[straube]

Looking for feedback on relay continuations for relay auctions.

To start, I'm still big on the idea of using the second step to suggest interest in signing off. This would replace 4D terminator puppet.

Say the auction goes 1C-3D showing 3415 specific...

3H-would ask controls or relay points
3S-would ask for a 3N bid unless responder had a super-accept (additional controls or relay points of a preset number).

One obvious advantage is that if responder can show a super-accept early, then there is more room for slam investigation.

One could even use this in misfit situations.

For example, 1C-2S showing 5/5 minors and opener is looking at a minimum with 5/5 in the majors.

2N-would ask controls or relay points
3C-would ask for a 3D bid unless responder had a super-accept...in which case he would revert to finishing his pattern and then show controls or relay points.

In the cases where pattern is being shown at the level of 3S or 3N, I like...

4C-asking controls or relay points
4D-terminator puppet
4M-optional RKC (pass with min, show RKC responses with max)
5m-slam invite in the minor

So two main questions. The first is how best to use that available space when pattern is shown by 3H.

The second is, assuming we are in a GF auction, what should bids other than step 1 or step 2 show? For example...

1C-3C showing 5422

3D-asks controls or relay points
3H-asks if responder has a super-accept or not
3S-?
4C-?

Some things I would like to have available include RKC and ways for opener to show shortness and especially voids.

I know most folks use 4D terminator puppet, so I'm interested in other relay breaks used with that as well as what I might use with a second step ask instead of the 4D terminator puppet.

[Free]

I think there's only an advantage if you're below 3♥, because then you can use the second step for signoff in 3NT or a suit (or even have extra options for 4m). Whenever you're higher, you need 3NT as a signoff, so it's better to use 4♦ as terminator instead of 4♣ imo.

[Zelandakh]

Assuming 3♦ is unlimited, I think the first step should ask for extras with zoom to controls and the second step should check for a stopper in the fragment suit (here spades). My rules here give:-

1♣ - 3♦
==
3♥ = relay
... - 3♠ = min
... - ... - 3NT = SO
... - ... - 4♣ = relay
... - 3NT+ = extras + controls according to std rules
3NT, 4M, 5m = to play
4♣ = puppet to 4♦, usually RKCB for ♦, ♥ or ♠.
4♦ = RKCB for clubs

You can easily adjust that for 4♦ terminator.

At the end of the day it is a matter of philosophy. If you cannot investigate the spade position you will be at a major disadvantage against natural on hands where this is an important factor. Against that your advantage on slam hands is even larger. The other main school uses the first step for AKQ points and the second for AK points. Or you can combine the two by using 3♠ as a stopper ask and 4♣ for the AK points.

As for your 1♣ - 2♠ case, why would you not use 2NT as a relay here? Does Opener not want to know the major suit distributions rather often here? Responder could still be (30)55 for one thing. For example, after 2NT: 3♣ = 0355/1255; 3♦ = 2155; 3♥ = 3055 and now you can still find out about extras below 3NT. This is an area where I suspect one of the "standard" forms is not such a bad idea.

Finally, the 5422 is the same for me as 3415 except now there are 2 fragments, so 3♥ is a club ask and 3♠ a diamond ask.

[straube]

Free, on 2015-May-24, 01:28, said:

I think there's only an advantage if you're below 3♥, because then you can use the second step for signoff in 3NT or a suit (or even have extra options for 4m). Whenever you're higher, you need 3NT as a signoff, so it's better to use 4♦ as terminator instead of 4♣ imo.

I wasn't thinking of using 3N as an ask; sometimes the second step would be 4C. Yes, I think being able to suggest a 4m contract is a plus, if only a small one.

[straube]

Zelandakh, on 2015-May-24, 08:34, said:

At the end of the day it is a matter of philosophy. If you cannot investigate the spade position you will be at a major disadvantage against natural on hands where this is an important factor. Against that your advantage on slam hands is even larger. The other main school uses the first step for AKQ points and the second for AK points. Or you can combine the two by using 3♠ as a stopper ask and 4♣ for the AK points.

