[PrecisionL]

Keylime and I are auditing our Copious Canape Club System, C3, for improvements and the question of switching from Beta to AKQ points has been suggested for slam exploration.

I have no experience with AKQ points (other than reading The Ultimate Club) and I am concerned that the switch would be difficult to accomplish. I prefer to readjust the Beta responses (1-2, 3, 4, 5 controls, etc).

We solicit comments and suggestions from those who have experience playing either scheme. Thanks.

Note: Currently use Turbo and Denial cue bidding.

[akhare]

 PrecisionL, on 2012-July-28, 15:24, said:

I have no experience with AKQ points (other than reading The Ultimate Club) and I am concerned that the switch would be difficult to accomplish. I prefer to readjust the Beta responses (1-2, 3, 4, 5 etc).

I don't have much experience with Beta responses, but IMO, AKQ (or Queen Points) in combination with symmetric relays pretty much trumps everything else out there.

Note that you will have to figure out the optimal values for the base QP for opener and responder. In our experience, a 16+ club goes pretty well with a 10 QP base for responder and we make a GF response with 5 QPs. Note that it's OK for opener to have fewer QPs than the base as long as reverse relays aren't employed.

Also, you may want to discuss things like the terminator puppet (most people use 4D for it), RKC after shape resolution (instead of QP ask), zoom and situations where responder may bid on after opener's attempted signoff.

Finally, you may want to decide which DCB scheme works best for you in conjunction with AKQ points. We started off with the Moscito scan (which scanned for A/K on the first round and Q on the second round etc. IIRC), but switched over to awm's parity scanning method.

[Free]

I haven't worked much with controls but I did some research in the past for MOSCITO. The methods I compared were:
A: AKQ-points with denial cuebidding (1st scan skips suits with 1 or 2 top honors)
B: AK-points with denial cuebidding (1st scan skips suits with A or K - not both)
Both methods started with a base, as far as I remember AK-base was 1/3 of AKQ-base (easy, because MOSCITO uses 3, 6 and 9 as bases).

My conclusion was the following:
Both methods work very similar and do the job most of the time. There were a few hands where AK did better (= stayed lower to get the necessary information) than AKQ, but there were more hands where AKQ did better than AK. I analyzed less than 500 hands, so I'm not sure if we can say one method is better than the other in general based on this set.
However, I discovered a pattern. AKQ is better for slams based on power (read balanced hands) because you quickly get an idea of the combined top honor strength. Also, the captain usually has his top honors scattered around, so he can determine what partner has in each suit. AK is better for distributional slams, for example when you have a running suit and need to find out if you have all the controls and maybe a Q somewhere. Note however that low-level RKC followed by SAB's can usually handle this equally well.

The biggest drawback of AKQ-points in my experience is that you can't make them as flexible as AK-points. You have to let go of HCP in many cases and just bid based on AKQ-points (hands with lots of Jacks and Queens for example).

Basically, once shape is determined (or almost determined - 12/14 cards - and you skip to slam bidding immediately), I concluded that the following list of methods would do the job very well when you have an accurate AKQ-base:
3NT = to play
4♦ = terminator (not always with 12/14 cards - needs agreements!)
step 1 = AKQ-ask followed by denial cuebids
step 2 = RKC in longest suit (when equal, use ♥>♠>♣>♦) followed by SAB's/TAB's
step 3 = RKC in 2nd longest suit (when equal, use ♥>♠>♣>♦) followed by SAB's/TAB's
step 4 = RKC in 3rd longest suit (when equal, use ♥>♠>♣>♦) followed by SAB's/TAB's
step 5 = RKC in shortest suit (when equal, use ♥>♠>♣>♦) followed by SAB's/TAB's
higher steps = absolute signoff, no zooming allowed.
When you really want to, you can replace step 2 with AK-ask and move the RKC's a step higher.

With 12/14 cards I mean:
4333 = 3333 (4-card unknown)
4432 = 4433 (doubleton unknown)
5332 = 5333 (doubleton unknown)
4441/5440 = 4440 (3-suited and short suit is known)
6322 = 6222 (3-card unknown) / can also be 7222/6322 -> 6222
etc.

Note that I only analyzed 1 method of denial cuebidding after both asks. Those were the only ones I knew at the time, maybe other methods make things more/less efficient in some cases.

[straube]

I think Free is obviously correct when he says that QPs work better for less distributional slams while controls work better for distributional slams. I read about a system where relay captain was able to ask for QPs (S1), controls (S2) or aces (S3). I suppose he was able to scan for other cards whichever route he took. Significant for me was that the system designer (Mark Abrahams) chose to give preference to QPs over controls.

