Monday, 18 June 2012

Slivers

It's a frantic week for me coming up, so although I have a couple of substantive subjects to blog about, for now I'll settle for tidying up a few little scraps here and there.

I've done some housekeeping on the website. There's some more material in the section on Fit Bidding, and I've briefly summarised Attitude Asks in the section on new ideas. While doing that it struck me as good idea to link back from the website to the relevant blogs, so I've started doing that with links from the page on shunts.

How do voids affect keycards asks? If you have a void which you have splintered, you can ask keycards and follow up with a feature ask to see if partner has the no-ace; and if partner has splintered, you can follow up similarly to see if that suit was void. The problem with both of these is that you often don't have the space to do it.

Consider the routine auction 1♠ - 4; 4 - 4NT; 5. Responder's splinter finds opener with values-for-five, and the keycard ask finds two. All very encouraging but - would you adam and eve it - you've gone past 5, so can't find out if one those keycards was the diamond ace opposite your void.

Or conversely: 1♣ - 2; 2 - 2NT; 4♣ - 4♠; 5. If you've read p124-125 you will know that 4♣ is a splinter, and 5 shows two keycards, but you've lost the chance to ask if the splinter was a void.

There has always been a solution to first sequence: the exclusion ask. Instead of asking with 4NT, use 5 instead, and partner will exclude the diamond ace from the answer. But the exclusion ask doesn't help in the second case.

Or does it? What does an exclusion ask in partner's splinter suit mean, when the splinter has already denied an ace in that suit?

The answer could be that responder should be adding one to the response to a normal keycard ask if void in a suit previously splintered. Playing that way, with an ace in partner's splinter suit you use an exclusion ask instead so that one is not added for a void. That all seems logical, and preserves the meaning of the exclusion ask, but it needs some testing, both for functional and psychological soundness.

Speak again soon
Alan

The Chilli bidding system is described and defined in the book What Can Possibly Go Wrong? and summarised on the Chilli bidding website.

11 comments:

Anonymous said...

Hi Alan,
I was analysing a bit the implications of the new 1D opening and found a type of hand I wouldn't know how to handle. Suppose the partership has this quite "vanilla" combination (where suits are in order SHDC):

Axx
Qxxx
x
AKQxx

and

KQJxx
xx
Axx
xx

In a modern natural system, the bidding sequence would probably be:

1C - 1S
2H (reverse) - 2S (five spades, one round forcing)
3S (three spades) - 4S

Let's see it in Chilli. If I have understood everything correctly (which I am not sure of):

1D (24 playing points, not five in a major) - 1S
2C (disturbing, because limited hand in front of unlimited) - ?

Now, if bidding was still neutral responder could say 2S, which in this case would work just as in the natural sequence (i.e. forcing with five spades); but it's disturbed, so 2S is to play - and I think opener has to pass, as responder could still have zero points! A waiting 2NT could do the trick, but 2NT would be strong agreement of clubs here, so I guess responder has to bif 3S to force, and opener can then bid 4S. But... a jump rebid of a five-card suit? Shouldn't it be a suit-setter with six cards? And if not, how could responder bid a six-card suit then?

I mentioned all this because that hand seems to me representative of a whole class of situations induced by the bidding often switching to disturbed at the second round.

Thanks and all the best,

Andrea

Alan Williams said...

Hi Andrea

There may be some hands where opener's disturbing rebid cause irritation, but this is not one of them. Opposite partner's 1S response you can raise to 2S.

This idea is mentioned on p37 of the book under 1H - 2H, but applies equally opposite any bid of one a major: a hand with minimum values, three-card support and a shortage should raise to two immediately.

So the auction goes 1D - 1S; 2S - 4S. As a bonus, the opening leader knows very little about dummy's shape.

Best wishes
Alan

Anonymous said...

Hi Alan,

this scheme doesn't look too hard to remember and has potential for gains.

The cost is perhaps where the splinter is just under the agreed suit. For example spades are agreed and 4H is the splinter. Say you need two key cards for slam, 4NT gets 5H or 5S with 2 key cards or one and a void with no room to ask for no ace while 5H gets 5N with just one keycard.