As for your 1♣ - 2♠ case, why would you not use 2NT as a relay here? Does Opener not want to know the major suit distributions rather often here? Responder could still be (30)55 for one thing. For example, after 2NT: 3♣ = 0355/1255; 3♦ = 2155; 3♥ = 3055 and now you can still find out about extras below 3NT. This is an area where I suspect one of the "standard" forms is not such a bad idea.

Finally, the 5422 is the same for me as 3415 except now there are 2 fragments, so 3♥ is a club ask and 3♠ a diamond ask.

I've just never taken to stopper inquiries. It seems like a lot of rules and brain strain for only occasional advantage. I've done ok by guessing and sometimes the opponents get off to the wrong lead. I'd be interested to hear how many folks really like this.

Hadn't thought about AK points. Or just aces? Aces are simpler in not requiring more rules. I could see...

S1-regular
S2-terminator
S3-aces
S4-RKC longest
S5-RKC second longest
etc

As for 1C-2S (5/5 minors), that may not be the best example. What if (however) we're at +1 and at the point of 2N partner is showing a 5/5 in a misfit situation? Now we've got more potential for going overboard if I make the normal ask. Or say partner shows a reverser with high short at the point of 2N and I'm not interested in the lower ranking suit. Now I ask with 3C and partner has some 47 and we are too high. Even if these are bad examples, I know I've been in uncomfortable spots where we risked getting too high. OTOH, sometimes partner has a plus hand and then we want to know.

[straube]

If responder is at 3S or 3N.....

4C-ask
4D-terminator puppet
4M-optional RKC ask
5m-slam invite

If responder is at 3H or lower....

3N-to play
4C-asks 4D
.....4D-
..........4M-optional RKC ask
..........5C-slam invite
S1-ask
S2-terminator puppet
S3-RKC longest
S4-RKC next longest
etc

So for example....

.....3D-4315
3H-control ask
3S-terminator puppet
3N-to play
4C-puppets 4D for slam invite
4D-RKC clubs
4H-RKC spades

or
.....3H-1633
3S-control ask
3N-to play
4C-puppets 4D for slam invite
4D-terminator puppet
4H-RKC hearts
4S-RKC diamonds
etc

So the terminator puppet would be weaker than the slam invite puppet and responder would accept with some hands that could not break the terminator puppet.

[benlessard]

I feel that with a 5521 or 6421 at 3S/3NT its simply better to us forget about the terminator puppet (that ive played for a long time).

5521.

4C=H keyc
4D=S keyc
4M to play
4NT D rkc
5C last train D.

this turn out so much better than 2 or 3 scheme where 4D is a signoff relay that ive tried even if the range of the 5521 hand is quite wide.

with 5152 or hand with a stiff or singleton in a majors its even better.

Im assuming that the opposite hand had a chance to break a relay with a good suit of its own and that only 2 or 3 trumps suit is possible.

[straube]

So 5152

4C-RKC D
4D-RKC S
4H-RKC C
4S-to play
4N-last train
5C-to play?

You break relay with 6-cd suits I assume. What are your continuations after opener shows an independent suit?

What's your scheme if pattern resolves before 3S?

[straube]

Resurrecting this post of Adam's. I'd like to have these shortness-showing bids.

awm, on 2013-May-01, 19:36, said:

There are several things you are missing about our actual methods. In particular:

1. When relayer shows shortness, we don't immediately jump to QP. We continue to show shape. So in the auction you gave, the sequence would go:

1♣ - 2♦ (strong, hearts and clubs)
2♥ - 2NT (GF relay, high shortage)
3♦ - 3♥ (short clubs, 1435)
3♠ - 4♠ (ask QP excluding ♣KQ, five)

2. Due to the above, relayer has many chances to sign off rather than ask for RP. The normal signoff sequences of 3NT (to play) and 4♦ (puppet to 4♥) are still available after pattern is shown. These are now conditional signoffs, such that describer is supposed to pull with weakness in the short suit (in the case of 3NT) or with maximum QP outside the shortage (in the case of 4♦). In effect opener's choice to ask for QP via 3♠ in the auction above (rather than bid 4♦ to start a signoff) is an indication that he is willing to play slam opposite at least some hands with 4 QP outside clubs (else he would bid 4♦ and describer would bid past 4♥ with 5-6 QP outside clubs). This means it's not a big deal to be "really high" here because we are virtually guaranteed to be slam-going.

