[straube]

Some of us have looked at this before. In the context of strong club...

1C-16+
.....1D-5-7 hcps
.....1H-GF, various
.....1S-DN
.....etc-GF various

After 1C-1D,
.....1H is natural or artificial GF
..........1S-bal, C/D, C, D, 3-suited short M
...............1N-GF relay
....................2C-bal
....................2D-C/D
....................2H-C
....................2S-3-suited short M
....................etc-D
...............other is natural and confirms hearts
..........1N-S, S/C, S /D
...............2C-GF relay
....................2D-S/C
....................2H-S
....................etc-S/D
...............other is natural and confirms hearts
..........2C-S/H or 3-suited short minor
..........2D-H/C
..........2H-H
..........etc-H/D

Some of the tradeoffs here...we are +1 on semipositive relays but are also usually +1
On GF relays. We seldom have the opportunity for reverse relays.

[dake50]

Putting a DN on the first page of responses uses a bid
for a hand case that is 0.7% freequent and almost surely
has no development bids. So a whole information branch is
severred with 1S=DN.
Seems a more frequent use for 1S is systemically logical.
How much fear do you have that a second round (even, as
nightmare, a 3rd round DN) will overheat your exploring auctions?
Again only a fraction of a percent is DN.

Edit from my Frost charts: 0:0.84 +1:1.92 +2:2.56 +3(no K):2.72 +4(no A, no K):2.60
= 10.64% is the maximum DN% when opener has exactly 16.
I actually use 1C-1D(0-7); 1H demands a 1S DN, else show something more.
My point is that this demand to show DN, is an asked 2nd response,
Rather than 10.64% of first responses, 10.64 of the 46.01 in 0-7.
How much does that take from the development of our auction?
Esp compared to a 1st-round bid reserved for DN, thus losing all the follow-ups
(2nd page,3rd page as I wont to say) to 1S defined another way?

[straube]

dake50, on 2012-November-24, 02:14, said:

Putting a DN on the first page of responses uses a bid
for a hand case that is 0.7% freequent and almost surely
has no development bids. So a whole information branch is
severred with 1S=DN.
Seems a more frequent use for 1S is systemically logical.
How much fear do you have that a second round (even, as
nightmare, a 3rd round DN) will overheat your exploring auctions?
Again only a fraction of a percent is DN.

A DN (0-4 hcps) is about 20% of responses. I've sampled them with the BBO generator and someone once ran a simulation. Even if they were .7%, you have to assign bidding space for them.

This scheme gives more space to semipositives at the expense of positives. DN has roughly the same, excepting that if I compare to 1C-1D, 1H-1S where 1H shows a bigger hand, then the DN in this latter sequence has information that 1C-1S lacks.

I think a lot of people would say that 1C-1D as 0-7 continuations are pretty awkward as 1) we haven't established that the partnership has more than 16 hcps and 2) we've exchanged no suit information. I know there are plenty of schemes to sort this out and some do better than others, but I think there's still that consensus that it is difficult.

[glen]

straube, on 2012-November-24, 09:15, said:

A DN (0-4 hcps) is about 20% of responses ...

.7% was strange, even if one forgets that opener has 16+, it would be closer to 7%, not 00.7% (I blame the Heineken).

[rbforster]

straube, on 2012-November-23, 15:39, said:

Some of the tradeoffs here...we are +1 on semipositive relays but are also usually +1
On GF relays.

I think +0 GF and +2 SP is clearly better than +1 each, given you're set on relaying both whenever you have sufficient values. this is because there is a lot more need for space to describe the values range of GF responses (since they are unlimited), while for SP hands their strength range is already quite narrow.

[hrothgar]

rbforster, on 2012-November-24, 11:39, said:

I think +0 GF and +2 SP is clearly better than +1 each, given you're set on relaying both whenever you have sufficient values. this is because there is a lot more need for space to describe the values range of GF responses (since they are unlimited), while for SP hands their strength range is already quite narrow.

It is for exactly this reason that you aren't relaying the semi positive hands nearly as often.

In an idea world the semi positive allows us to quickly bash to game or break relays and transition to natural bidding to intelligently explore 3NT.

Relays are great for slams, but not so much for exploring game.

