[Valardent]

dake50, on Oct 5 2010, 09:52 AM, said:

Try Yellow Rose of Texas follow-ups to Texas.
Shape shown/known, 4 or 5 controls guaranteed.
CONFIT else.

Sorry to ask but, u see, I'm not an english "native" and what u posted is simply incomprehensible to me.

Could you clear that up?

Tx,

Val

[Valardent]

---

1NT 15-17 2♠
2NT(mini) 3♦
3NT(5ctrl) 4♣
4♦ 4♥
5♣ 5♦
5♠2344 5NT
6♦♣Q 6♠
7♣no♠K end.

If teams, one should settle for 6m on 5♠.

If pairs and strong field, and u think 6nt will be popular, u might decide to ask all the hand.

Well, pbbly not the best example but gives an idea how it works.

Val

[Cascade]

Siegmund, on Oct 1 2010, 02:58 PM, said:

Just as a random tidbit...

If you choose to relay out opener's full shape -- the obvious approach, to me, would be to designate one rebid after Stayman as a relay (I am guessing 3C is most easily spared), and take advantage of SOME information exchange at the 2-level. The 2S immediate response can be used instead to plug whatever you lose from your Stayman structure (presumably longer clubs and a 4-card major, if 1NT-2C-2X-3C is the sequence you give up.)

Is there some reason nobody seems to be doing that? (For those of you who still play off-the-shelf Stayman, not the NF variation in onefer's last post.)

We do this:

1NT 2♣ Stayman
any 3♣ is a relay

The first relay asks for controls - responder is always balanced distributional hands go some other way.

After which we bid suits searching for a fit.

3♣ has the additional constraint that we have not yet uncovered a fit. With a major fit we immediately confirm the fit with 3 other Major.

The 3♣ relay seems to work well. Potentially we can use it with any mildly slammish hand or better without a singleton or void. This includes hands with six-card suits.

[Cascade]

Valardent, on Oct 5 2010, 11:07 PM, said:

straube, on Oct 4 2010, 06:22 PM, said:

I didn't understand that last paragraph.  Can you elaborate and give examples?

I did a frequency analysis of a few hands for our 14-16 NT. 

controls....%
3.............7
4.............28
5.............35
6.............25
7.............5

I had the thought that one might use either/or responses as for RKC.  Our ask starts at 3H.

3S-3 or 6
3N-4 or 7
other-5

but I think it leaves too much ambiguity.  Perhaps if we had room for 3 or 7, 4, 5, 6 .  I think you have that room with the extra step.

Hi,

7% for 3 ctrl seems low to me.

These are the numbers I get for 14-16 hcp.

They are exact (hopefully) for that range.

However I have not constrained the hands to be balanced but I doubt that that would make much difference:

1 Control 7.35025E-05
2 Controls 0.005876804
3 Controls 0.064862823
4 Controls 0.254499652
5 Controls 0.40022026
6 Controls 0.232985381
7 Controls 0.040377505
8 Controls 0.001104072

[Zelandakh]

Val, what prevents Opener from having Kx/AKx/Kxxx/Qxxx in your example? Your sequence is not fully explained but what I can decipher of it seems to be consistent with this hand which would seem to be better played in a small slam.

On these Stayman follow-up structures, I am still not really convinced that it is not better to locate a minor-suit fit at the 3 level before asking for controls. You can achieve this by using 2S over 2D or 2H, and by either 3C or 3D over 2S depending on what hands you like to filter through Stayman. Of course you could always use 3C instead of 2S for this too but that seems less efficient on bidding space since this way you also get a bonus min/max piece of information along the way as well as being a step lower.

[Cascade]

The disadvantage of asking for shape first is that you give additional information to the defense on the more frequent occasions where there is no slam and you end up just wanting to play in 3NT.

[Zelandakh]

Cascade, on Oct 5 2010, 11:23 PM, said:

The disadvantage of asking for shape first is that you give additional information to the defense on the more frequent occasions where there is no slam and you end up just wanting to play in 3NT.

My experience of these hands, albeit mostly with a weak NT, is that we are good enough to be looking at slam if a fit is uncovered but not without a fit. Perhaps I am only using this sequence with stronger pairs of hands than you though. In a weak NT system these sequences are typically used with ~18hcp; for a 14-16 NT that would be ~16. Is it really worth looking for slam by Blue Controls with 2 balanced hands weaker than this (assuming appropriate up-/downgradings)?

