[Cyberyeti]

Imagine you have a fairly tightly defined auction for one of the hands which is then asked by partner about number of keycards.

Some responses may now become impossible.

eg you open a weak 2♠ and over 2N you show "good suit bad hand" then at some point partner keycards, if QJ1098x is not a good suit for you, then 0/3 is impossible, you would bid 3N with AKQ so 2+Q is also not happening either (either it's 3N or it's a good hand or it's a 1 opener).

Do you modify your keycard responses ? and if so how ?

Another situation is after 2♣-2♦ (negative not waiting)

[mw64ahw]

Its not come up for me, but logically a good suit bad hand could be AK or AQ so your keycard responses would be the same.
With a 3NT AKQ reply there would be no need to ask for keycards with only Js and lower being of possible interest.

With 2♣-2♦ (negative <8hcp rather than double negative) there is still the opportunity to have keycards so I doubt I would adjust.

[apollo1201]

I know there are some « adapted » BW answers for a hand that preempted who shouldn’t have 2 A e.g.

IIRC, but I might be wrong, it is 5C no A (then 5D asks about K and Q of trumps), 5D and above showing 1 A and the presence or absence of KQ or both.

But I’m all but expert😅

[DavidKok]

I don't do this, and would recommend not doing this. Of course it makes sense once it comes up - you can logically never have certain bids, and especially if those are the cheapest answers that's a waste of bidding space. But slam auctions over your own weak two are rare, and key cards are maybe not the most important question to ask. I think optimising this is a waste of time and memory, and I'd much rather invest in something more frequent (e.g. discussing which bids are and are not suitable for a weak two, or which responses are NF and how this changes in competition) or better (e.g. control asking bids, sweep/denial cue bids or some form of kickback).

[bluenikki]

Cyberyeti, on 2022-August-13, 04:06, said:

Imagine you have a fairly tightly defined auction for one of the hands which is then asked by partner about number of keycards.

Some responses may now become impossible.

eg you open a weak 2♠ and over 2N you show "good suit bad hand" then at some point partner keycards, if QJ1098x is not a good suit for you, then 0/3 is impossible, you would bid 3N with AKQ so 2+Q is also not happening either (either it's 3N or it's a good hand or it's a 1 opener).

Do you modify your keycard responses ? and if so how ?

Another situation is after 2♣-2♦ (negative not waiting)

When the teller is known to be weak, first step should be 0 or 2, second 1 without queen, third 1 with queen.

As to Ogust, there is no excuse for a scheme of replies that leaves ambiguity about what is defined to be a good suit in a bad hand.

[pescetom]

bluenikki, on 2022-August-13, 07:50, said:

When the teller is known to be weak, first step should be 0 or 2, second 1 without queen, third 1 with queen.

Some people play a 4♧ response to a weak two as a modified keycard ask with these or similar "Poor man" replies. IMO it's simpler and almost as effective just to play it as a normal 1430 RKCB kickback with subsequent Queen Ask.

[pescetom]

bluenikki, on 2022-August-13, 07:50, said:

As to Ogust, there is no excuse for a scheme of replies that leaves ambiguity about what is defined to be a good suit in a bad hand.

It's up to the partnership how and how tightly they define good suit or good strength, which makes sense as it has to reflect their agreement about what is a weak two opening in the first place.

[mikeh]

It’s fairly standard, at least in expert circles, to modify keycard over preempts

We play (in both my partnerships) that over 3D/H/S 4Cis keycard, with the responses being 0, 1, 1+ Q, 2, 2+Q

If we open 3C, then 4D is the keycard ask, same structure

Otherwise, no except that in exclusion the usual response structure is 0314, rather than 1430.

Edit

Btw, the reason we don’t do this over weak two bids is that we don’t play weak two bids. If we did, we’d use this structure.

[LBengtsson]

What is the point of having a RKCB convention as it stands over a weak two opening? Partner is never going to have 3 or 4 aces, or 3 keycards, and 99% of the time not likely to have two aces. As mikeh said, you need to modify RKCB structure and responses in certain auctions. That is just logic.

[pescetom]

LBengtsson, on 2022-August-13, 12:23, said:

What is the point of having a RKCB convention as it stands over a weak two opening? Partner is never going to have 3 or 4 aces, or 3 keycards, and 99% of the time not likely to have two aces. As mikeh said, you need to modify RKCB structure and responses in certain auctions. That is just logic.