Akhare and I use a combination of QP ask (S1) and RKC, but we decided to simply arrange the RKC suits C/D/H/S regardless of partner's distribution. I've found it very useful for investigating minor suit slams. It's just very difficult or impossible for either DCB or PCB to determine which cards partner has before the 5m level. At least RKC helps us avoid hopeless slams. Not saying this is necessarily optimized, but when we're interested in a major we have much more room to employ QP ask and Parity Cue Bidding and usually choose this method over RKC.

[chasetb]

I have no experience with AKQ points (if anyone would like to teach me, please message me!) but I do know a little about Beta. The system I play it in isn't too complicated. We use Beta acceptance, with a little twist - if Responder has a wide range at the time Beta is used (5+ HCP), the first step is 1-2 controls (assuming you haven't promised 4+) or a min in HCP.

This was added pretty quickly, after a board went 1♣-1NT (8-13 bal) ; 2♣ - 2♥ (4 spades, not 4 hearts) ; 2♠ (Beta acc.) - 3♦ (4 controls)... We ended up 6♠ down 2. I had something like AT9x JTx xx ATxx, we didn't have enough HCP and there were 2 bad splits on top of that. Since then, it's worked pretty well.

Over the Super Unusual Positives, we modified it so that 1st step = 4 controls (0 or 2 Aces), 2nd step shows 4 (1 Ace), and then 5, etc.

[Zelandakh]

Free, out of interest, when you did your analysis did you remember to include relay breaks for Queen asks in the AK method (eg: 4/5NT)? These are extremely important. I also think it is useful to have a second slam exploration method with AK controls, either RKCB or AKQ points, as sometimes the Q ask itself is not enough. When I checked this, the combination of AK + RKCB + Q asks was better than AKQ points alone. It may well be that some combination of AKQ points plus assigning the unused relay breaks to the correct thing is better - but I have not examined the possibilities exhaustively.

As for the OP question, AKQ points are good. They are certainly optimal for very strong hands such as would open 2♣ in Standard. On the other hand, I have (personally) not found this to be the case for weaker hands. In your current method I would ask what you use NT bids for. I strongly recommend that a relay break to the lowest NT bid be used as a non-forcing queen ask when using AK points. That is, partner bids the lowest queen that they do not hold and you now bid the trump suit. If partner has the trump queen they bid on, if not they pass. You can then use a jump in NT as a forcing queen ask, committing the partnership to 6NT while investigating side queens. It is tricks like this which keeps the AK method competitive with the AKQ method.

Naturally there are still some problem hands. In general, if side queens are important then AKQ points will tend to have a small edge. Where kings and the trump queen are sufficient then AK points will sometimes have an edge, especially in more extreme distributions that finish shape relays at a higher level. This goes back to the observation about strength - a very strong Opener tends to have lots of aces and kings and is interested in hearing about queens; a weaker Opener is more often concerned primarily with partner's aces and kings to fill in the gaps. AKQ points also have an issue in often being quite expensive on the first round. The more optimal methods carry a lot of information leaving you well placed if you are not already too high; unfortunately in a bind if you are. AK points seem to work better in finding out if you have a slam, especially a minor slam, but are much less consistent in determining which slam is best, even when employing the NT relay breaks.

One final point. Several posters have already mentioned using 4♦ as a Terminator. This was one of the very first things I removed from consideration in my system. It cannot be right to give the opponents a free lead-directing double on every hand! Instead I use the 4m bids more cleverly to achieve the same effect:

4♣ is a puppet to 4♦, then
4♥ = RKCB for diamonds
4♠ = RKCB for hearts
4NT = RKCB for spades
5♣ = 6KCB for the longest 2 suits, and so on

4♦ instead of 4♣ is now RKCB for clubs.
and
3NT, 4M, 4NT and 5m are now natural.

If the last relay response was 3♠ or 3NT then 4♥ is RKCB for clubs and 4♦ has to carry the additional load of handling a heart sign off. It is still better than the classic 4-way KCB method though (IMO anyway).

Naturally, if you were willing to give up the calls below 3NT for slam purposes, you could adapt this scheme to that. For example, if the last call was 3♦ you might play:

3♥ = relay continuation (AKQ points or whatever)
3♠ = puppet to 3NT, then
4♣ = RKCB for diamonds
4♦ = RKCB for hearts
4♥ = RKCB for spades
4♠ = 6KCB for the longest 2 suits, and so on

with 4♣ as RKCB for clubs
and 3NT, 4M, 4NT, 5m natural.

