[CSGibson]

So I was toying with Q ask responses, and I came up with a simple step response different than normal that may have some merit. In context, this is a precision partnership which uses a lot of artificial low-middle-high steps in response to "asking" relays.

My simple Q ask responses would be:
1st step: No Q
2nd step: Q, no K
3rd step: Q & K of lowest ranking suit
4th step: Q, K of middle suit, denies K of lowest ranking suit
5th step: Q, K of highest suit, denies other two kings
6th step: (If available) Q and an undisclosed source of tricks (in context of the auction, obviously; still toying with this response).

The theoretical advantage in my mind comes from the 2nd step, you can still explore grands with 3rd round control asks, but also you can sometimes stop in 5 when partner doesn't have his expected allotment of kings.

Potential problems include some memory burden at the very least, and maybe something in the wash in the ordinal of the king responses - I'm not sure. I'm posting this to invite comments.

[FrancesHinden]

I'll tell you what we play, which I think is slightly superior if you are responding to 'normal' blackwood i.e. most of your answers take you past the 5-level (or maybe if you were a level lower it's also relevant as they take you past game).

Simple case: no room below 5M. Example, spades are trumps, the Q ask is 5H. Now we play what you suggest i.e. 5S = no queen, 5NT = queen, no side king; others = lowest side king (following which a non-sign off asks for the king of the bid suit)
If there is more room, e.g. Q ask is 5D, now 5H = I have the queen, 5S = I don't, 5NT+ = I have a source of tricks in the bid suit (5NT = hearts) - usually this is a suit already bid by the responder in which case it promises KQ (KQJ if KQ implied already)
Over 5H, 5S asks for specific kings while a new suit asks for 3rd round control in the suit (5NT asks about hearts)

[Zelandakh]

I think it is usually best to look at this kind of idea by comparing directly with the alternatives. In the case of RKCB the example used should normally be spades, since if the idea has merit the other cases can always be converted by using Kickback if it is found to have merit. Here we have 2 auctions to consider, a 5♣ response and a 5♦ response. Taking the latter first, the proposal is:-

4NT - 5♦ - 5♥
=============
5♠ = no ♠Q
5NT = ♠Q, no side king (6♣♦♥ = SSTs)
6♣ = ♠Q, ♣K (6♦♥ = SKA)
6♦ = ♠Q, ♦K, no ♣K (6♥ = SKA)
6♥ = ♠Q, ♥K, no minor king
6♠ = ♠Q plus trick source

This looks remarkably close to "natural" responses to RKCB, ie

5♠ = no ♠Q
5NT = ♠Q, no side king, something extra (6♣♦♥ = interest in feature of this suit + maybe higher)
6♣ = ♠Q, ♣K (6♦♥ = SKA)
6♦ = ♠Q, ♦K, no ♣K (6♥ = SKA)
6♥ = ♠Q, ♥K, no minor king
6♠ = ♠Q, nothing extra

There is a small movement between the hands with ♠Q and no side king but nothing major. Given that the trick source hands are actually removed from 5NT I woudl have thought the feature ask style of follow-ups was actually even better for the proposed method. That would definitels be something I would want to look into.

Now we can compare with the Kings first method:

5♠ = no ♠Q
5NT = ♠Q, ♥K (6♣= SmKA, 6♦♥ = SSTs)
6♣ = ♠Q, ♣K, no ♥K (6♦ = SKA; 6♥ = general try)
6♦ = ♠Q, ♦K, no ♣K, no ♥K (6♥ = general try)
6♥ = ♠Q, no side king, something extra
6♠ = ♠Q, nothing extra

Here we have gained additional grand tries after partner shows a side king but lost them when partner shows no king. Which is more important? Well your experience may vary but I think the grand tries come up more often when partner shows a king than no king, so I prefer the last method. The other justofocation for this is that the "extra" being shown is often obvious from the context of the auction making those asks less valuable. Note also that the SSTs after 5NT do not necessarily have to be in diamonds and hearts. For example, after 1♠ - 4♦; 4NT - 5♦; 5♥ - 5NT, it should be completely obvious that 5♦ is an SST in clubs, not diamonds.

Let's move on to a 5♣ response. The proposed method here is

4NT - 5♣ - 5♦
=============
5♥ = ♠Q, no side king (6♣♦♥ = SSTs, 5NT = ? (SQA? general try?), 5♠ = ? (sign-off? something else?)
5♠ = no ♠Q
5NT = ♠Q, ♣K (6♣ = king ask, 6♦♥ = SSTs)
6♣ = ♠Q, ♦K, no ♣K (6♦ = ♥K ask, 6♥ = ?general try?)
6♦ = ♠Q, ♥K, no minor king (6♥ = ?general try?)
6♥ = ♠Q plus trick source
6♠ = ?♠Q, nothing extra?

Again, comparing with "natural" follow-ups:

5♥ = ♠Q, ♥K (5NT = SmKA, 6♣♦♥ = SSTs, 5♠ = ?)
5♠ = no ♠Q
5NT = ♠Q, no side king, something extra (6♣♦♥ = feature asks)
6♣ = ♠Q, ♣K, no ♥K (6♦ = SKA, 6♥ = general try)
6♦ = ♠Q, ♦K, no other side K (6♥ = general try)
6♥ = undefined, could be ♠Q + trick source if desired
6♠ = ♠Q, nothing extra

This structure looks to me to be clearly better. We have tries everywhere and have the correct amount of space devoted to the Q + no side king hand types. In fact, this strongly suggests that the optimal response for this hand type if wanting to be able to resolve the side feature is 5X + 1 step, where X is our side suit.