[straube]

Valardent, on Oct 5 2010, 01:52 PM, said:

---

1NT 15-17 2♠
2NT(mini) 3♦
3NT(5ctrl) 4♣
4♦ 4♥
5♣ 5♦
5♠2344 5NT
6♦♣Q 6♠
7♣no♠K end.

If teams, one should settle for 6m on 5♠.

If pairs and strong field, and u think 6nt will be popular, u might decide to ask all the hand.

Well, pbbly not the best example but gives an idea how it works.

Val

Thanks for the example. I'm concerned that our NT includes off-shapes 5422s and 6322s. Also not sure how your dcb works. What cards are you searching for? What's the order? Seems like it's very how for dcb to be effective.

[straube]

Cascade, on Oct 5 2010, 04:02 PM, said:

Valardent, on Oct 5 2010, 11:07 PM, said:

straube, on Oct 4 2010, 06:22 PM, said:

I didn't understand that last paragraph.  Can you elaborate and give examples?

I did a frequency analysis of a few hands for our 14-16 NT. 

controls....%
3.............7
4.............28
5.............35
6.............25
7.............5

I had the thought that one might use either/or responses as for RKC.  Our ask starts at 3H.

3S-3 or 6
3N-4 or 7
other-5

but I think it leaves too much ambiguity.  Perhaps if we had room for 3 or 7, 4, 5, 6 .  I think you have that room with the extra step.

Hi,

7% for 3 ctrl seems low to me.

These are the numbers I get for 14-16 hcp.

They are exact (hopefully) for that range.

However I have not constrained the hands to be balanced but I doubt that that would make much difference:

1 Control 7.35025E-05
2 Controls 0.005876804
3 Controls 0.064862823
4 Controls 0.254499652
5 Controls 0.40022026
6 Controls 0.232985381
7 Controls 0.040377505
8 Controls 0.001104072

Thanks again Cascade. I'm amazed at how close my tally was to the actual numbers.

After our ask (3H) I'm thinking

3S-4 or 7+
3N-5
4L-6
4N-3 or fewer

I know, I know. We're too high right for when we only have 3 controls? But it's a frequency issue. We'll have 3 or fewer controls only 7% of the time...and 4N will make most days with a combined point count of say 31 hcps.

Thoughts?

[Zelandakh]

straube, on Oct 6 2010, 07:30 AM, said:

After our ask (3H) I'm thinking

3S-4 or 7+
3N-5
4L-6
4N-3 or fewer

You could use a little trick with
3S = 0-3 or 6
3N = 4
4C = 5
4D+ = 7-8

Over 3S, 3NT is now to play opposite 0-3 or a relay opposite 6. This saves you a step and enables you to get out in 3NT opposite all soft-valued openers. Naturally you could reverse the 6 and 7-8 hands if you think that makes the structure easier to remember (although that does not seem to be a major factor for you).

[straube]

Zelandakh, on Oct 6 2010, 09:03 AM, said:

straube, on Oct 6 2010, 07:30 AM, said:

After our ask (3H) I'm thinking

3S-4 or 7+
3N-5
4L-6
4N-3 or fewer

You could use a little trick with
3S = 0-3 or 6
3N = 4
4C = 5
4D+ = 7-8

Over 3S, 3NT is now to play opposite 0-3 or a relay opposite 6. This saves you a step and enables you to get out in 3NT opposite all soft-valued openers. Naturally you could reverse the 6 and 7-8 hands if you think that makes the structure easier to remember (although that does not seem to be a major factor for you).

That's really not bad at all. I was hoping that 4C could be the start of natural bidding (mostly for ease of memory). Yours is technically superior since you avoid playing 4N with the worst response.

[Valardent]

Zelandakh, on Oct 5 2010, 05:59 PM, said:

Val, what prevents Opener from having Kx/AKx/Kxxx/Qxxx in your example? Your sequence is not fully explained but what I can decipher of it seems to be consistent with this hand which would seem to be better played in a small slam.

On 5♣, one kows to end up in a small minor slam at least.

So the question is : What are the risks to investigate 7?

opener cld even have KQ/AKx/Kxxx/xxxx (not odds on)

So 6NT does not rate to be that worse than 6m.

What are the tools on 5♣?

3 sort of relays once shape is partially (44xx, 5xxx, 6xxx) knowned.