[straube]

Nice to hear from you Rob. Seldom see you online any more. I agree I'd rather have 0 and 2. I'm actually pretty much like our 1H semipositive response, but I'm always on the lookout for better. It's a little awkward that we have to rebid 1N with off-shapes like 1453 or whatever. Getting to the right game is almost never a problem, but we can languish in a poor part score on occasion. Like if opener has 1543, he'll transfer to 2H (1C-1H, 2D) and responder with a minimum 4045 has little choice but to accept. I mean, these things never seem to come up, but they could.

[rbforster]

straube, on 2012-November-24, 11:48, said:

Nice to hear from you Rob. Seldom see you online any more. I agree I'd rather have 0 and 2. I'm actually pretty much like our 1H semipositive response, but I'm always on the lookout for better. It's a little awkward that we have to rebid 1N with off-shapes like 1453 or whatever. Getting to the right game is almost never a problem, but we can languish in a poor part score on occasion. Like if opener has 1543, he'll transfer to 2H (1C-1H, 2D) and responder with a minimum 4045 has little choice but to accept. I mean, these things never seem to come up, but they could.

thanks for the welcome, sorry I've been too busy with stocks and not so much with cards lately. There was another reason for 0/2 vs 1/1 that I remembered. Again assuming you will relay both GF and SP hands when you have GF total values, SP relays are less frequent since the probabilities favor more even distributions of HCPs. So having 20 vs 6 is less likely than 16 vs 10, and so forth, even after you condition on the strong club opener. this means that you'll get more value out of GF relays than SP ones on a frequency basis, which favors lower resolution for those GF ones.

You may recall that my relay system took this one step farther and didn't relay SP hands at all, in order to get -1 on GF but worse for SPs.

[PrecisionL]

straube, on 2012-November-24, 09:15, said:

A DN (0-4 hcps) is about 20% of responses. I've sampled them with the BBO generator and someone once ran a simulation. Even if they were .7%, you have to assign bidding space for them.

I agree, 0-4 hcp = 20.5 % (using bridge odds complete by Frost, Kibler, Telfer, & Traub) in an Excel Spread Sheet.

Also, 5-7 = 32 %, 8+ = 47.5 %, 8-11 = 35 % and 12+ = 12.5 %.

The sequence 1♣ - 1♦ = 0-7 hcp is an important part of half of the responses to a strong 1♣ and the space is needed to find a good partial contract or even a game invitation (especially in Match Point Pairs competition).

Keylime and I do NOT relay after a 1♦ response (except 1♥ has a 20+ hcp balanced option), but do have a follow-up scheme similar to Cohen-Berkowitz Precision, but we have condensed mini-splinters and added fit jumps by responder to show 3 kinds of fit for opener's 1-level major suit rebid.

[Zelandakh]

rbforster, on 2012-November-24, 13:34, said:

You may recall that my relay system took this one step farther and didn't relay SP hands at all, in order to get -1 on GF but worse for SPs.

Am interested what you did to accomodate this. I do not relay SPs either but GF hands are at +0.

I have found that relays are only bad for game bidding if you tailor your relay breaks to slams. If you use the relay breaks to improve the game bidding then they are at least as good as natural bidding in most auctions. This is actually one of their attractions for me, and my system is fairly heavily geared towards game bidding, trusting that the advantages of relays will be enough to keep a small edge on slam hands.

[rbforster]

Zelandakh, on 2012-November-26, 03:39, said:

Am interested what you did to accomodate this. I do not relay SPs either but GF hands are at +0.

there was some shuffling of the 3-suiters, but basically I put the GF H/H+m hands into 1C-1D along with the 0-7 hands. There were more artificial responses after that, but it was designed so that when responder had the GF, opener would usually bid 1H (catch all but no extras, and completing the transfer). Then 1N+ over 1H relayed the GF shapes at -1, while 1S confirmed the 0-7 hands and natural nonforcing bidding followed since opener was limited. Having removed those GFs, there was space for everything else over 1C to be GF at -1, and nice transfer symmetric relays.

[Zelandakh]

OK, so you effectively have a (delayed) 1♠ response as any non-GF. I imagine that is difficult to handle and is unfortunately not something I can use since my 1♣ can be a flat 15. I thought you were saying that you could relay GF hands at -1 and still accomodate SPs. That would be something of a Holy Grail for me!

[rbforster]

Zelandakh, on 2012-November-27, 08:05, said:

OK, so you effectively have a (delayed) 1♠ response as any non-GF. I imagine that is difficult to handle and is unfortunately not something I can use since my 1♣ can be a flat 15. I thought you were saying that you could relay GF hands at -1 and still accomodate SPs. That would be something of a Holy Grail for me!