Finally, let's compare with Kings first:

5♥ = ♠Q, ♥K (5♠ = SmKA, 5NT, 6♣♦ = SSTs, 6♥ = general try)
5♠ = no ♠Q
5NT = ♠Q, ♦K, no ♥K (6♣ = SKA, 6♦♥ = SSTs)
6♣ = ♠Q, ♣K, no other side K (6♦♥ = SSTs)
6♦ = ♠Q, no side king, something extra (6♥ = general try)
6♥ = undefined, could be ♠Q + trick source if desired
6♠ = ♠Q, nothing extra

Once again, the same trade-off. Many more grand slam tries if partner shows a king witht he queen but inability to resolve what the specific extra is when partner has no king.

Overall, the structure looks playable but I would definitely suggest moving the "queen + no side king" hand to one step above 5 of the trump suit. I also think that you will gain more from feature ask bids over this than proper SSTs. Switching the "no extras" and "trick source" hands round between 5NT and 6 of the trump suit looks to be fine and might even be better on rare hands. If you do go ahead and play this I suggest you keep track of how often you needed a specific feature ask (or SST) after the Q, no K response (needed, not a nice-to-have) and how often you needed one after a Q + K response but were unable to. If the latter occurs more frequently then consider switching over to the Kings first method.

(Extra: Frances' method is also good (obviously) - it would be worth the time comparing that against the alternatives too.)

[CSGibson]

Zelandakh, on 2012-September-12, 02:58, said:

Let's move on to a 5♣ response. The proposed method here is

4NT - 5♣ - 5♦
=============
5♥ = ♠Q, no side king (6♣♦♥ = SSTs, 5NT = ? (SQA? general try?), 5♠ = ? (sign-off? something else?)
5♠ = no ♠Q
5NT = ♠Q, ♣K (6♣ = king ask, 6♦♥ = SSTs)
6♣ = ♠Q, ♦K, no ♣K (6♦ = ♥K ask, 6♥ = ?general try?)
6♦ = ♠Q, ♥K, no minor king (6♥ = ?general try?)
6♥ = ♠Q plus trick source
6♠ = ?♠Q, nothing extra?

Actually, the proposed method is:
4NT - 5♣ - 5♦
=============
5♥ = no ♠Q
5♠ = ♠Q, no side king (6♣♦♥ = SSTs, 5NT = ? (SQA? general try?), 5♠ = ? (sign-off? something else?)
5NT = ♠Q, ♣K (6♣ = king ask, 6♦♥ = SSTs)
6♣ = ♠Q, ♦K, no ♣K (6♦ = ♥K ask, 6♥ = ?general try?)
6♦ = ♠Q, ♥K, no minor king (6♥ = ?general try?)
6♥ = ♠Q plus trick source
6♠ = ?♠Q, nothing extra?

Thinking about this more, I would think that 5N over 5♠ would be asking for specific queens, while 6 level bids are grand slam tries for any 3rd round control.

On a frequency basis, I want to know whether partner has a Q for my grand try if I have AKJxx as my side suit, but any 3rd round control might do with AKxxxx. Having a way to distinguish between the two might prove very useful in grand tries.

And I assume that 6♥ after any of the king showing responses would be a last-train bid, confirming all keycards.

[Quantumcat]

Knowing about side suit stuff is important, alternatively you can just have keycarder make asking bids in the suits they care about. 1st step = lowest suit, 2nd step = middle suit, 3rd step = higher suit. Ignoring suits known to be short. Steps don't include trump suit or 5NT (trump suit yes if pard can have 3 or more keycards). Can continue making more asking bids (step lowest of remaining, 2nd step higher of remaining).

Responses: (aces already shown count as an x)
1st step = nothing (xxx+)
2nd = Q or doubleton (step by keycarder asks which)
3rd = Kxx+
4th = x (if possible)
5th = Kx
6th = KQ

Now keycarder uses the room they have to ask the questions that they find important, rather than the room being used to tell them stuff they may not care about.

[SteveMoe]

Chris,

In a Precision context, you have 1N (over majors) and perhaps 4♣ as beta/top control asks (though the latter is old-fashioned).
Thinking in terms of the 3 available triggers 1N, 4♣, 4N (or the kickback variant of your choice), once top controls are known, the asker can use the next trigger to elicit Queens using Roman responses: 03 | 14 | 2 Color | 2 Rank | 2 Shape (or 1430 if you prefer)

So for 1♣ - P - 1 Major - P - 1N (beta ask), a later 4♣ triggers the Roman Queen step responses. If 4♣ is unavailable (trump suit or epsilon ask) then 4N/Kickback serves the purpose. Of course, the Queen ask is OFF if pard responds 0-2 controls. We prefer repeat beta - but you can make your choice here.

I think simple is good.

PS - I play 1♥ nebulous GF and use 1N/1[{spades] and 2♣/1N(=♥) as beta.

Now about asking for Jacks....