This scheme would free up 4♦ for any other purpose that might be wanted. Personally, if I were going to give up 3♠ for slam purposes here I would proably choose to use 3♥ as an AKQ ask and 3♠ as an AK ask, with 4m covering RKCB as above and 4M/4NT/5m natural. This gives you the best of all worlds on slam hands at the cost of hurting your game exploration occasionally.

At the end of the day you have to determine what you want your system to be optimised for. Game bidding; game versus slam; or choice of slam. This choice leads naturally towards certain combinations.

A final note. I have written all of this under the assumption that the system in use involves shape relays (eg symmetric relay) followed by control asks. If the trump suit is already known (which the use of Turbo suggests) then the above is gibberish. Could you perhaps clarify how the rest of your system works as I think everyone has made this same assumption and the replies will likely be quite different if you are talking about using AKQ points as a follow-up to Alphas.

[rbforster]

 Zelandakh, on 2012-July-30, 01:35, said:

One final point. Several posters have already mentioned using 4♦ as a Terminator. This was one of the very first things I removed from consideration in my system. It cannot be right to give the opponents a free lead-directing double on every hand! Instead I use the 4m bids more cleverly to achieve the same effect:

4♣ is a puppet to 4♦, then
4♥ = RKCB for diamonds
4♠ = RKCB for hearts
4NT = RKCB for spades
5♣ = 6KCB for the longest 2 suits, and so on
♦
4♦ instead of 4♣ is now RKCB for clubs.
and
3NT, 4M, 4NT and 5m are now natural.

This scheme would free up 4♦ for any other purpose that might be wanted.

It doesn't hurt as much if they don't know the strain or who will be leading, and besides most relay systems give plenty of artificial bids to double for lead and yet almost everyone agrees that symmetric shape relays are very good and don't cite this as a big drawback.

In addition, giving up the 4♦ terminator in favor of direct bids of 4M or 5m to play has a real cost too - responder with undisclosed extras can show them a lot more economically with 4♠+ instead of bidding 4♥ as requested, while bidding over openers game sign off is guaranteed to put you one level higher.

[Zelandakh]

 rbforster, on 2012-July-30, 18:05, said:

It doesn't hurt as much if they don't know the strain or who will be leading, and besides most relay systems give plenty of artificial bids to double for lead and yet almost everyone agrees that symmetric shape relays are very good and don't cite this as a big drawback.

In addition, giving up the 4♦ terminator in favor of direct bids of 4M or 5m to play has a real cost too - responder with undisclosed extras can show them a lot more economically with 4♠+ instead of bidding 4♥ as requested, while bidding over openers game sign off is guaranteed to put you one level higher.

Absolutely agree with the first paragraph except to say that good relayers also know when to just go directly to game and not give the opps any extra information. There is also the point that it is a lot safer to double a bid of 4♦ for the lead than 2♦.

The point in the second paragraph is, I think, overblown. For me at least, when Responder has extras then they are also Serious. Knowing the trump suit here allows Responder to make an intelligent break, for example by showing key cards where appropriate. Going to the 5 level is no big thing here - I am hoping to go to the 6 or 7 level! If you compare this with the 4♦ Terminator then breaking to 4♠ either runs the risk of being passed, and therefore is not possible with a Serious hand, or commite the partnership to the 5 level anyway. Worse, since the trump suit is unknown it is more difficult to make an intelligent move to slam.

You have to look at both sides here. My (strong) opinion is that the 4m method is generally superior to the 4♦ Terminator method overall. Naturally, within the context of a different system the pro and con levels might be different. That is why it is sometimes difficult to make generalisations. Nonetheless, for most systems I would personally not recommend the 4♦ Terminator + 4-way RKCB approach, despite its popularity.

[Free]

 Zelandakh, on 2012-July-30, 01:35, said:

Free, out of interest, when you did your analysis did you remember to include relay breaks for Queen asks in the AK method (eg: 4/5NT)? These are extremely important. I also think it is useful to have a second slam exploration method with AK controls, either RKCB or AKQ points, as sometimes the Q ask itself is not enough. When I checked this, the combination of AK + RKCB + Q asks was better than AKQ points alone. It may well be that some combination of AKQ points plus assigning the unused relay breaks to the correct thing is better - but I have not examined the possibilities exhaustively.

I used similar methods like after AKQ-ask for the AK-ask.
After AK-ask:
- determine a certain suit order to scan
- skipping a step during the first scan means you have 1/2 controls in that suit (similar to 1/2 top honors after AKQ-ask), bidding it shows 0/3 controls (basically skip with 1 top honor, bid with 0/2)
- skipping a step during the second scan means you have the Q in that suit, bidding the step means you don't have it. (remark: it may be more efficient to switch this around)
- 3rd scan is similar as the 2nd scan, but with Jacks
- etc. (if possible)
I didn't include specific relay breaks for Q ask, as it's included in the scans. KISS Bidding anything else but the relay was signoff.