1st step : ask to complete shape

2nd step : morphing scan (quacks location)

3rd step : Kings location (by step if 1 of 3 or on a sort of crash principle if 2)

(3rd step must have to be made at or below 5NT, otherwise it's not available. 6x wld be sign off)

The most usefull realy is the morphing scan. (complete shape is rarely crucial and kings location is knowned +/- half of the time)

If a step is by-passed, no way to ask back for it. The kings location must always be made at or below 5NT otherwise it's not available. In the example, 6♠ is an impossible bid (no ♠ fit can exist) and is a sort of patch relay for getting more information in the ♠ suit (1st step would show the king, others no)

The morphing scan works like this (a better method might exist, tx for letting me know) :

- one knows opener is min or max + total of 1-2 ctrl points

One adds to the total hcp AK points 3 hcp (= taking into account opener could hv 3 jacks as 4 woud be too rare).

In the example : 1 Ace + 3 Kings = 13hcp + 3 = 16 hcp

One knows opener is minimum, so 16 hcp with 3 jacks is possible which means in the worst case opener has no queen.

The 1st step to a morphing scan ask = worst scenario case for number of queens

2nd+ step = zoom location of other scenario(s) (only one possible here : one queen since with 2, opener wld be max)

On 5NT morphing ask :
6♣ = no queen
6♦ = ♣ queen
6♥ = ♦ queen
... (priority for Q of longest suit, if equal order is ♥♠♣♦)

Once queen(s) location is ended, a new relay below 6NT is dcb for jacks. (Answers to relays ending above 6NT must bid 7♣.)

Back to the example, to investigate 7, complete shape is crucial (since with 3♠2♥ = no 7) which take some needed space but that's not always the case.

On 5♠, one must judge if 6NT will be good enough.

Not the easiest/precised method but in a natural context of a 1NT opening, it works pretty fine with some (of course) unresolved cases ...

I also use the same structure after
1m 1NT(balanced GF)
2NT (12-14 if 1♣, 18-20 if 1♦)

where the min/max concerned is even more precised, reducing the need to use the morphing scan and making quacks location easier.

To conclude, this is far from perfect but much better than the basic confit method.

If one has examples of balanced slam+ hands to bid, do not hesitate to post it. I'll bid those and one will have a more "impartial" judgement of the relevance of all of this post.

Val

[Valardent]

Hi,

I was just looking the last 16 round.

Board 4 is interesting.

In most of the matches, 6♥ is played going 1 down after ♣ lead and then ♣ruff.

Only Kopecky-Kourka managed to play in 6NT after (it seems) a Heman type seq. where opener could suggest bad trumps. (1NT 12-14 2♦4+♥/2NT(4♥) 3♣/3NT 6NT)

This is the board in question :

---

I'm not that good , I need more bids :

1♣bal 12-14 or ♣ 1NT (GF balanced)
....2NT(max) 3♦ confit ask
3♥ 3 or 4 3♠
3NT 3 4♣ relay
4♥ 4♠
5♣3424 ?

2NT(max) is a (non 12hcp) purish hand which means :

All 13hcp hands without a jack are max (2 exceptions : no queen may be located in a 2 card suit and a 5332/6322 may have the jack of the long suit)

All 14hcp hands with at most one jack are max (even those with a queen in a 2 card suit). One can always downgrade a QJ stiff or the likes. One exception: the 4333 case. This one is max only if the J is in the 4 card suit.

All 14hcp hands with 2 jacks are considered as max if both jacks are located in
4(+) cards suits.

Ouf !

So, at first glance, responder doesn't look better placed since opener could have
AK or KKK with 3 queens in the 1st case and 2 in the second, leaving a lot of possible combinations.

A closer look :

Case 1 : AK + 3Q

impossible unless 14hcp (thus with ♣J) bcs of the ♦Q located in the 2 card suit

6♥ pbbly better than 6NT

Case 2 : KKK + 2Q

odds on with pbbly the black queens (6NT better) or less likely one blackQ/♦Q/♣J (6NT better)

So all one needs is to ask for kings location to distinguish the 2 cases.

...5♣3424 5♥
5♠ (♥K or 3 others) 6NT

Convincing?

Val

[Zelandakh]

Valardent, on Oct 10 2010, 04:20 PM, said:

---

1♣bal 12-14 or ♣ 1NT (GF balanced)
....2NT(max) 3♦ confit ask
3♥ 3 or 4 3♠
3NT 3 4♣ relay
4♥ 4♠
5♣3424 ?

