[kenrexford]

I have developed what I believe to be an improvement on ace-asking that beats RKCB.

The idea starts with grouping replier's holding into one of three types:
A. Weak Hands (typically less than opening strength, where "two with the Queen" is the maximum pssible response
B. Moderate/Medium hands (neither weak nor strong)
C. Strong Hands (where expectations are that partner usually has at least 2 key cards

These groupings are pre-determined by agreement. For example, "weak hands" include passed hands, preempts, simple raises, etc., while "strong hands" include 2♣ openings, jump shifts, reverses, "Serious 3NT" bidders, etc. "Medium" hands are any hands that do not qualify for Weak or Strong treatment.

For each grouping, a specific key card structure is used: WKCB, MKCB, or SKCB (Weak, Medium, or Strong Key Card Blackwood).

For each group, there is a specific response with which replier forces slam and goes immediately to the "next question." The response causing that wrap-around is 2+Q for WKCB, 3+Q for MCKB, and 4+Q for SKCB. So, for example, a MKCB replier with 3+Q immediately skips to specific Kings (e.g., 4NT-P-6♣).

If Replier has the "go directly to slam" holding, but minus the Queen, he indicates whether he has the trump King in his key card count or not. E.g., a WKCB replier with "two without the Queen" bids 5♥ with A+K but 5♠ with A+A.

If replier has one less key card than the "go directly to slam" holding (whether with or without the Queen), he bids the cheapest reply available (5♣ if 4NT was the ask -- note that KICKBACK is needed for this to work). Asker can then ask for the Queen (+1); if replier has it, he bids pas the 5-level signoff; without it, he signs off with the trump King as one of his key cards or bids +1 with all Aces as key cards. For example, if 4NT(WKCB)-P-5♣, 5♦ asks; 5♥=A, 5♠=K, 5NT+=1+Q. If asker needs specifically Aces+Q, he bids +2 (4NT-5♣, 5♥?).

With one key card worse, replier bids 5♦ (if 4NT ask, parallel if kickback); +1 asks for Queen (cannot distinguish King from Aces).

So far, this covers expected holdings and allows the "is the key the trump King?" when it MOST LIKELY MATTERS.

What about holdings outside of expectations?

With WKCB, this is not possible.

In MKCB or SKCB, this is possible either way. What you do is to apply the principle of 0 or 3, 1 or 4, 2 or 5 and find the corresponding bid. If the corresponding bid is ask+1, you get the Queen and King ask. If the corresponding bid is ask+2, you only get the Queen-ask. If the corresponding bid is ask+3 or ask+4, you pick one of these two to indicate whether you have or do not have the trump Queen. Thus, for example, consider SKCB. You do not expect a SKCB to only have one key card. None of the normal responses show this. But, the SKCB with that holding can bid 5♥ or 5♠ as either "4 without the Queen, indicating if trump King or not" or "1 key card, indicating whether the trump Queen is also held." Still 1 or 4, but a different message attached.

This description seems complicated, but if you thinkl it through, it somewhat means to asnwer WKCB always, but to subtract 1 from your key card count if MKCB, -2 if SKCB, and applying the 0/3, 1/4, and 2/5 rules logically. Or, memorize a table of answers.

The end result of this basic approach is to allow you to tell whether the trump King is a key card when it most often will matter.

This can then be expanded. After a Queen-ask, Queen denied, +1 asks for the trump Jack instead (with either 5NT or six of an alternative strain as the "no" answer).

Also, if the trump holding is already known (maybe cuebidding or asking bids), then redundancy avoidance kicks in. Suppose, for instance, that replier already established that he holds the trump King and Queen. Showing those cards twice is dumb. So, if he "shows the trump Queen" he is known to have (or known to not have), that shows the "first priority" critical card. Denying the King (a stronger position than having the King) shows the "second priority" critical card. As an example, a replier who has shown the trump King and Queen earlier might "show the trump Queen" with, instead, the King of partner's side suit and "deny the trump King" with partner's side suit Queen. A crazy example might be ...4NT-5♣, 5♥ (do you have the King and Queen of my diamond suit?). Priorities are established by a pre-agreed list.

I am working on a write-up of this with examples and the like. But, I thought I would share this, to get any feedback, questions, comments, and the like that folks might be willing to share.

[PrecisionL]

Interesting, Ken but RKC 1430 is hard enough for my many club partners. I don't think I will tell them about your idea(s).

I gave up RKC with my tournament partner years ago. Now Kantar will have to revise his RKC book again!

