[straube]

Looking at 1M-2D again. 1M is 5+ and 10-15 while 2M is 7-11 such that opener will either have a plus hand with a 6-cd major or will be 6M/4OM.

1M-2C GF relay or club invite
.....2D-side minors (edit: or 5332)
..........2S-GI, 4+/4+ minors
..........2N-GI, only 4+ clubs
.....2H-OM
..........2N-GI
.....2S-4C
..........3C-GI clubs
.....etc-6M, GF

My general want is to eliminate the GI hands from the 1N semiforcing response. Thinking that if 1M-2D promises 5 diamonds that this can be accomplished for anything except 1S-2D when responder has specifically 2443 (because no 4 clubs there).

I've also remember Adam's point about responder needing to know about a diamond fit, but we were talking then about a 2D response that promised only 4 diamonds (consequently raising with 4) and here I'm talking about responding only with 5 diamonds...so maybe responder can pass with 2 or 3 diamonds and a minimum.

1H-2D (GI 5+ diamonds)
.....2H-minimum, 5 only hearts unless 4S/6H
.....2S-minimum, but has 4 diamonds or 5 clubs
..........2N-relays
...............3C-5 clubs
...............3D-4 diamonds
.....2N-artificial GF, may have 3 diamonds or 6 hearts
.....3C-5 clubs, GF
.....3D-fit, GF
.....3H-6H, nf if misfitting

1S-2D (GI 5+ diamonds)
.....2H-minimum, artificial, could have 4 diamonds or 5 of other suit
..........2S-2-fit
...............2N-5 hearts
...............3C-5 clubs
...............3D-4 diamonds
.....2S-minimum, 4 hearts
.....2N-artificial GF, may have 3 diamonds or 6 hearts
.....3C-5 clubs, GF
.....3D-fit, GF
.....3H-5 hearts, GF
.....3S-6S, nf if misfitting

So this is quite a bit simpler than what I had before. Any suggestions?

[Zelandakh]

What are you doing with balanced and 3-suited invitational hands without 3 card support?

[straube]

Zelandakh, on 2014-February-13, 11:12, said:

What are you doing with balanced and 3-suited invitational hands without 3 card support?

If they have 4 clubs, they are responding 2C. I may have the math wrong but I think the only pattern with 12-13 that couldn't respond to 1S is 2443.

[Zelandakh]

Hmmm, so you have a 1444 invite and respond 2♣. Opener with a bare minimum one-suiter rebids...what? I guess we have similar issues with (13)45 hands and shortage in the major opened. For sure these are icky hands but the rest of the field is stopping in 2M or 2NT so being forced to the 3 level is not going to score well. How did the old structure handle these?

[straube]

Zelandakh, on 2014-February-13, 11:35, said:

Hmmm, so you have a 1444 invite and respond 2♣. Opener with a bare minimum one-suiter rebids...what? I guess we have similar issues with (13)45 hands and shortage in the major opened. For sure these are icky hands but the rest of the field is stopping in 2M or 2NT so being forced to the 3 level is not going to score well. How did the old structure handle these?

With 1444

1S-2C,
.....2D (various including a 5332)
..........2S-4/4+ minors, GI
...............2N-yuck

Before...

1S-1N, P

For sure we lose on this deal (although we can comfortably pass 1N now with our 13 hcp 5332s), but the idea is to separate the heavy from the light invites. I could even sacrifice 1S-3C for that 2443 12-13 and 1N would always be 11 or less. I think this adds clarity to auctions such as...1S-1N, 2H-3H which is more a distributional raise limited to 11 hcps. I know it's good to stay out of 2N, but let's say we had 1S-1N, 2D-? and I have a 1525 10 ct...I might just bid 2N as the least of evils.

1M-2C doesn't let us stop shy of 2N, but 1M-2D lets us frequently stop in 2D or 2M (the invitational strength message having been delivered) and 1S-2H (GI only) also lets us stop in 2H or 2S.

[awm]

The 1M-2D seems good to me. Dunno how efficient your GF auctions will be over 1M-2C though; it seems like you have lost some space for relays...

