[FrancesHinden]

lexlogan, on Jul 8 2008, 02:51 PM, said:

This hand is a bit better than 8 hcp, due to the 10's and 9's. Double-dummy analysis, for what it's worth, emphatically rejects adding a point for a 5 card suit at notrump. Peter Cheung's website, for example, shows 24 hcp with a five-card suit to be only about 2% better than without -- and well under 50%.

It's worth virtually nothing.
(I know, I'm like one of Pavlov's dogs)

The reason that DD analysis does not value the 5 card suit is that when you are in a light HCP 3NT with a long(ish) suit in one of the hands the result will depend hugely on the choice of opening lead - much more so than on declarer play. So they don't balance out at all; the DD analysis will favour the defence.

Take this example hand opposite a strong NT opener:

xx
xx
AKQxxx
xxx

Any bridge player would raise to 3NT. I am certain that single dummy analysis shows that this is right and that it will make far more than 50% of the time (I've done the analysis for about 100 deals and it is pretty conclusive)

But I bet you DD analysis shows that 3NT is going off the majority of the time.

Alternatively, when you are in a light HCP 3NT bid on the basis of lots of good honour combinations and 10s and 9s with every suit well stopped, the DD analysis will favour declarer who will get every guess right.

[kenrexford]

pclayton, on Jul 8 2008, 11:42 AM, said:

Very interesting idea Ken.

The 4-3 major fit did not occur to me, but there's a lot of reasons why a Moysian could be a good idea:

--we have a ruffing value
--our points are sharp

Heck, what about the 5-3 fit? Partner could easily have some 5332 with a major.

Another point that was mentioned was that 2♣ also has that (rare?) potential benefit of blasting the opponents out of the auction. Of course, that also might be a bad thing if nobody makes nuttin'.

[jdonn]

Another reason a 5 card suit adds more value in practice is that if you can run the suit you force the opponents to make many discards, and often give away an unmakeable game.

[nige1]

kenrexford, on Jul 7 2008, 10:35 PM, said:

♠109x ♥109x ♦Ax ♣Kxxxx
Partner opens 1NT.  15-17, first seat.

IMO 6♣=10, 4N=9, 5♣=5, 3N=4, 4♠=3
My simulations prove that slam is playable opposite as little as ♠ AKQ8 ♥ Ax ♦ xx ♣ Axxxx
(Seriously, I do understand your point, Ken)

[kenrexford]

nige1, on Jul 8 2008, 09:52 PM, said:

kenrexford, on Jul 7 2008, 10:35 PM, said:

♠109x ♥109x ♦Ax ♣Kxxxx
Partner opens 1NT.  15-17, first seat.

IMO 6♣=10, 4N=9, 5♣=5, 3N=4, 4♠=3
My simulations prove that slam is playable opposite as little as ♠ AKQ8 ♥ Ax ♦ xx ♣ Axxxx
(Seriously, I do understand your point, Ken)

Don't you think that the spade Queen is overkill, though?

[skjaeran]

kenrexford, on Jul 9 2008, 05:45 AM, said:

nige1, on Jul 8 2008, 09:52 PM, said:

kenrexford, on Jul 7 2008, 10:35 PM, said:

♠109x ♥109x ♦Ax ♣Kxxxx
Partner opens 1NT.  15-17, first seat.

IMO 6♣=10, 4N=9, 5♣=5, 3N=4, 4♠=3
My simulations prove that slam is playable opposite as little as ♠ AKQ8 ♥ Ax ♦ xx ♣ Axxxx
(Seriously, I do understand your point, Ken)

Don't you think that the spade Queen is overkill, though?

That's why he said 'as little as'...

[benlessard]

Pass Wtp. Even if the club suit run we might be short of tricks. If the hand is a bit stronger i would bid 2C and
bid 3Nt over 2D knowing partner is likely to have 3C.
Pass 2M. Wich will lead to a 4-3 that will be as good as 1Nt.

[Stephen Tu]

FrancesHinden, on Jul 8 2008, 10:44 AM, said:

Take this example hand opposite a strong NT opener:

xx
xx
AKQxxx
xxx

But I bet you DD analysis shows that 3NT is going off the majority of the time.

You lose? MY DD over 2000 hands shows making ~70% of the time with this hand.

Original problem:
makes DD only 19% of the time. So to me it's impossible to support blasting 3nt over passing & I highly doubt inviting works better either.

I'll investigate the other approaches (inv 2nt, ken's play moysian/invite over 2d strategy) later when I have time.

[nige1]

kenrexford said:

♠109x ♥109x ♦Ax ♣Kxxxx
Partner opens 1NT.  15-17, first seat.

nige1, on Jul 8 2008, 09:52 PM, said:

IMO 6♣=10, 4N=9, 5♣=5, 3N=4, 4♠=3
My simulations prove that slam is playable opposite as little as ♠ AKQ8 ♥ Ax ♦ xx ♣ Axxxx.
(Seriously, I do understand your point, Ken)

kenrexford, on Jul 8 2008, 10:45 PM, said:

Don't you think that the spade Queen is overkill, though? 