3. In fact relayer rarely has to ask for the exact location of honors in these sequences. With the short suit already excluded usually a count of QP is enough to make a slam decision.

4. We don't exclude the ace of the short suit from the QP ask. However, opener can skip a step instead of asking for QP such as:

1♣ - 2♦ (strong, hearts and clubs)
2♥ - 2NT (GF relay, high shortage)
3♦ - 3♥ (short clubs, 1435)
4♣? (void clubs ask controls outside clubs)

This void-showing asking bid takes the place of our usual relay break (which asks controls outside some particular suit).

5. We use relay breaks after pattern has been shown but below 3NT as stopper asks. So 1♣-2♦-2♥-2NT-3♣-3♦-3♠ asks for diamond help.

6. The original idea behind these asks was to have opener show shortness in the "suit just shown" such that breaking relays at different points would show different shortage. So for example you could have 1♣-1♥ (spades)-1NT (short spades) or 1♣-1♥ (spades)-1♠ (relay)-2♣ (diamonds)-2♥ (short diamonds). However, this only works in a system with GF relay responses; for us we need the first relay break to show a minimum opener so we rarely have more than one chance to use the shortness showing relay break. With that said, I have found that this shortness showing break is surprisingly rarely needed (i.e. usually I reach the same contract without it).

[Zelandakh]

The way to have shortage relay breaks is to arrange the responses so as to avoid first round jumps to the 3 level with "normal" hands. With standard symmetric you can encompass all of the hands with the responses 1♥-2♠ if you want to.

At the 3♠/3NT level, Ben's scheme gains 2 steps for RKCB sequences at the cost of losing relay continuations. That seems to me to be a bad deal. My scheme is:

4♣ = relay
4♦ = marionette to 4♥ (Responder can break this with extras) then 4♠/4NT/5♣ = RKCB for ♦♥♠
4♥ = RKCB for clubs
4♠, 5m = to play

...but obviously you can rearrange the suits if one or more are out of scope. This is functionally equivalent to a 4♦ terminator with 4♥+ being 4SKCB.

My view is simple - giving up the relay continuations is also giving up your advantage over natural. You have better shape resolution but in turn have given up cue bidding before key carding, which is a big deal. Whatever else you include, I think you should keep 4♣ as a relay. Those 2 steps are just not important enough to give up your main weapon.

[straube]

I agree about retaining the first step for normal relay continuations. It apparently works for Ben to use RKC instead, but he breaks relay when opener has an independent suit and since we don't we're really talking about two very different situations.

I'm actually very happy with queen points and parity cue bidding and especially so when we have a major suit fit. One thing that I like of Zelandakh's structure is using the RKC order for C,D, H, and S because I probably would only choose RKC when I'm deciding between 5m and 6m. PCB often takes until the mid 5-level to resolve partner's cards, but then again knowing the total queen points is often enough.

I'm still interested in showing shortness. Adam pointed out to me that we might try (in a GF auction) skipping one step for shortness in responder's longest suit but waiting and skipping one step in responder's next longest suit. I.e. this would be instead of skipping one step for the longest and two steps for the second longest. The advantage here would be that in the case of shortness for the second longest, we would only be costing one step instead of two.