RKC: The Final, Final Word (6th edition, 2013?) - thanks to Ken

This post has been edited by PrecisionL: 2012-October-11, 20:28

[kenrexford]

PrecisionL, on 2012-October-11, 14:21, said:

Interesting, Ken but RKC 1430 is hard enough for my many club partners. I don't think I will tell them about your idea(s).

I gave up RKC with my tournament partner years ago. Now Kantar will have to revise his RKC book again!

This is what I hear constantly, and it is true. My idea is not for those people. I doubt they could handle the cuebidding, either!

[bluecalm]

My intuition is that RKCB is good convention and it rarely causes problems if rest of the system is well designed.
That's why I am reluctant to analyze your idea, the potential gain is just too small even if it's the best possible solution.

[CarlRitner]

My only comment is that you'd probably want to limit comments to those playing at a level where the gain from using this method will definitely outweigh the losses from a forget or misunderstanding. I think you are talking expert+ where the plus indicates no problems with very memory intensive methods.

It might not suffer too much from the "use it or lose it" syndrome, since keycard asks are fairly common.

[PrecisionL]

Interesting, Ken but RKC 1430 is hard enough for my many club partners. I don't think I will tell them about your idea(s).

I gave up RKC with my tournament partner years ago. Now Kantar will have to revise his RKC book again! [6th ed. ?]

kenrexford, on 2012-October-11, 14:28, said:

This is what I hear constantly, and it is true. My idea is not for those people. I doubt they could handle the cuebidding, either!

Danny Kleinman self published a book on 100 RKC hands, it is in my libarary somewhere. Anyway, 50 bid by club players and 50 by experts and they both had accidents or used it incorrectly about the same amount, ~45% if I remember correctly.

[cloa513]

You have to split the slam potential from the strength- they are related but not the same
(any suits) Axxxx Axxx xx xx has moderate slam potential if you have 5-4 fit but low strength
Axxxx Axxx xxxx has even higher slam potential but no more strength.

[kenrexford]

cloa513, on 2012-October-11, 21:44, said:

You have to split the slam potential from the strength- they are related but not the same
(any suits) Axxxx Axxx xx xx has moderate slam potential if you have 5-4 fit but low strength
Axxxx Axxx xxxx has even higher slam potential but no more strength.

True, but I am not sure what conclusion to draw from this fact.

[Zelandakh]

Ken, I think you should start with a write-up of MKCB and then give WKCB and SKCB as optional variations, much as Kantar gives 1430 as the basic form and then adds 3041 for strong hands, void-showing, etc as extras. So, to see if I have understood, here is my tentative write-up of MKCB...

(spades agreed) - 4NT
===
5♣ = 2 (5♦ = asks for ♠K and/or ♠Q, 5♥ = asks for ♠K and ♠Q)
5♦ = 0 or 1 (5♥ = asks for ♠Q)
5♥ = 3 - ♠Q including ♠K
5♠ = 3 - ♠Q not including ♠K
5NT = 3 + ♠Q with no side king but an extra
6♣♦♥ = 3 + ♠Q with corresponding side king
6♠ = 3 + ♠Q only

Is this right? If so, what does Asker do if they need to know the difference between 0 and 1? Also, what about voids? If not right then please tell me where I went wrong!

[bluecalm]

So now we are bidding 5NT/6X just to say if one of our keycards is K of trumps (info which is about never needed imo). We are losing other stuff though comparing to standard rkcb where the last answer is 5S. This stuff is:
a)showing voids with even/uneven number of keycards
b)various asks from RKCB bidder side (grand slam invite for xx/Qx in a suit, choice of slams)

If I am getting it right it looks terrible to me.

[Tomi2]

I am quite confortable with Tomiwood responses to the keycard ask
1st step: 0, 2 or 4 Keycards, next step asks the trumpqueen
2nd step: 1KC with OR 3KC w/o OR 5 with trump queen
3rd step: 1KC w/o OR 3 with OR 5 w/o trump queen

after the 2nd and 3rd step, because of using kickback and minorwood we can ask for kings below 5suit very often

additionally we gain the 4th response step for "I don't have too many keycards but I do like my hand very much" for example if you have a weak hand with no keys but a void and long trumps or some outside tricksource partner can not see
jumps in Tomiwood show "enough keycards and a specific void in this suit"

5suit signoffs after all steps are "pass or correct" same as in 14-30 when you sometimes cant read the 1 or 4... actually because of showing the queen directly I had fewer spots where i could not read partners resonse compared to the 1 or 4 accidents. A player with 2 keycards and now trump queen in his hand must be cautious because he might not be able to read partners response

[kenrexford]

Zelandakh, on 2012-October-12, 03:16, said:

Ken, I think you should start with a write-up of MKCB and then give WKCB and SKCB as optional variations, much as Kantar gives 1430 as the basic form and then adds 3041 for strong hands, void-showing, etc as extras. So, to see if I have understood, here is my tentative write-up of MKCB...