Perhaps, on a good day, ♠ AQ8x ♥ Ax ♦ xx ♣ Axxxx

[NickRW]

On a small sample (50 deals), Ken's idea (2C-2D-3N):
7 or less tricks=21, 8=13, 9=16

Normal invite and acceptance (1N-2N-3N sort of thing - or however you invite if indeed you can):
7 or less tricks=12, 8=24, 9=14

Combination of the two (1N-2C-2D-2N-3N):
7 or less tricks=9, 8=16, 9=25

Opener's acceptance criteria = 17HCP or 16 and at least one of the remaining 10s.

The latter strategy would seem to be just about worth it at IMPs. At MP, given that there would be many unaccepted 2N deals that only ended up with 7 or less tricks, still not worth it.

Nick

P.S. Later edit. I should add, that even if you are playing IMPs with this hand, your stayman+invite would only really be worth it if partner's acceptance criteria are as high as better 16s and 17s. If pard's standard is lower, then you're obviously on dodgier ground.

And even if pard's 3N acceptance criteria are as high as a better 16, 15s and poor 16s are probably commoner than 17s and better 16s - so even here the cost factor of having some failing 2N contracts needs to be weighted relatively heavily.

[kenrexford]

NickRW, on Jul 10 2008, 09:20 PM, said:

On a small sample (50 deals), Ken's idea (2C-2D-3N):
7 or less tricks=21, 8=13, 9=16

Normal invite and acceptance (1N-2N-3N sort of thing - or however you invite if indeed you can):
7 or less tricks=12, 8=24, 9=14

Combination of the two (1N-2C-2D-2N-3N):
7 or less tricks=9, 8=16, 9=25

Opener's acceptance criteria = 17HCP or 16 and at least one of the remaining 10s.

The latter strategy would seem to be just about worth it at IMPs.  At MP, given that there would be many unaccepted 2N deals that only ended up with 7 or less tricks, still not worth it.

Nick

P.S. Later edit.  I should add, that even if you are playing IMPs with this hand, your stayman+invite would only really be worth it if partner's acceptance criteria are as high as better 16s and 17s.  If pard's standard is lower, then you're obviously on dodgier ground.

And even if pard's 3N acceptance criteria are as high as a better 16, 15s and poor 16s are probably commoner than 17s and better 16s - so even here the cost factor of having some failing 2N contracts needs to be weighted relatively heavily.

A few minor points of clarification.

"Ken's idea" was actually "Ken's idea," but a different Ken, relayed through Ken to the forum.

Also, the actual idea was the combination, it seems. Bid 2♣, planning to pass 2M but bid an invitational 2NT if Opener bids 2♦.

To assess the merits of this idea, however, it seems that you need to determine the effect of three likely contracts:

2M as the end contract.
3NT as the end contract, when Opener raises.
2NT as the end contract, when Opener passes.

You also want to determine the rate at which passing will yield competition, and what happens to that competition.

[NickRW]

kenrexford, on Jul 11 2008, 03:31 AM, said:

A few minor points of clarification.

"Ken's idea" was actually "Ken's idea," but a different Ken, relayed through Ken to the forum.

Yes. I realise "Ken's idea" was shorthand.

Quote

Also, the actual idea was the combination, it seems.  Bid 2♣, planning to pass 2M but bid an invitational 2NT if Opener bids 2♦.

To assess the merits of this idea, however, it seems that you need to determine the effect of three likely contracts:

2M as the end contract.
3NT as the end contract, when Opener raises.
2NT as the end contract, when Opener passes.

You also want to determine the rate at which passing will yield competition, and what happens to that competition.

Yes. To do full justice to the analysis requires several hundreds of hands all where opener is simply 15-17 bal and then follow up what would happen on each one. Frankly I couldn't be bothered last night. What I posted was the nearest I could get to a proper analysis quickly.

Nick

[kenrexford]

I know. I hate problems like that. The decision as to whether to pass or to bid 2♣ is not simply one of hand evaluation for a contemplated specific contract (like 1NT-P-3NT, good or bad).

The possible continuations that I can imagine are actually wildly disparate.

1. Partner bids 2M and this ends the auction. In that event, we need to consider 4-card and 5-card fits, for starters, and how they play out in comparison to 1NT. However, we also need to consider (with me) that rare 2M contract with a 3-3 fit, as I or my partner might, on rare occasion, opt to bid 2M after Stayman with 3-2 in the majors, Honors in the fragment but xx in the doubleton. Plus, we need to consider whether 1NT would have bouth the contract, and, if not, what would have happened otherwise, and whether 2M buys the contract, as opponents sometimes balance after 2M (now at a presumably higher level). We might even see a rare 2♣X sequence, or even a 3♣ or 3♦ response from partner to Stayman.