There's no problem with the strong club being 15 balanced, although that's 1 hcp weaker than I play it, you just raise your GF response standard by the same amount. The SP hands are handled by a combination of artificial rebids over 1C-1D and relay breaks by opener. Admittedly they aren't handled as well as they could be when 1D is purely weak, but I think with better methods over 1C-1D they can be handled about as well as in normal precision (not great, but ok).

There's no holy grail unless you start using tempo breaks for extra bidding space - there's only so much space and you try to find the best probability and score-maximizing adjusted methods in that context.

[straube]

Surprised really not to see more interest in the idea behind this thread which is the 1D as 5-7. I mean, on balance, I don't think I like it, but I haven't spent much time with it. For sure, the 5-7 range bidding is vastly improved, but at the cost of GF relays. Most of the GF relays are +1 but a lot are +0 and there's still some potential for reverse relays.

1C-1D,
.....1H-1N (hearts or GF relay, spades)
.....2S fit

1C-1D,
.....2C-2H (natural nf, natural nf)


1C-1D,
.....1S-2C (natural F1, natural nf)

By using 1D for 5-7, responder can often bid fairly naturally after opener's own natural rebid. We use 1H as our semipositive and it definitely cramps opener from showing distribution. He can't even get out in a minor until the 3-level. I'm almost certain Moscito has the same difficulty. OTOH, Precision 1C-1D, 2C is natural but probably a huge range. Here it's..."I have 16, you have 5-7, I have clubs and am content to play 2C" Big difference.


I mean the idea isn't about relaying SPs at +1. The idea is to use 1C-1D, 1H as a two-way bid...natural or GF...and orchestrate responder's answers such that when he holds hearts he bids at 2H or higher...which is news that will be welcome to opener whether he has hearts or a GF relay. Isn't this a little bit new? We looked at it before, but I had the responses after 1C-1D, 1H all wrong.

[straube]

How is this Zelandakh?

Positive responses...

1H-bal, S, S/C, S/D, S/H, S/H/m
.....1S-asks
..........1N-S/H, S/H/m, bal w/out major
...............2C-asks
....................2D-bal w/out major
....................2H-S/H/m
....................etc-S/H
..........2C-bal
..........2D-S/C
..........2H-S
..........etc-S/D
1N-H, H/C, H/D
2C-C/D
2D-C
2H-C/D/M
etc-D

What I like about this is that it preserves room for hands that are more likely to wind up in 3N...namely those with minors or those that are balanced. Hands with majors likely have more room to ask controls on the way to 4M.

After 1C-1H (which handles balanced hands as well as spade hands), opener may reverse relay at +1

[awm]

I'm reluctant to post because I feel like I've already solved this problem, and have yet to see significant negatives to my current approach.

The style you propose here seems a certain loser when responder has 0-4 (because 1♠ double negative is so high) while also sacrificing space to describe positive hands (because you lose the 1♠ bid). The only gains when compared to other systems would be on the semi-positives themselves, but these are so tightly constrained in terms of strength that they rarely give much trouble in an uncontested auction, and the complete lack of shape information (combined with the lack of a forcing pass) means you will not see very significant benefits in competitive situations either.

I've played against a lot of pairs who use an "any 5-7" double in competition and my experience is that the method is absolutely terrible. Obviously the situation here is slightly different, but I think not very much so.

[Zelandakh]

rbforster, on 2012-November-27, 20:17, said:

There's no problem with the strong club being 15 balanced, although that's 1 hcp weaker than I play it, you just raise your GF response standard by the same amount.

The problem is that the weaker your 1♣ opening, the more common negatives and semi-positives become. This means that my (effectively 0-8) 1♦ response is essentially overloaded. I handle hands after it just fine without interference but it is a large range if, say, 4th hand suddenly jumps to 3♠. SP responses are designed to ease this situation and it would be great to include them in my system if I could do so without giving up too much.

straube, on 2012-November-27, 21:35, said:

How is this Zelandakh?

I like this. The 1♥ response is effectively "spades or balanced", which is about as good as you can make it within the confines you have set up, and it fits together logically. I do have one last trick for you though. Instead of 2♣ = minors; and 2♠+ = diamonds you might try 2♣ = 5+ diamonds, 0-4 clubs; 2♠ = 5-5 minors; and 2NT+ = 5+ clubs, 4 diamonds. That improves both the right-siding and the homogeneity slightly.