[straube]

Thank you. Yeah, I've given up showing 5440s, and the 5332s resolve at 3H, 3S, and 3N. The rest are +1.

I still kind of prefer to relay and feel I'm helped by opener being a limited hand (the 5332s between 11 and 13 for example). Occasionally responder has the distributional GF hand and wants to show.

I ran a simulation for 1S and counted only the 1N, 2C, 2D, 2H and 2S responses. They were...

1N-35
2C-28
2D-4
2H-6
2S-29

so rather disappointed with the frequency of 2D and 2H responses. Still interesting.

[akhare]

straube, on 2014-February-13, 10:59, said:

Looking at 1M-2D again. 1M is 5+ and 10-15 while 2M is 7-11 such that opener will either have a plus hand with a 6-cd major or will be 6M/4OM.
So this is quite a bit simpler than what I had before. Any suggestions?

IMO, the flaw in this method is that it's trying to slice things too thin:

• The 2M range as 7-11 is too conservative and while it's OK on some hands, 10 and 11 point hands should strongly consider opening 1M
  
• Balanced 10-counts should be allowed to open 1M especially when NV
  
• The 2♣ relay method is nice, but may reveal too much information about opener's hand when limit is game and can't stop below 2N
  
• The mnemonic ease of such methods is always a cause for concern and it's unclear that the gains are sufficient to outweigh the downsides
  
• We may be poorly placed over a 1M - 2♣ response in case of interference. Note that it's different from a strong♣ interference because opener can be much weaker

That said, removing the invitational hands from 1N is the right approach, but it might be better to use all 2 level responses as GI (a la SAYC) and take it from there.

My idea of the structure would be something along these lines:

1M:

1N: Semi-forcing
2C: GI+, F1
2D: GI+, F1
2♥: GI+ over 1♠, F1
2M: Constructive
2N: Various raises, not necessarily GF
3m: Natural invite with good suit, not enough for 2m (or some other raise if you prefer, but I don't think it's necessary)
3M: Preemptive when NV, mixed raised when vul.

Over 2x bids, opener's rebids can be mostly natural, with 2M / raise to 3m showing minimum hands. This means that opener can use 2N / 3-level bids, etc. to show GF sequences. Similarly, responder still has several options to establish a cheap GF and / or ask for more information.

There's still plenty of time to find out about shortness on the way and while it doesn't give the complete pattern, it should still be very playable.

[straube]

Here's my tally for two structures. I made the 3-levels the same for both...


Passing with fit.........................relay....................natural
....................................................0...........................4

1N 7-13......................................15
1N-7-11..................................................................12

2C-GF or inv 4H or 3-cd LR........29
2C-natural, 12+.......................................................5

2D-hearts...................................7
2D-natural, 12+......................................................11

2H-8-11 raise............................15
2H-natural, 12+......................................................13

2S-0-7 raise...............................18
2S-simple or constructive........................................26

2N-4-cd LR.................................1
2N-all 3+ cd LR.......................................................11

3C-natural invite..........................1
3C-natural invite......................................................1

3D-mixed raise............................3
3D-mixed raise.........................................................3

3H-light invite.............................2
3H-light invite..........................................................2

3S-preempt.................................4
3S-preempt..............................................................7

4H-to play....................................1
4H-to play.................................................................1

4S-to play....................................4
4S-to play.................................................................4

I agree with you Atul about wanting to open 5332 10 cts when NV. I'm having difficulty trying to spend too much effort on the GI (12-13) hands that don't have a fit, don't have four hearts and don't have 6 clubs. They just aren't that frequent and very often opener makes a rebid and an invitation can be made. I've tried 1S-2D nf and 1S-2H nf and the tallies were just very low for both.

It's hard to switch gears and have responder start to describe his hand after opener has already shown 5/13 of is pattern...which is kind of what we're trying to do when a 2/1 is both natural and forcing. Add to this the difficulty of needing to decide what further bids start a GF and which do not.