Opener has:

AQxx
Axx
KQxx
Qx

1D - 1S
2S - 4H
?

Responder could be either of the following:

Jxxx
-
AJxxx
KJxx

Jxxx
x
Axxxx
AKx

The first hand is a reasonable splinter, if opener's heart ace was the spade king or club ace then slam is a pretty good bet.

I'll be interested to hear how they work out in practice.

Michael

Anonymous said...

Another question about counting a void. The other night this hand came up.that was misbid by our opponents to 7S

AQxxxx
AKQJ
x
xx

Kxxx
x
Axx
AQJxx

I think the chilli auction would have gone

1C - 1S
2N - 4H
4S - 4N 0/3 you should have asked
5C - 5S no Q & no DK & no H void
6C - 6S no CK

Now if you should have asked can contain a void this hand is more difficult to bid.

Since partner has already shown disinterest opposite shortage in the suit, a void probably won't be of much use. For this reason I think you have asked is best played as only key cards.

By the way you should have asked is a bit cumbersome, I can't see it catching on:) Maybe key card tell?

So simple rule - when asked count the void, when telling only key cards.

Michael

Anonymous said...

Actually I've realized that 4H in the example auction is not a splinter because 3S is not a waiting bid. Anyway my point remains that key card tells don't count the void.

Michael

Alan Williams said...

Hi Michael

On the 23 June comment ...

These examples do highlight the possible downside when the splinter is one below trumps.

As an aside, opener's club queen is almost irrelevant opposite both hands, and without it perhaps opener - while liking the splinter fit - might not have enough to justify a keycard ask, although it could be argued that, since opener is already limited, the 4H splinter shows values for five+.

On that last point I've been rethinking quite a bit around continuations after splinters, but haven't come to any good conclusion yet.

On the 24th June comment ...

A few points from your auction ...
as your correction says, 4H is not a splinter, but even had it been then it was never my intention that 4NT would be You should have asked, which was only meant to be after an explicit values-for-five bid and partner's negative response.

I can see the logic, but extending it to any situation where partner declines one's first effort pulls in all sorts of other auctions such as 1H - 2NT; 4H - 4S/4NT/5C and indeed your auction in the previous comment.

Also you continue with a 6C 'ask' - but in book Chilli it's an alternative place to play, as all slam bids are (other than responses to asks). I prefer this approach because of the possibility of hidden fits, but allowing asks to roam above 5NT is a logical alternative which will gain on some hands.

I will think some more about your substantive point ... (yes, keycard tell is better!)

Alan

Anonymous said...

Fair point about the 6C, wishful thinking doesn't make it so:)

I wouldn't include that last auction. 2NT by itself isn't showing values for 5 whereas the splinter does, at least if partner has no wasted values.

The other point is that a splinter shows a quite specific hand type as does the 4H bid directly opposite 2NT. It makes sense that the better described hand continues to describe their hand rather than switch to asking (or being told).

Michael

Anonymous said...

Thanks a lot for your answer, Alan. I had not thought of your simple and elegant support-with-3-cards solution.
Being the fastidious nature I am, I asked myself: what if opener's hand has no singleton? I think in that case opener's rebid after 1D-1S could always be 1NT, so bidding stays neutral and responder can serenely bid his forcing 2S.
Best wishes,
Andrea

Andrea said...

No, sorry, I was wrong. I was thinking in original Chilli. Now 2S opposite 1NT should be a "sandpit" raw bid, so the right bid should be the compelling 2D instead.

P.S.
The "captcha" makes posting here a nigthmare. I'm sure a robot would fare better.

Anonymous said...

Just thinking out loud. Assuming diamonds are agreed with 4H, values for 5 & partner signs off in 4NT are the Key Card Tell bids as follows?

5C - 0/3 KC
5D - I really prefer diamonds to NT
5H - 1/4 KC
5S - 2 KC

5D seems a likely spot to want to play but there is memory overhead. Perhaps not that much, partner just jogged our memory that 4NT is special in a minor fit.

Michael

Alan Williams said...

Hi Michael

Yes, when using a keycard tell you have to skip sign-offs. 5D here is a better place to play.

Alan