I will say that I am with Adam in not being a big fan of the 1♠ DN response but we have been over that ground often enough in the past. You are obviously comfortable with it and the best idea is to find an optimal structure, or as close to optimal as we can guess, and then to test it against the alternatives and see. Whether one loses more on the DNs than is gained on the SPs is something I just could not say. It may well depend on the range of the 1♣ opening, which goes back to the answer in paragraph one.

[straube]

The thing about reverse relays (strong hand shows) is that systems like Moscito and Scream allow space for patterns that really should not be shown...specifically the 5422s, 6322s, and 7222s. Of course, either Moscito or Scream would benefit as much as what I'm attempting here...

After 1C-1H (GF spades OR balanced)
.....1S-relays
..........1N-various
................2C-relays
....................2D-bal w/out major
.........................2H-relays
.........................2S-H/S/m
................2D-C/D
................2H-C
................2S-C/D/M
................2N-D
..........2C-bal w/ major
...............2H-C
...............2S-C/D/M
...............2N-D
.....1N-S, S/C, S/D
.....2C-H, H/C
.....2D-H/D
.....2H-H
.....etc-S/H

so +1 here for the 2-suited and 3-suited. However, by omitting the 5422s we are +0 for all of the short-legged low short patterns. For example...

1C-1H, 3D would be 5-4-3-1 but the pattern 5-4-1-3 would be shown at the point of 3H.

We are +0 for the single-suited (6+) patterns, but by omitting the 6332s and 7222s, we are actually -1 for the single-suited compared to standard symmetric.

[straube]

I'm wondering whether we could borrow a little from Imprecision. Here 1D is all semipositives or GF balanced and the similarity is that responder can relay opener's unbalanced minimum. In the event that opener establishes a GF himself....no problem holding the balanced hand...you just show your shape and super-accept if you're a QP or two over GF strength. More differences than similarities by far with Imprecision in that 1) this uses a DN and 2) the semipositives are lumped all together and impart no suit information. Also, no accommodation for opener's 5M332 which have to rebid 1N instead of showing the major. Still, I think it has some nice things. The semipositive auctions seem more natural. Most of the GF auctions are +0. Semipositives are +1. We establish immediate GFs when responder has an unbalanced hand. We establish likely point majority when responder has a semipositive or GF balanced hand.

1C-16+
.....1D-5+ balanced or 5-7 unbalanced
.....1H-GF, S+ at +0
.....1S-DN
.....1N-GF, C, D, C/D, C/D/M all at +1
.....2C-H/C
.....2D-H
.....etc-H/D

After 1C-1D,
.....1H is natural or artificial GF
..........1S-bal, C/D, C, D, 3-suited short M
...............1N-GF relay
....................2C-bal
....................2D-C/D
....................2H-C
....................2S-3-suited short M
....................etc-D
...............2C-H/C or H/C/O
....................2D-relay
...............2D-D/H
....................2H-relay
...............2H-H
..........1N-S, S/C, S/D
...............2C-GF relay
....................2D-S/C
....................2H-S
....................etc-S/D
...............other is natural and confirms hearts
..........2C-S/H or 3-suited short minor
..........2D-H/C
..........2H-H
..........etc-H/D
.....1S-4+ spades
..........1N-forcing
...............2C-S/C or S/C/O
....................2D-relay
...............2D-S/D
....................2H-relay
...............2H-5S/4H
....................2S-relay
...............2S-5S/5H
....................2N-relay
..........other-natural semipositive
.....1N-17-19
.....2C-6C or 4+D/5C
..........2D-relay
...............2H-4D/5C
...............etc-C
.....2D-6D or 5D/4C
...............2H-relay
...................2S-5D/4C
...................etc-D
.....2H-4S/5H
...............2S-relay
.....2S-5D/5C
...............2N-relay

[rbforster]

straube, on 2012-December-03, 22:15, said:

1C-16+
.....1D-5+ balanced or 5-7 unbalanced
.....1H-GF, S+ at +0
.....1S-DN
.....1N-GF, C, D, C/D, C/D/M all at +1
.....2C-H/C
.....2D-H
.....etc-H/D

I might restructure this to have the majors at +1 and the minors at +0, or closer to that anyway for the GF responses. Theres less space below 3N than 4M, and the minor hands are usually headed to 3N and hence could use a little more room. Also, by zooming into minor(s) with 2C+, you may avoid wrong siding NT. it's likely better to have opener, who is both stronger and less described, declaring 3N in these cases.