Separating constructive from less than constructive raises seems really useful to do. If we're NV and I have Qxx xx xxxx xxxx I want to raise spades, but I can't also do this then with Axx xxxx Kxx Kxx. But I don't want to go to the 3-level with the latter hand if opener can have a 5332 ten count. If we're vul and I hold Jxxx xx Jxxxx xx I might prefer to bid 2S than choose between pass and 3S.

[akhare]

straube, on 2014-February-16, 15:25, said:

It's hard to switch gears and have responder start to describe his hand after opener has already shown 5/13 of is pattern...which is kind of what we're trying to do when a 2/1 is both natural and forcing. Add to this the difficulty of needing to decide what further bids start a GF and which do not.

The basic differences between the approaches is strength first or pattern first. Natural forcing responses limit opener's strength (basically establish the (in)ability to accept game opposite a routine 12+ invite). This can be important considering that opener's 10-15 can be 9-15 with compensating shape.

The conjecture here is that sufficient shape information can always be gleaned later if opener can create a GF or if responder has enough to force game opposite opener's denial. The natural approach caters better to the 12 vs. 12 or 13 vs. 12 balanced hands which may get passed out over 1M - 1N (semi-forcing).

Also, since responder can simply establish a cheap GF (or simply ask for more information), there's really no need to switch gears to show responder's hand. For example, after say 1♠ - 2♣ - 2♦ (natural or 5M332), responder's impossible 2♥ simple asks for more, after which opener's 2♠ can show the balanced hand.

Granted, some sequences are more compressed and 1♠ - 2♥ - 2♠ can have a variety of hands that can't insist on game. However, this still allows responder to force further with 3♣ (artificial), get out in 2N with a misfit (if applicable), etc.

[Siegmund]

Nice to see someone is experimenting with what to do after a non-GF 2/1.

I've been playing something similar in one partnership. The gist of it is that 1M-2m is natural inv+, and won't include 4 of the unbid major unless GF. 1M-2m-2M is now passable with six hearts, and 1M-2m-2OM says "If you bid 2NT or 3m or 3M, I might pass", while 1M-2m-2N and 1M-2m-3m remain GF, and the 2OM bid takes care of some stopper-fishing too.

I could squeeze a bit more out of the system especially when two steps rather than one are available below 2NT but went for a bit of simplicity.

[akhare]

straube, on 2014-February-16, 15:25, said:

Separating constructive from less than constructive raises seems really useful to do. If we're NV and I have Qxx xx xxxx xxxx I want to raise spades, but I can't also do this then with Axx xxxx Kxx Kxx. But I don't want to go to the 3-level with the latter hand if opener can have a 5332 ten count. If we're vul and I hold Jxxx xx Jxxxx xx I might prefer to bid 2S than choose between pass and 3S.

IMO, you are overestimating the ability of the weak raise to shut out the opposition. Assuming the constructive range is 7-10, the weak raise will win over a 1N response only when responder has exactly 0-6 points and 3-4 ♠ and when both LHO and RHO can't make a convenient bid.

Granted, there will be a few hands when all conditions are satisfied, but does it happen with sufficient frequency? What about the potential disadvantages of a cheap leading directing X of the artificial raise and the 2M cue bid? Also, the 1M - 2M (weak) occasionally shutouts a big side suit fit, which may might be discovered over the semi-forcing 1N response.

Regarding the 1M - 2x responses, 1M - 2♣ is great for GI+ ♣ hands (or balanced hands that prefer to ask), but 1M - 2♦ and 1♠ - 2♥ as GI+ might work out better for unbalanced hands that are better off describing themselves to an ostensibly balanced 1M opener.

[straube]

Just to clarify, but I'm using 1S-2H as 8-11 instead of 7-10. It's only a point difference, but I think it's still significant. I got 18 2S (0-7) responses out of my 100 responding hand which is pretty significant. The 1S-2H response occurred 15 times.

1S-2S leaves us in a good position (not likely going to game) and the opponents in a more difficult position...because it may easily be right for them to balance or not, be in game or not.