I'm also wondering if instead of showing shortness and then skipping a step later to show a void, one might show shortness and then ask responder to later give QPs counting an ace in the short suit as 1 QP instead of 3 and not counting the K or Q in that suit at all. The value of that ace is probably just shy of 2 QPs opposite a singleton and just north of 0 QPs opposite a void and so we would average it 1. Opener then makes his decision whether to start PCB and this suit is examined first. Odd shows the ace and even denies. After examining the other suits, it would be as simple a matter as other schemes to locate other high honors.

[benlessard]

Quote

but in turn have given up cue bidding before key carding, which is a big deal.

Why ? when you have all the keycard or you missing 2 keycards the cuebidding mostly help the opponents rather than yourself. Its only when your missing exactly one keycard that cuebidding/zooming can dominate scanning methods. When you scan in some case there is only one suit that you can be worried for a control so you can also ask for a control rather than cuebid stuff that youll know after RKC anyway.

Note that its one of the reasons why im working into a 6keyc scheme to know if im lacking the K of trumps or an ace is quite useful to scan without being worried about a suit wide open.

I dont have a lot of experienced with QP but for me a Q in a unknown suit is not that useful.

[Zelandakh]

We have discussed this idea before and it seems reasonable if we cna agree the trump suit at the 3 level. Then we can play the first step as a slam try (Frivolous) and the second as, effectively, RKCG. By having the follow-ups work as control asks the end result is parity. But in a full relay system the situation is different. Most of the time we will have more distibutional information but have to make more compromises in the follow-up structure due to lack of space. Because of the cost of agreeing a suit, it works out better on average to use a control point scheme, either AK or AKQ points. At least that was my experience when I tried testing some ideas a little while back.

[benlessard]

For non fit hand or NT hands I agree that showing general values of the hand with a AKQ scheme is sensible but usually when your end in 3S-3NT you have a lot of shape.

Ex I hold a 6421-5521 is the Q of D or a stiff king of club really useful cards for slam ? I would just prefer to not count these cards and know how many keycards my partner have rather than his overall strenght or his QP.

As i see it whats important

1- fixing trumps suit
2- keycards + Q of trumps
3- overall strenght ask/QP
4- cuebids/control ask.
5- specific card ask.

Your goal is mostly to avoid duplication (giving the same informations twice) or giving useless information (information that will be used against you especially).

I feel that 1-2 first is always useful and rarely redundant. While cuebidding or strenght ask followed by keycarding later is often redundant or pointless. When you learn you have all the keycards the control showing bids information is worthless. When your missing 2 keycards strenght ask is only useful for the opps.

Even if I know my partner got the Ks (we have all the keycards and he cue in a suit where I had the A and i know hes not a stiff) when im going to ask for side K he will still show me this king. If i ask for a specific king I can skip this card that i know by inference but i wont save a step anyway.


ex (HSCD is the order)

4C= ask for KS
4D= ask for KC

4C--??
4D no ks
4H Ks no Kc
4S Ks+Kc no Kd
etc

I know by inference that my partner got the Ks.
So he will respond 4H without the Kc and 4S with KC no Kd

If I skip the Ks ask.

4D= ask for KC --- ??

4H no Kc
4S Kc no Kd

So knowing by inference that my partner had the Ks was worthless. So cuebidding followed by rkc was redundant.

Another cases is when your lacking 2 keycards,knowing if my partner is maximum or minimum is also pointless and information that will be only useful for the opps.

I have no problems with strenght ask but why not as a last train just after the keycards ?

Even with jacoby 2NT (that I hate) responding keycards or showing voids rather than showing min/max.singleton is sensible because the keycard information is more likely to be useful than overall strenght because we are in a suit contract.

Also I feel that hands without fit or NT hand you can just wing it and use 4NT quant by either players and it will work well. I dont remember last time I went down in 4NT while 3NT would have made (made a try for 6nt, unlikely 4nt go down). In trumps however going down in 5M can happen (made a try for 6M, 5M can go down if your a bit unlucky.)

I dont exactly know how to explain it but in NT the number of tricks is less random while suit contracts the number of tricks available got more variability.