It's not that clear cut, especially if the partnership is not regular or expert level (my apologies that I didn't note this was posted in expert forum). Partner is likely to have 1 keycard and certainly not 3 or more, but the normal 1430 structure can show 1,0 and 2 plus expose the Queen. Occasionally it will push too high to stop safely, especially in a low ranking suit, but you can foresee that risk and decide whether or not to take it (or agree that partner will suppress some replies) just like any other keycard ask. And of course you have the significant advantage of reduced memory load for a relatively low frequency ask.

[pescetom]

Out of interest I did a quick comparison (might well miss something) between Mike's custom scheme of 0,1-Q,1+Q,2-Q,2+Q (let's call it 011Q) and the classic 14,03,2-Q,2+Q scheme (1430), both with ask 4C.

Over a spades preempt, both can discover 0,1-Q,1+Q within 4S. But 1430 can also discover 2-Q within 4S and discover no specific king within 5S.

Over a hearts preempt, both can discover 0 within 4H. 011Q can also discover 1-Q within 4H but not 1+Q, 1430 can discover 1 but without information on the Q.

Over a diamonds preempt, 011Q can stop in 4D with 0 but not 1, 1430 with 1 but not 0.

So marginal gains/losses in specific situations for both, partly linked to preempt level and style, but I can certainly live with 1430.

[TylerE]

From my experience the impossible happens at least once a session.

[Cyberyeti]

Another thought for those sim minded say partner opens a strong NT (or 1x-1y-2N, or 2N) what point range of balanced hand makes 3 keycards more likely than 1 or 4 so you ought to switch to 0314 from 1430 ? I don't exactly know how you would sim it as knowing what the king of trumps would be is awkward, I'm guessing you'd need to do some post processing and say that 3 aces and 2 kings is 50% chance of 3 and 50% chance of 4.

[DavidKok]

If opener is strong balanced you want responder to describe their hand type, so that opener has captaincy on the slam auction. You're bidding blackwood with the wrong hand.

[Cyberyeti]

DavidKok, on 2022-August-17, 03:06, said:

If opener is strong balanced you want responder to describe their hand type, so that opener has captaincy on the slam auction. You're bidding blackwood with the wrong hand.

Absolutely wrong, the NT opener has the more tightly defined hand so you want their partner to ask.

[DavidKok]

I think you have it backwards. Shapely opposite balanced means the balanced hand is better placed to evaluate the wasted values of the 26 cards. Also strong hands take longer to describe, even after being limited to a narrow range. This is called the 'balanced hand principle', on scientific auctions with (strong) balanced opposite shapely the shapely hand should describe.

[Cyberyeti]

DavidKok, on 2022-August-17, 04:30, said:

I think you have it backwards. Shapely opposite balanced means the balanced hand is better placed to evaluate the wasted values of the 26 cards. Also strong hands take longer to describe, even after being limited to a narrow range. This is called the 'balanced hand principle', on scientific auctions with (strong) balanced opposite shapely the shapely hand should describe.

The problem is that the balnced hand often only knows a slam is in the offing when their partner asks. Also if you have 15, so often does partner if both hands are balanced.

[DavidKok]

Cyberyeti, on 2022-August-17, 05:51, said:

The problem is that the balnced hand often only knows a slam is in the offing when their partner asks. Also if you have 15, so often does partner if both hands are balanced.

That's why responder describes the hand. For example, 'interested in a slam with a balanced hand', or 'forcing with at least slam interest and a long diamond suit and heart shortness', etc.
When both hands have approximately 15-17 points in a balanced shape it does not matter which hand does the asking. When one of the two does not, there is benefit to describing the shapely, often weaker, hand.

[Cyberyeti]

DavidKok, on 2022-August-17, 09:44, said:

That's why responder describes the hand. For example, 'interested in a slam with a balanced hand', or 'forcing with at least slam interest and a long diamond suit and heart shortness', etc.
When both hands have approximately 15-17 points in a balanced shape it does not matter which hand does the asking. When one of the two does not, there is benefit to describing the shapely, often weaker, hand.

We play weak NT so tend to have auctions like:

1♣(4+)-1♥
1N(15-bad 19)-2♣(asking)
3♥(exactly 4324 17-bad 19 due to not opening 1♦ and not making other rebids)-4♣/♦ to set hearts, or keycard

Where responder knows an awful lot about opener's hand (which then hits the table) and is likely to do the asking while concealing his own shape so as to not leak info.

[Gilithin]

As Mike has mentioned, the main time this happens is after a preempt. Some pairs also play 3041 opposite strong or very weak hands and 1430 otherwise. If you want something more tailored than this, I would suggest searching out Ken Rexford's method, which uses a variety of different response structures according to the strength of the hand. I think he even made a self-published book on the convention but he certainly posted the main points on BBF if you search back a few years.