(spades agreed) - 4NT
===
5♣ = 2 (5♦ = asks for ♠K and/or ♠Q, 5♥ = asks for ♠K and ♠Q)
5♦ = 0 or 1 (5♥ = asks for ♠Q)
5♥ = 3 - ♠Q including ♠K
5♠ = 3 - ♠Q not including ♠K
5NT = 3 + ♠Q with no side king but an extra
6♣♦♥ = 3 + ♠Q with corresponding side king
6♠ = 3 + ♠Q only

Is this right? If so, what does Asker do if they need to know the difference between 0 and 1? Also, what about voids? If not right then please tell me where I went wrong!

In mkcb, with 0 bid 5H as 0 or 5S 0+Q. 5H & 5S are thus 0 or 3.

[kenrexford]

bluecalm, on 2012-October-12, 03:39, said:

So now we are bidding 5NT/6X just to say if one of our keycards is K of trumps (info which is about never needed imo). We are losing other stuff though comparing to standard rkcb where the last answer is 5S. This stuff is:
a)showing voids with even/uneven number of keycards
b)various asks from RKCB bidder side (grand slam invite for xx/Qx in a suit, choice of slams)

If I am getting it right it looks terrible to me.

the K issue is more important than you think, imo. Plus, some of your concerns about losses actually are addressed in my draft writeup already. But, I will make sure that I consider these ideas more.

[bluecalm]

It would for sure go long way to convince me (and others) to your idea if you give some example hands.
I can't remember one hand I needed to know if partner has K of trumps among his key cards while I rememeber many when I needed to make grand slam try in a suit or offer choice of slam (or wish I had that available).

[Fluffy]

I will ask my partner if he wants to memorice all this methods.

Looks like some slams will be played missing KQJ or KQ10 of trumps.

[DinDIP]

Just to show -- as though further evidence were needed -- that good (and bad) ideas have many authors, the idea of varying responses to a keycard ask on the basis of strength already shown was advocated by Jeff Rubens in the original article on Kickback (The Bridge World, January 1981). He too suggested bidding beyond the safety threshold (5T), to show kings, with hands with the maximum number of keycards that can reasonably be expected, rather than using that space to show hands with a void.

David

[kenrexford]

DinDIP, on 2012-October-14, 04:42, said:

Just to show -- as though further evidence were needed -- that good (and bad) ideas have many authors, the idea of varying responses to a keycard ask on the basis of strength already shown was advocated by Jeff Rubens in the original article on Kickback (The Bridge World, January 1981). He too suggested bidding beyond the safety threshold (5T), to show kings, with hands with the maximum number of keycards that can reasonably be expected, rather than using that space to show hands with a void.

David

Wow. Do you know how he used the extra space?

[glen]

DinDIP, on 2012-October-14, 04:42, said:

... January 1981 ...

The thread title needed to be Varying replies to RKCB is so old-fashioned. In the February 1981 issue JR outlines the Kickback Slam Convention, but in regard to replying above 5 of the trump suit only suggests "profitable revision possible", and "If Kickback becomes popular I'll publish more about it"

[Siegmund]

Will be interested to hear more about it. There is certainly room to greatly improve its efficiency.

I am slightly scared of it, because of a similar experience. With two of my regular partners, we use a variation on Brashler's Sweep Cues, and use 5NT GSF a lot... with different scales of responses according to whether the the person who replies has said nothing about trump quality, already promised good trumps, or already promised bad trumps... and it's the single hardest-to-remember part of the whole system for us.

[dake50]

Two comments:
I have cherished the 5 or 7 hand where the missing K-trump is finessable.
Finesse wins 13 tricks; loses not even small slam makes.
Try the game theoric payoff matrix at IMP.
But how to pinpoint which A ot K-trump is missing??
***
I have emulated systems where slam info is demanded/shown much before 4NT.
Thus little need for RKCB.
Of course this is more discussion than "Key card,partner?"