1S-2H does let the opponents have a cheap double and a cue bid. Probably their double should be takeout of spades as opposed to showing just hearts and probably their 2S bid should be Michaels. So I think they get a net gain of Michaels. Opponents are always going to make some use of artificial bids whether we're talking Jacoby transfers or Bergen raises or even an IMprecision 1H response.

I think we get a lot for this bid. We don't have to give a limit raise with 11 that takes us to the 3-level opposite a potential 5332 10-ct. Our opener will be able to move slightly more often than if the range were 7-10. The main thing though is separating out the good from the bad raise. 1S-1N with a trash raise leaves a lot more room for the opponents. It's true that we'll occasionally benefit (like finding a superior heart fit) but this room seems much more likely to benefit the opposition.

I kind of like 1S-1N to communicate values and no fit. I think it helps opener in deciding whether to double contracts and how to defend. Say it goes 1S P 1N (4H) ? does responder have something or just a weak hand with spades? Can opener safely double or will responder run? 1S-2S and opener has a better idea of what's going on. Maybe think of it this way. Lots of folks play 1S-3S as weak. If I wanted 1S-3S to be limit and 1S-1N, 2L-3S to be weak, would that make sense?

I think it's a good point that sometimes responder will be unbalanced and sometimes opener will be balanced and in those cases it would be better to have the unbalanced hand be described. However, it's just very hard for responder to pattern properly. There isn't room. When opener is balanced he will be minimum and will seldom be thinking of slam. Again, there just isn't room for responder to show his hand and then communicate extra values. Plus there's no guarantee that opener will stay out of responder's way. Say it goes 1S-2H, 3D. Each partner is trying to describe to the other but at this point probably responder is captain and he just doesn't have the shape information that he would have had he started with a 2C relay. Also, there's a bit of confusion as to who really [/u}is{/u} captain and which strain should be trump.

Perhaps it can be fixed, but at present the natural 2/1 responses we looked at endplay us into 4H whenever an 8-cd fit is discovered. If 1S-2H is GI I would want to be able to stop in 3H. Same for 1S-2m, 2H-3H. How do you show fit for both slam exploration (a forcing 3H) and to limit your hand (a nf 3H)? Maybe we just can't open Axxxx Kxxx Kx xx.



.

[akhare]

straube, on 2014-February-17, 15:56, said:

Just to clarify, but I'm using 1S-2H as 8-11 instead of 7-10. It's only a point difference, but I think it's still significant. I got 18 2S (0-7) responses out of my 100 responding hand which is pretty significant. The 1S-2H response occurred 15 times. 1S-2S leaves us in a good position (not likely going to game) and the opponents in a more difficult position...because it may easily be right for them to balance or not, be in game or not.

OK -- let's talk specific numbers. Out of those 18 2S responses, how many of them were 7 HCPs? In how many cases could both LHO and RHO not make a convenient bid over 1♠ or 1♠ - 2♠?

straube, on 2014-February-17, 15:56, said:

Perhaps it can be fixed, but at present the natural 2/1 responses we looked at endplay us into 4H whenever an 8-cd fit is discovered. If 1S-2H is GI I would want to be able to stop in 3H. Same for 1S-2m, 2H-3H. How do you show fit for both slam exploration (a forcing 3H) and to limit your hand (a nf 3H)? Maybe we just can't open Axxxx Kxxx Kx xx.

Perhaps Adam can comment more on how to handle the 1♠ - 2♥ sequences. One possibility is to simply temporize with 2N (GF).

Edit: Actually, this isn't a difficult auction at all. 1♠ - 2♥ - 3♥ isn't GF. Opener bids 4♥ with no slam interest, splinters with 4+♥, bids 3m (GF) with 4+m and 3♥, 2N (GF) with 5+♠ and 3♥ and 3N with 5332♥ (12-13).

The hand in question should bid 4♥ given that we have a 9-card fit and good controls opposite a 12+ responder (unless feeling especially conservative).

[straube]

oh yeah. 1S-1N (4H) and opener would take the save except he is unaware of spade support.

