[jillybean]

---

Please tell me about transfers over 2nt. Thanks.

[mw64ahw]

4♣ as some Gerber variation or 6N is sufficient over 2N with that run of tricks and hcp count.

[mycroft]

Gerber is uncomfortable - that ♠K could easily be ♠QJ. Frankly, I'd prefer your other option - a "guess right T1" 6NT. But...

Sure you could play transfers here, and they're probably better - but "standard" in my playing areas is "after 2NT 18-19, the only NF call below game is PASS." Yes, it bites you very occasionally (but even then, they frequently have to guess right T1) but here, 3♥ is both standout and simple - provided partner won't pass it!

[DavidKok]

Transfers over 2NT, Wolff and 'everything forces game' are all fine but not amazing. Pick your favourite. The transfers are a little bit better, I think. You can overload the transfer to 3♦ with some artificiality, especially in a Walsh context.

Personally I think that all these solutions come one round too late. I want to tell partner I have a garbage hand with a suit before they jump to 2NT, and before fourth seat interferes. My beloved Dutch Doubleton handles the rest automatically, but even without that simple weak jump shifts reduce the problem by so much that I think you'll be fine without a gadget here.

[mw64ahw]

mycroft, on 2026-August-03, 09:35, said:

'that ♠K could easily be ♠QJ'

The probability of two uncontrolled suits in an 18-19 balanced hand is sub 1%.

[mikeh]

Assume 1x 1M 2N, uncontested, just because it makes it simpler to describe. But there is zero difference if there has been competition so long as opener’s last bid was a strength showing natural 2N and his LHO has passed.

3C is a transfer to diamonds, regardless of what opener bid at the one level. It is mandatory because responder could be about to pass. Note I’m assuming some form of Walsh whereby one responds to 1C by bidding the major with weak hands with a 4 card major and longer diamonds.

If responder bids over 3D, it’s gf and bidding is natural, with responder having shown diamonds plus the major in which he responded.

3D is a transfer to hearts

If responder bid 1H, then the 3D transfer is mandatory…responder could be about to pass. If he bids anything else, it’s gf

If over 3H, responder bids 3S, that’s a 4 card spade suit and 5+ hearts, gf. With 4=4 majors, bid 3H, transferring to spades.

Now after 1m 1S 2N 3D:

Opener only bids 3H with 4. With 2 or 3, he bids 3S with 3 and otherwise 3N. In this auction, responder can transfer again over 3N. So 1C 1S 2N 3D 3N 4D is 5=5 or better in the majors, unlimited

After a 1H response, 3H by responder, over 2N, shows 4=4 majors. Opener bids 3S with 4 and 3N otherwise (there are variants but I’ll keep it simple)

After a 1S response, 3H forces 3S, since responder may be passing

3S over 2N shows clubs with some reason for not bidding 3N. Opener assumes responder has a 4 card major and some interest in a club slam. Opener usually bids 3N but can go past 3N with a hand that is slam suitable within the context of the auction

It may sound complicated but it really isn’t.

Transfers to diamonds or responder’s major are mandatory because responder may by trying to get out.

Transfers to a new major are accepted only with 4 card support.

[pescetom]

 mw64ahw, on 2026-August-03, 10:36, said:

The probability of two uncontrolled suits in an 18-19 balanced hand is sub 1%.

Sub 0% looking at the hand of North. .

[mycroft]

mw64ahw, on 2026-August-03, 10:36, said:

The probability of two uncontrolled suits in an 18-19 balanced hand is sub 1%.

And wheeling out Ole Black (okay, the G entity - same thing) is going to find out about that 1% - how? Are you really going to bid 7 opposite "3 aces and 6 more points"?

Or, you're concerned that the 8 points you're missing are "A, A". Which surely is not much more likely than "AK suited".

And, on this hand, you get the joy of ...4♣-4♠-(X). Which, if it is off ♠AK...

All of which explains the rest of that paragraph you didn't quote:

Quote

Frankly, I'd prefer your other option - a "guess right T1" 6NT.

Oh, and as a director, I expect that the number of auctions that go ...4♣-4♠; (long pause) 4NT-6NT ("don't worry, I have the K in the other suit") in the fraction of the 99% that only have two aces to be *at least* 1%. And the number of those that get a director call... And the number of *those* that get awarded +690 by said director...

