus 1NT (i.e. on this scale you get 1 IMP for scoring 20-40) not vulnerable, averaged over 5000 hands per example. A positive score indicates that 2♠ is better on average.
5332 distribution:
Weak suit, 6-8 HCP: +23.1
Weak suit, 9-10 HCP: +11.3
Strong suit, 6-8 HCP: +21.0
Strong suit, 9-10 HCP: +15.6

5134 distribution:
Weak suit, 6-8 HCP: +73.5
Weak suit, 9-10 HCP: +54.0
Strong suit, 6-8 HCP: +59.6
Strong suit, 9-10 HCP: +48.1

If people are interested I'd be happy to share the dealer script, seeds for the simulations, and/or discuss this in more detail.
To me the inferences are clear: pulling to 2M is always better, suit quality and HCP don't matter enough to influence that decision (though notice the effect of suit quality with unbalanced 6-8) and if they do matter it is not clear in which direction (compare 9-10 balanced, where stronger suit = better for pulling, with 9-10 unbalanced, where stronger suit = smaller edge for pulling).

If there is a particular example of a hand where you think it is better to sit I'd be happy to discuss or simulate that as well. Simulations like this are a bit crude, and double dummy defence against 1NT in particular is very difficult. Still I don't really see a reason to pass 1NT.

Test

I didn’t address unbalanced hands but I would have thought it implicit that one only passes with 5332….it didn’t occur to me that anyone would think passing was correct with 5431 shape. Oh well, goes to show that one has to spell out everything.

What were your criteria for opener’s rebid? Does your 1N rebid include hands on which a 3 card raise was possible?

It is common in some circles (I play transfer Walsh and an unbalanced 1D in my serious partnerships so this entire topic is irrelevant but I certainly used to raise with Hxx and a small doubleton) to raise the major with Hxx and a small doubleton on the side…if you left those in your 1N rebid my sense is that that might well explain your results.

Kxx xx AQxxx Axx after a standard 1D 1S, I’d raise to 2S rather than bid 1N. If you’d rather rebid 1N, then that would, imo, tilt the analysis in favour of responder bidding 2M.

Finally, as I think is pretty obvious to anyone who has read, for example, the Bird and Anthias books on opening leads, double dummy analysis has huge flaws.

[DavidKok]

Thank you for your comment. I've removed the mentions of the 5431 shape to reduce the amount you have to spell out in the future.
I've rerun the simulations with the condition you specified for opener's concentration of values and shape. I won't post the results here, if you don't think there is merit in a double dummy investigation the numbers have no meaning.
The Bird and Anthias books make a lot of interesting points, and I think a lot can be learned from them. They oversell their results a lot and the reader needs to be very cautious when going through those books, but it is equally ignorant to disregard them entirely.

I stand by what I said earlier. I think rebidding the 5cM is better in the long run, and have not yet found a hand where I think passing is percentage.

[kgr]

Hi David,
Can you exclude hands on which opps are not passing in your simulation?
If opps are balanced then possibly passing 1NT is a better choice.
If you are able to run this simulation: can you then also add a count for the number of hands were passing is better? I would rather pass 1NT at MP's.

[awm]

I'll go with "never."

My style is to raise quite often with three-card support and a side doubleton, and to rebid 1NT with singleton on awkward patterns. This means rebidding the major will catch a singleton about as often as it catches three (two-card support is most frequent of course). I've also noticed that opponents quite frequently misdefend 1NT (it's a very difficult contract to lead against, for one thing) so I tend to do better than double-dummy would indicate (2M is less frequently misdefended).

Of course, hands with 5-4 majors or with 4-card support for opener's minor are removing 1NT anyway (to the second suit).

[bluenikki]

DavidKok, on 2023-December-28, 16:00, said:

I always rebid the 5cM. Similar to always giving a Jacoby transfer with a weak hand with 5 when partner opened 1NT, it is percentage to do so here. Knowing that partner doesn't have 4 makes the odds slightly worse but they are still favourable.

However, this is in part because I will never rebid 1NT with a singleton in partner's suit. If you systemically rebid 1NT with a suitable minimum with a singleton, e.g. a 1=4=3=5 minimum on 1♣-1♠, or a 1=4=4=4, or the likes, it becomes more interesting to pass.

There are other considerations that have a minor impact on the auction, such as whether you play Walsh or not and what meaning you assign to 1m-2M, but I don't think it changes the approach (merely the inferences).

Here is a different consideration.

If the rebid of the major can be 5-card, then for practical purposes it _is_ 5-card.

That forces opener to pass with a max and two trump honors doubleton.

[mycroft]

And why would opener do anything else? He has already said "I have [this many] HCP and a balanced hand, without 4 trumps [or whatever 3-card support would raise immediately]." Partner has said "okay, let's play 2M." She didn't invite, she didn't ask for advice, she said "let's play 2M".

You're down just as much with Ace-to-6 opposite-KQ and 7 points outside in 3 as you would be with Ace-to-5 in 2; and hands that can make game opposite KQ tight that wouldn't opposite Jx and the other 4 points outside seem hard to come by.

Sure, when they compete over 2M, it's harder to decide whether to raise to 3 - say with Hxx and potential ruffs; but it's harder to compete over 2M when it's a "non-fit auction", too.

Yes, "thinker's game", yes. "re-evaluate after every call", but 1NT is the ultimate "Captaincy" bid. If there was much to re-evaluate, there would have been a better rebid than 1NT.

[mikeh]

bluenikki, on 2023-December-30, 11:32, said:

Here is a different consideration.

If the rebid of the major can be 5-card, then for practical purposes it _is_ 5-card.

That forces opener to pass with a max and two trump honors doubleton.