[Zelandakh]

 Free, on 2012-July-31, 09:43, said:

I didn't include specific relay breaks for Q ask, as it's included in the scans. KISS Bidding anything else but the relay was signoff.

I think this is a problem as I find the queen asks come up regularly using AK controls. Here's a forum hand from a year ago which caused some problems for almost everyone (there are surely better examples around but this is one I remember):-

---

Dealer South, IMPs.

In my sequence, South finds out about the 6412 shape and lack of extras at 3♠ and 3 controls at 4♠. Now relaying further gets you nowhere as you run out of space before you can locate the major suit queens. However, how can a 5♥ bid over 3♠ or a 5♠ bid over 4♠ be natural here? This makes no sense. Instead, using these as different forms of queen asks does make sense. So my sequence was to bid 5♠ over 4♠ as a non-forcing ask for the spade queen. That works here but not when North has the heart queen instead. Distinguishing between 6♥ and 6♠ on this hand is difficult for these methods and this is what made the hand interesting for me.

This hand is almost certainly much easier in an AKQ system - hands where the key is in the queens usually are. Nonetheless, the relay breaks help to even out some of that disadvantage without losing the hands where the AK points do have an advantage. It is not difficult to create consistent rules for these breaks which are at least as simple as those you are using for relay scanning (which are already more complicated than I am using) and being able to get specific information quickly is good, even if you could get the same information by relaying eventually. This is the same argument for playing both RKCB and relay scans in the same system - different slam tools for different hands makes for a more effective system.

[benlessard]

I dont have enough experience with AKQ pts. But I would gladly participate in a comparison studies. My method is spiral scan (jumping right into keycards) and a inferential control ask rather than cue bids/denial cue. I also dont show a specific amount of pts by both sides so the side who know about partner pts must sometimes lie on keycards to induce a sign off.

In Chapi
http://bridge.downag...=1CP1HP1NP2DP2H

1C---1H (15+ -- Spades W or GF)
1NT--2D (denies 3S 15-20-- Hearts W or GF)
2H---3H (forced--6412)
4C---4D (H RKC--1/4)
4H---??? (signoff--???)

here responder can bid ---4S im slammish but I dont have Q of H wich will lead to an easy 6S or pass. He knows that declarer has 4H and not too much waste in D and not a bad minimum. I think he should bid 4S despite significant chances of lacking 2 keycards.



If north open

1S--1nt (nat---art relay)
2D--2NT (5S+4H --- ART GF)
3Nt--4D (6412 -- H keyc)
4H--4S (1/4--Q of H?)
4Nt--5C (no--K of S?)
5H--5NT (yes but no K of clubs--PAS)
6S
here you are gambling that partner doesnt have a stiff K or Q of D rather than the Q or J of S.

Tough but I think 6S should be bid.

[Free]

 Zelandakh, on 2012-August-01, 02:17, said:

I think this is a problem as I find the queen asks come up regularly using AK controls. Here's a forum hand from a year ago which caused some problems for almost everyone (there are surely better examples around but this is one I remember):-

---

Dealer South, IMPs.

In my sequence, South finds out about the 6412 shape and lack of extras at 3♠ and 3 controls at 4♠. Now relaying further gets you nowhere as you run out of space before you can locate the major suit queens.

As always, the relay system should be efficient. 3♠ is quite nice to show a 6421, however, you need to bid 4♣ (2 steps instead of 1) to relay further. Whenever you can show shape below 3♠ you win twice: you're lower to show exact shape and you can continue relays with the first step.

In MOSCITO we ended at 3♥ to show a 6421. This saves 2 valuable steps, which makes a huge difference on some hands.
On a side note, this hand is also awful for RKC ♥ when you ask for the ♥Q. So I guess this is a hand you could definitely classify under "AKQ better". I think it's quite safe to say that whenever Queens are important, AKQ is better. When Queens are not important then AK just does the job.

[PrecisionL]

♠: KQJT92 ♥: JT72 ♦: 5 ♣: A3

♠: A5 ♥: AK63 ♦: J63 ♣: KQJ5

Newly Revised Copious Canape Club bidding of this 6-4-1-2 hand:

1♣ (16+) - 3♥ (Semi-solid 6-cd transfer or better) - 3♠ (asking for controls outside suit) - 4♣ (has-to-be-the ace) - 4♦ (old Precision CAB) - 5♣ (4 steps = 1) - 6♠ - - -

If dummy has 1 or 2 or 3 or 4 clubs, there are NO ♥ losers!