But having said all that, there are surely better uses of, say, the entire 3 level to, maybe, trigger cuebidding that will find out if it is that 1% (or "off two aces")? Maybe even opener ending up keycarding to find ♥AKQxxx+ instead (with responder correcting to NT)?

Which, to roll us right back to the beginning, is the OP's question!

[mycroft]

Okay, so, I did the thing. And here's what I'm getting (after *I think* getting rid of all the bugs. I'm not looking at 10 000 hands (but I looked at several sets of 100 hands) to see :-):

• 6NT off two aces: 346, off suited AK: 855, off for other reasons: 290, making: 8509
  
• 6NT off two aces: 347, off suited AK: 901, off for other reasons: 309, making: 8443
  
• 6NT off two aces: 315, off suited AK: 868, off for other reasons: 289, making: 8528

So, you know, not in fact "much more likely", in fact "about a third as likely". In fact, 6NT off a suited AK beats *all other reasons 6NT goes down double-dummy* by about a 3-2 margin.

Now, argue that "all" hands off two aces are going to go down, as you're not often going to have 5 tricks in the other suit to go with 7 hearts, (but how often do you get "spade to the A, spade back, take 3 spades and 9 reds"?) And I will freely admit that you're going to get a *lot more* misleads off a suited AK split between the defence. And that you're going to make more "other" contracts on the wrong opening lead too (looks like most of these are "♦Ax(x) and off a black ace", some "clubs or spades are AJxx(x) and off an ace.")

So, let's be generous and say, you go down in 6NT off two aces as often as you do for any other reason. That means "blast 6NT" makes ~93% of the time, and beats "get to 6NT after any other investigation" (which could help the opening lead). Does it beat "bid 4NT if off two aces, 6NT if not" often enough to make up for the "help"? Does it beat "play in 4NT/4♥ if we find out we're off suited AK, or 6♥ if we're off ♣AK or need to set up the clubs by ruffing, or 6NT if none of the above or (as in this hand) we need to protect the spades?" that you get by describing your hand and letting partner take control? Don't know, probably depends on the defenders.

But, of course, I bet RL defenders aren't as generous as I am positing here.

(really crappy) code in spoiler (to avoid screenspam):

Spoiler

"""From BBO: how often (and how) 6NT goes down on 18-opposite-7solid."""

import collections
import pprint

from redeal import Card, Deal, Evaluator, Shape, SmartStack, Suit, hcp, controls

DEBUG = False

# Auction went 1C-1H; 2NT
two_nt_rebid = Shape.from_cond(
    lambda s, h, d, c: min(c, d, h, s) > 1  # no singletons or voids
    and max(c, d) <= 5  # okay, this one's arguable
    and h < 4  # no raise
    and s < 5  # no 1S opener
    and (c > d or c == d == 3)  # no 1D opener
)
predeal = {"N": "65 AKQJT42 Q95 8", "S": SmartStack(two_nt_rebid, hcp, (18, 19))}
aces = Evaluator(1)
hands = collections.defaultdict(list)


def do(deal: Deal) -> None:
    south = deal.south
    result = deal.dd_score("6NS", vul=True)
    if result < 0:  # contract does not make
        if any(suit.controls == 0 for suit in {south.spades, south.diamonds, south.clubs}):
            type = "AKs"
        elif aces(south) < 2:
            type = "AA"
        else:
            type = "other"
    else:  # making
        type = "making"
    hands[type].append(str(south))


def final(n_deals: int) -> None:
    print(
        f"6NT off two aces: {len(hands["AA"])}, "
        f"off suited AK: {len(hands["AKs"])}, "
        f"off for other reasons: {len(hands["other"])}, "
        f"making: {len(hands["making"])}"
    )
    if DEBUG:
        for k, v in sorted(hands.items()):
            if k != "making":
                print(k)
                pprint.pprint(v)

P.S. Did the "arguable" in the code and allowed 6322s:

• 6NT off two aces: 334, off suited AK: 874, off for other reasons: 263, making: 8529
  
• 6NT off two aces: 362, off suited AK: 883, off for other reasons: 282, making: 8473
  
• 6NT off two aces: 334, off suited AK: 827, off for other reasons: 300, making: 8539

Doesn't seem to matter much.

P.P.S. For S+G, I looked at 6♥. Looks like, in a void, it's about 5% worse - usually because you've got two spade losers playing in North (6H by South, if you're playing transfer responses to clubs, is about 1% better than 6NT). Only ran 3000 hands for each direction, but they were pretty consistent.