...5♣3424 5♥
5♠ (♥K or 3 others) 6NT

Convincing?

On the system...
After you find out openers' suits your next 3 steps are all slam tries. That essentially means that you are pretty much committed to slamming when you start these relays because so few bids are assigned to sign-offs. Yet one of the most useful functions for this kind of sequence is being able to investigate for a fit on "close to slam" and having safe sign off in 3 or 4NT available where no fit exists. Is this not a problem?

On this specific hand.... no, not at all convincing! First of all I cannot see why 5C shows 3424 and cannot be 2434 - in the previous example you needed a further relay here for exact shape. Secondly, 5H (2nd step) is your morphing relay so 5S (3rd step) is the king ask. Finally, you need to know king location and queen location to bid this hand to 6NT with confidence.

Myself I can reach 6NT if East is opener but only 6H with West opener.

E: 1C = 15+ bal/nat or 18+any
W: 2D = 9+ bal, 4-5 hearts, 2-3 spades
E: 2H = relay, 18+ any
W: 4C = 3=4=2=4, extras (12+), 3 controls
E: 4D = relay
W: 4H = denies heart control
E: 4S = relay
W: 5H = denies HQ, shows controls in other 3 suits
E: 5S = relay
W: 6D = denies DQ, shows SQ, CQ
E: 6N

and the rather simple

W: 1N = 12-14
E: 2C = asks for 5-card majors
W: 2D = no 5-card major
E: 3C = asks for 4-card majors
W: 3H = 4 hearts
E: 4S = RKCB
W: 5C = 0 or 3
E: 6H

You will notice that I use denial bids in my scans. I spent some time comparing the 2 methods and found denials much more efficient, mostly as it is alot easier to calculate when you are safe from getting overboard. You will also see that the 1C auction is more than 1 level lower than the corresponding Confit-like auction and still only just has enough space to gather all the required information. This seems to me to be the Achilles heel of this particular method. The idea suggested earlier of using a bid after Stayman for shape relays, or the long-established method of using 1NT - 2D - 2H - 2S in this way, seems to me to be beneficial as you can arrive at the same point having saved a large amount of bidding space which is pretty much gold-dust when you start scanning.

[Free]

Suppose you play confit in a weak NT:
1NT - 2♠
3♣ - 3♦
3♥ - 3♠ (2-3 controls ; 4+♠)
3NT - 4♥ (2 contols ; 4♥)
4NT - 6NT (doesn't support because of missing top honour ; NT it is)

Only bid 4 card suits if you have a top honour in them. That's why opener doesn't support.

[Valardent]

[Zelandakh]

Valardent, on Oct 11 2010, 09:05 AM, said:

I would be interested to have a look at your complete scan structure, particularly the way you cope with limit slam hands where scanning might get one overboard too often. Did you forsee alternatives ways to describe such hands?

My scan structure is essentially the same as Italian cue bids but done in denial form. Suit order is by length with highest first from ties. Void suits are obviously not scanned. Then bidding the corresponding step denies the appropriate holding. All non-jump suit bids of game from Captain are sign-offs. NT calls in san auctions are used for queen asks. First pass, you give a positive if you hold an ace or a king in the appropriate suit discounting K bare. 2nd pass you give a positive in a suit with the queen, AK or K bare. You always need to calculate if a given scan might get you overboard - I have no special mechanism for handling such hands.

There are of course always alternatives. Most hands have the option of agreeing the major at the 3 level for a normal cue auction. Sometimes it is possible to do this in a minor too. It is also generally also possible to use RKCB in any suit. Sometimes a RKCB auction can be started by 3M but this is rare. These natural routes are usually used on hands that need to cue before looking for controls (open suit) or are missing the trump queen (RKCB often more efficient). Note also that in 'natural' auctions the known hand gives keycards directly even without being asked by 4NT. If a void can be held in such an auction then this is shown by the bid that would have been Kickback. An example is probably good here...

1C - 1H (4+ spades)
1S - 2D (1-suiter)
3S (agrees spades, slam interest)
... - 3N = frivolous (poor slam hand)
... - 4C/D/H = good slam hand, no control in suit bid
... - 4S = total garbage
... - 4N = controls in all suits, good slam hand, side void
... - 5C = controls in all suits, good slam hand, 1 or 4 keycards
etc

That is about it. Much simpler rules than many but I am finding them effective enough and alot easier in practise than 'standard'.