(Note for semi-solid and better suits we give up finding exact distribution.)

[awm]

 PrecisionL, on 2012-August-02, 21:10, said:

♠: KQJT92 ♥: JT72 ♦: 5 ♣: A3

♠: A5 ♥: AK63 ♦: J63 ♣: KQJ5

Newly Revised Copious Canape Club bidding of this 6-4-1-2 hand:

1♣ (16+) - 3♥ (Semi-solid 6-cd transfer or better) - 3♠ (asking for controls outside suit) - 4♣ (has-to-be-the ace) - 4♦ (old Precision CAB) - 5♣ (4 steps = 1) - 6♠ - - -

If dummy has 1 or 2 or 3 or 4 clubs, there are NO ♥ losers!

(Note for semi-solid and better suits we give up finding exact distribution.)

Maybe a stupid question, but how did opener know that the semi-solid suit wasn't diamonds?

[PrecisionL]

 awm, on 2012-August-02, 22:31, said:

Maybe a stupid question, but how did opener know that the semi-solid suit wasn't diamonds?

1♣ (16+) - 3♥ (Semi-solid 6-cd transfer or better)

Sorry, I posted after driving home from a Regional last night.

Bids @ the 3-level are TRANSFERS into the next higher suit, so the above auction shows Spades.

[benlessard]

Treating the north hand as a singlesuiter is IMO just better, partner would need a S void and 4H for a H contract to be significantly better than a S contract. In Chapi

1C--2H (15+--6S W or GF)
2S--3D (15-20 -- stiff D)
3H--4D (S RKC--2+Q no K of H & Slammish)
4Nt--5C (bypassed K of club and ask Q of H ?--no)
5H--?? (bypassed Q of clubs & asking for extras)--??

Here the obvious extra is the J of S. Even if responder doesnt have the J of S 6S can be bid.
This would be the worse hand and responder shouldnt treat it as slammish
KQxxxx
xxx
A
xxx

[Free]

 benlessard, on 2012-August-03, 07:38, said:

Treating the north hand as a singlesuiter is IMO just better, partner would need a S void and 4H for a H contract to be significantly better than a S contract.

Consider these 2 hands:
♠KQJT92 ♥JT72 ♦5 ♣A3
♠A5 ♥AKQ9 ♦A63 ♣Q654
If we ask you how you bid these 2 hands, I'm pretty sure you'll claim that the North hand will be better treated as a 2-suiter and end up in 7♥. Nice DD system you have there!

[PrecisionL]

 Free, on 2012-August-05, 05:11, said:

Consider these 2 hands:
♠KQJT92 ♥JT72 ♦5 ♣A3
♠A5 ♥AKQ9 ♦A63 ♣Q654
If we ask you how you bid these 2 hands, I'm pretty sure you'll claim that the North hand will be better treated as a 2-suiter and end up in 7♥. Nice DD system you have there!

Yep, there is always an exception and this is one.

[Rethinking my system with a side Jxxx suit - it could be a minor instead of hearts.]

[benlessard]

Quote

If we ask you how you bid these 2 hands, I'm pretty sure you'll claim that the North hand will be better treated as a 2-suiter and end up in 7♥. Nice DD system you have there!

Well I agree with you that there is difference between practice and IRL I was really in "my system" state of mind rather than making an all system comments.

In chapi the transfer are weak (0-bad 5) or GF, opener will make normal NF bids with 15-20, so i can only relay economically if opener is in the 15-20 range meaning that looking for 7 is far from being a priority holding only 11 pts. Even game might go down so aiming at 4S rather than playing 3Nt/4H when opener got a S in S is pertinant IMO. Rightsiding is a factor too, 1C--2H(spades) will rightside 98% all the spades contracts while 1C--1H(spades) opener need 3 to bid 1S so we might easily wrong side both of the M contracts.

Its not an easy question to guess if the S tricks for a clubs discard in a Hearts contract is worth more than being able to avoid a slow loser in H with a club/diamond discards. Holding JT in H the only H loser is a slow Q and need 2 discards to be avoided.

Also where you are coming out is important too. With a singlesuiter with S diamonds my chance to get out in 3D are excellent. However if I bid 1H (showing 4/5S then opener might bid 2C (natural) and my hand wont be relayed. If my hand is relayed it will come out at 3H.

Put the hand a little bit stronger and I would show both M for sure. Im going to ask my partner what he think of the north hand.

[dake50]

Instead of 4-3-2 AKQ, use 5-3-1 for AKQ.
Only Jacks in "our" suits are counted.