I don't know how many were 7 point hands. I just know that you wind up with significantly more 2-level spade raises this way than playing 1S-2S. You get 1) the weak hands that otherwise pass or respond 1N and 2) maybe a third of the limit raise hands that would otherwise be shown at the 3-level and 3) a couple more that would have made an over-bid of a preemptive 3S raise or a mixed raise.

The hand I offered, Axxxx Kxxx Kx xx was meant for 1S-2C, 2H-? in which case responder has a 4-cd fit and a balanced 12 count. If I'm not mistaken, there isn't a way to stop in 3H there either.

I feel like apologizing to awm at this point because I don't mean to be critical of a structure that Atul is more or less basing his on. If I'm not mistaken, Adam probably opens lighter but more distributional hands than we do and passes some of the balanced trash that we open. He uses some form of loser count and that's not really how we communicate.

I like 1N semiforcing to be hands that are willing to play 1N. Why take pd out of spades and use hardly any room and risk actually playing 1N? I mean I understand that opener is supposed to find a rebid and the opponents are supposed to be quiet and then responder is supposed to reluctantly correct to 2S and a lot of times it works out ok, but I think a more "honest" and direct approach works better.

[awm]

I do think our opening style is a bit different from yours, in that we upgrade shapely hands more aggressively and we don't really open balanced 10s. There are a few points I can mention about our structure:

1. We don't play "constructive raises" nor do we bid 1NT on any hands with 3+ support for the major.
2. We believe that various shape upgrades justify forcing game on some of the hands you mention. For example, give responder xx AQxxx Axx xxx (which is not even a 2/1 bid) opposite your hand of Axxxx Kxxx Kx xx and game is excellent. Of course if you change opener's hand to Axxxx Kxx Kxx xx (completely flat) then game is bad, but we would not open that hand (maybe you would).
3. The usual way of avoiding game after 1♠-2♥ is to use 2♠ as a "slow down" bid. Of course this becomes trickier if 2♠ is not forcing and can conceal a heart fit.

[akhare]

awm, on 2014-February-17, 20:57, said:

1. We don't play "constructive raises" nor do we bid 1NT on any hands with 3+ support for the major.

This seems to be an interesting treatment.

Does responder typically pass with a flat hand hand and < 5 points when vulnerable and aggressively raise with 3-card support and any hand at any other vulnerability?

Does opener assume a 6-11 range when making a GT?

[awm]

akhare, on 2014-February-17, 22:00, said:

This seems to be an interesting treatment.

Does responder typically pass with a flat hand hand and < 5 points when vulnerable and aggressively raise with 3-card support and any hand at any other vulnerability?

Does opener assume a 6-11 range when making a GT?

We play a wide-ranging single raise (it's about 5+ to 10-). We make make 3-card limit raises directly to the three-level (using cheap jump shifts with space to inquire). With three card support and 0-4 we pass the opening. With four-card support we play 1M-3M preemptive at NV (mixed at V).

[straube]

Here's my latest tally. I'm basing it on how I would attempt to bid hands based on what I know of the two structures.

................relay.....................natural
P..............16........................19
1N............26........................26
2C............19........................5
2D............6..........................7
2H............13........................9
2S.............10.......................18
2N.............1.........................7

3H.............1..........................0
3S.............3...........................3
4D.............1..........................1
4S..............8..........................8

This isn't anything like science. I don't quite know how Atul or Adam would handle certain hands. I think the relay structure has a distribution graph that is favorable. On a side note, I think it's interesting how infrequent the mixed raise comes up and I might change my structure so that 1S-3D is just GI diamonds. I'm also curious what the rest of awm's structure is like (especially what 1S-2N shows) and what Atul would like to see for 2N on up.

Thanks for answering questions, Adam.

[akhare]

straube, on 2014-February-17, 22:35, said:

I'm also curious what the rest of awm's structure is like (especially what 1S-2N shows) and what Atul would like to see for 2N on up.

2N: 4-card, LR+ (including mixed raise when NV)
3C: 3-card LR
3D: Natural invite with good suit, but not sufficient to warrant a 1M - 2D response
3M: Weak raise when NV, mixed raise when vul