So what?

Ok, semi-facetious question but your question reveals a pretty basic misunderstanding of standard bidding.

2M says ‘I have zero interest in what you have beyond what you have announced. I have zero interest in game.I do NOT want a raise or a 2N rebid. I want…no, I require…that you say the magic word….pass.’

[sfi]

DavidKok, on 2023-December-29, 02:48, said:

If there is a particular example of a hand where you think it is better to sit I'd be happy to discuss or simulate that as well. Simulations like this are a bit crude, and double dummy defence against 1NT in particular is very difficult. Still I don't really see a reason to pass 1NT.

The sort of hand I'm thinking of is

---

I'd be surprised if 2S plays better than 1NT, particularly if partner is allowed to raise on a three-card suit. Even if it comes ahead double-dummy (and I don't think it will), my experience is that 1NT is better in practice.

I do know that bidding 2S would lead to a different contract than the other table in any serious match around here, so that may be valuable at times.

[pescetom]

sfi, on 2023-December-30, 15:44, said:

The sort of hand I'm thinking of is

---

I'd be surprised if 2S plays better than 1NT, particularly if partner is allowed to raise on a three-card suit. Even if it comes ahead double-dummy (and I don't think it will), my experience is that 1NT is better in practice.

I do know that bidding 2S would lead to a different contract than the other table in any serious match around here, so that may be valuable at times.

Around here most of the field would not be in 2S and I would rather be with them.

But it's an extreme and low frequency example. Change the spades to even just JT532 and it starts to look better.

If BBO has made a New Year resolution to restore the arbitrarily disabled 'tricks' function in Dealer then I will be happy to simulate.

[DavidKok]

kgr, on 2023-December-30, 06:15, said:

Hi David,
Can you exclude hands on which opps are not passing in your simulation?
If opps are balanced then possibly passing 1NT is a better choice.
If you are able to run this simulation: can you then also add a count for the number of hands were passing is better? I would rather pass 1NT at MP's.

I most certainly can. It depends a bit on how you define an overcall (what is the minimum, which hands are suitable for a takeout double etc.) but I can plug some mainstream ideas and hope to approximate the style of actual players, hopefully splitting the difference between erring on the side of caution and erring on the side of aggression.
For simplicity I defined an overcall at the 1-level as a five card suit with 8+ HCP and allowed takeout doubles starting at 11 points and at most two clubs (I'm going with a balanced club system here, sorry) or 14 points and a balanced shape. Fourth hand can overcall a six card suit at the two level starting at 10 points, or make a takeout double with at least 54 in the reds either way (the auction will have gone (1♣)-P-(1♠)-? from their perspective) or any 15+ HCP balanced hand.

I've kept the other parameters the same, so South has a 5=3=3=2 and the good and bad suit are as I described earlier. I simulated the average score of 2♠ minus that of 1NT, and the frequency with which 2♠ outscores 1NT. The results can be found below (note: a previous version of this post had a mistake here, and I've updated the numbers by splitting them into Win/Draw/Loss, with 'Win' being 2♠ being better and 'Loss' being 1NT being better):
Weak suit, 6-8 HCP: +19.9, 50.5%/23.9%/25.6%
Weak suit, 9-10 HCP: +9.8, 56.2%/7.4%/36.4%
Strong suit, 6-8 HCP: +16.9, 52.1%/18.1%/29.8%
Strong suit, 9-10 HCP: +12.8, 55.5%/4.4%/40.1%

As an aid to interpreting these numbers, the standard deviation on this average difference of scores is approximately 1.0-1.3. This means little to nothing for possibly flawed methodology, but at least my sample size (5000 deals per test) is sufficient, making all these results statistically significant at the 8-15 sigma level.

sfi, on 2023-December-30, 15:44, said:

The sort of hand I'm thinking of is

---

I'd be surprised if 2S plays better than 1NT, particularly if partner is allowed to raise on a three-card suit. Even if it comes ahead double-dummy (and I don't think it will), my experience is that 1NT is better in practice.

I do know that bidding 2S would lead to a different contract than the other table in any serious match around here, so that may be valuable at times.

I ran this one as well, with a result of -0.2, 50.7%/8.6%/40.7%. A very nice example, and my simulation shows no statistically significant deviation from zero difference between the two contracts at my sample size. At MPs it seems that pulling is considerably better for double dummy play though. This is somewhat surprising to me, I expected a hand this extreme to score well when passing.

Edit: I got curious about that last one so I ran it again with double the sample size (10,000 fresh deals with South having that exact hand and the other three hands matching the descriptions given above). The result was +2.0, 52.1%/8.2%/39.7%. The higher sample size means that the standard deviation is a bit lower, but I think this should still be interpreted as zero at IMPs for all intents and purposes.

[sfi]

DavidKok, on 2024-January-01, 16:04, said:

I ran this one as well, with a result of -0.2, 59.3%. A very nice example, and my simulation shows no statistically significant deviation from zero difference between the two contracts at my sample size. At MPs it seems that pulling is considerably better for double dummy play though. This is somewhat surprising to me, I expected a hand this extreme to score well when passing.

The difference in scoring methods is interesting indeed. There is one other factor that (probably) comes down on the side of passing, which is the 1NT rebid with a singleton honour. But doing that is not necessarily mainstream and would be part of partnership discussions beforehand.

[sfi]

pescetom, on 2024-January-01, 15:14, said:

But it's an extreme and low frequency example. Change the spades to even just JT532 and it starts to look better.

Absolutely - I'm happy to agree that the examples I'm thinking of are both extreme and low frequency. After all, my first sentence was "you won't go too far wrong if you always rebid the major."