2. If partner bids 2♦, it is not assured that this will stop competition either. RHO might double or bid here, strangely.

3. If we do bid 2NT after 2♦, the opponents doing something weird seems rare. However, partner may well introduce a six-card minor, which would be nice, albeit equally weird. But, in the event of a pass, we need to know what happens to 2NT, not just as a contract but against possible alternative contracts had the opponents intervened in the 1NT-P-P sequence.

4. 3NT.

[pigpenz]

simulations shows that:
80% 8 tricks
24% 9 tricks

if behind in a match and needed to gain might be a hand to push on

[rogerclee]

pigpenz, on Jul 11 2008, 08:01 AM, said:

simulations shows that:
80% 8 tricks
24% 9 tricks

if behind in a match and needed to gain might be a hand to push on

What about the other negative 4% of the time?

[NickRW]

kenrexford, on Jul 11 2008, 02:07 PM, said:

I know. I hate problems like that.

You're getting a little too complicated for me Ken. There is the matter of what defence to 1N the opps are playing.

To me, the only way to tackle this without getting hopelessly lost, would be to assume a head to head team of 4 match and that, at the other table, the auction is 1N passed out. Your team mates play a given. recognised and fairly aggressive defence and decline to use it - say something like Lionel as that allows all 44 shapes to be overcalled apart from 44 in the minors, plus the 5M and 6m shapes and (if you're up for it) to do so with not too much HCP - therefore the opp hands can be assumed to not be any of the types that the given defence would allow. Further, one assumes that one's team mate's defence covers all the hands our opps would overcall with at this table. While this assumption is not necessarily realistic, it simplifies the analysis as the option of us bidding normal Stayman obviously disappears if there is, say, a 2♥ overcall from RHO - so the question of what would happen if we chose the Stayman route to deal with this responding hand becomes academic.

Maybe I'll give it some time later. Or maybe not!

Nick

[Stephen Tu]

DD results, 1000 deal samples:

straight invite: (1nt-2nt), opener accepts 16+
makes 3nt (game swing): 17.2%
goes down in 2nt/3nt when 1nt makes (partial swing): 53.1%
goes down in 2nt/3nt when 1nt still goes down (1-2 imps assume not x'd): 11.4%
makes 2nt (push): 18.3%

conclusion:
passing >> inviting > blasting all forms of scoring.

stayman, pass 2M, 2nt over 2d:
makes 3nt: 6.6%
makes 2nt (push): 5.7%
goes down in 2/3NT when 1nt makes: 13.9%
goes down in 2/3NT when 1nt down: 2.2%

makes 2M when 1nt down: 4.4%
down 2M when 1nt makes: 7.9%
both 2M / 1NT make, M matchpoints better: 36.8%
both 2M / 1NT make, NT matchpoints better: 19.1%
both 2M / 1NT down, M down less: .7%
both 2M / 1NT down, NT down less: .4%
both 2M / 1NT down, same score: 2.3%

conclusion:
IMPS - clearly passing best
MP - might be worth taking shot for major partial if partner routinely opens 5cM in range with 1nt (the sim condition). Didn't do sim with more restrictive 5cM criteria, maybe later.

[NickRW]

Stephen Tu, on Jul 11 2008, 11:00 PM, said:

DD results, 1000 deal samples:
<snip>

I see, you beat me to it. I started a deeper analysis but preliminary results show fairly similar figures. I do get a slightly higher frequency of our contracts making in general, but this may well be due to the restrictions I placed on the opp hands to simulate them being quiet in the auction - the upshot would be that opener holds 17 HCP slightly more frequently and the breaks would tend be a bit more benign on average.

Even so, the times when we make 3N after 2C-2D-2N-3N do not appear to be enough to compensate for the times when 2N is off and also the cases where 2M fails but 1N makes (which appears to be more likely than 2M making and 1N failing)

Nick

[pigpenz]

Jlall, on Jul 11 2008, 08:19 PM, said:

Stephen Tu, on Jul 11 2008, 06:00 PM, said:

conclusion:
passing >> inviting > blasting all forms of scoring.

I don't think you can say this about single dummy though. 1N-2N I'm sure is better than 1N-3N, but most people have to bid 2C which will give away a lot of info about partner's hand, or bid 2S which will help them lead. I would bet in real life 3N is better than 2C then 2N since it's easier to make them misdefend on that auction. Also, most people (people I play with) accept even more aggressively than your sim suggests. This means that you don't end up in 2N when it is cold very often anyways, so I would hate to give away the info which might lead to 3N going down instead of making to cater to a possibility that doesn't happen that often. Anyways, agree with passing being a clear favorite.

yes if you had 8hcp and those 109 combinations in 2 diferrent suits the best bid would be a direct 3nt.