[mikeh]

The only scary thing about this thread is that there are so many lemmings refusing to pass.

I love getting into auctions, but what on earth suggests doing so on this foot?

[jdonn]

Double, I have trouble explaining why but I swear it's good.

Before losing your heads over a 4-2 club fit, consider that we have a really good dummy for a 5-2 club fit. In fact, we have a really good dummy for anything that will happen on offense or defense except a 4-2 club fit (and not even a bad dummy for that but don't tell.) While I can't just ignore the possibility we may reach a silly club contract, all I can say is I'm not even that scared about it because so many good things can happen and they might not catch us anyway.

Ok maybe I don't have trouble explaining why I strongly believe double is good.

[whereagles]

Dbl. Not perfect, but I think I can live with a 2/3♣ response. Anything else is love over gold.

[Apollo81]

I'd still pass, although I am less scared of double/2♥ than you probably are.

[ArtK78]

kgr, on Jan 14 2010, 10:05 AM, said:

2♥ or Pass, depending on the day.

I agree. I would pass on any day ending in a "y," and bid 2♥ on all of the other days.

[aguahombre]

ArtK78, on Jan 15 2010, 09:18 AM, said:

kgr, on Jan 14 2010, 10:05 AM, said:

2♥ or Pass, depending on the day.

I agree. I would pass on any day ending in a "y," and bid 2♥ on all of the other days.

like April Fools or Christmas.

[mikeh]

I earlier posted a criticism of any action but pass, and I remain of the view that pass is (clearly) the best call.

However, if I were forced to bid, I would double....

Double can succeed in virtually all circumstances where 2♥ works except for those hands on which we have a playable 5-3 heart fit....and with this hand, a lot of 5-3 fits will not play well. In the meantime, 2♥ loses almost all of the diamond fits we may have as well as reducing the chances to play in clubs when that is best. I concede that 2♥ also improves the chances of reaching notrump when that is right, since advancer will be more optimistic about our having a spade card after 2♥ than after a double.

But in the meantime, 2♥ risks anything from a moderate loss to catastrophe.

We are red/white. If partner can't raise 2♥... and it gets passed out.... I suspect most of us, before dummy hits, would predict that we are going down...on such a sequence, LHO has short spades and rho some heart length...unless we are lucky, we have to play everything from our hand and LHO will shortly be leading high spades through us while LHO still has a trump or two.

What makes this even worse is that we probably won't like the hand much if partner raises....unless he has 4 card support, in which case double would usually have found the fit anyway.

And of course a bad day could see us going 800 or 1100 at the 2-level when we had a much safer minor suit contract...either undoubleable or only 200-500 against their game.

Double risks a 4-2 club fit, which could be catastrophic but that is only one, low frequency, possibility.

[aguahombre]

I would think that a 4-2 club contract is more likely than "low frequency". And I don't get to ELC into a four-card diamond suit. Hence, "scary" is correct for all non-pass actions.

As Josh said, though, this hand might play well in a 4-2 club fit. I don't have as much practice playing that kind of fit, and would probably screw it up. So, maybe the answer is to double if it is me doubling --but ask partner not to if she has this hand.

[jdonn]

mikeh, on Jan 15 2010, 12:17 PM, said:

Double risks a 4-2 club fit, which could be catastrophic but that is only one, low frequency, possibility.

You said it yourself. So then why is pass "clearly" the best call?

[ArtK78]

aguahombre, on Jan 15 2010, 12:05 PM, said:

ArtK78, on Jan 15 2010, 09:18 AM, said:

kgr, on Jan 14 2010, 10:05 AM, said:

2♥ or Pass, depending on the day.

I agree. I would pass on any day ending in a "y," and bid 2♥ on all of the other days.

like April Fools or Christmas.

April Fools DAY and Christmas DAY.

[mikeh]

I just finished a 30 deal simulation....and I don't do double dummy simulations...which would be particularly misleading in this kind of scenario since it would entail not only dd play but also dd bidding. As it is, my simulation is to some degree subjective since I have to assess how I think the bidding and play would likely go...and I am sure that some would disagree with my views on most hands.

Subject to that, my simulation, leaning over backwards to avoid my pre-judgment that pass was best, showed that on 30 hands:

Pass worked out best on 12.
Either pass or double would lead to a good outcome, but 2♥ would not on 3
Either double or 2♥ would lead to a good outcome but pass would not on 4
double was the only winning call on 1
2♥ was the only winning call on 2
Any action would lead to the same outcome on 8

The latter category included hands on which partner had a good hand and LHO passed 1♠ or hands on which E-W bid successfully to a high spade contract regardless of what we did....altho on one that outcome required that partner with a 2=5=2=4 hand with KQJxx in hearts not take the phantom red/white save.

To the extent that this is reasonably accurate, this suggests that pass is indeed the safest choice...being the only winning call 12/30 times and being a winning call on 15/22....omitting the 8 on which any action breaks even. In addition, there is the problem that when action is better, we sometimes need to guess the right action.

I did not try to analyze the size of losses arising from action, but it was clear that on several hands 2♥ led to large losses....altho in fairness, on one of the poor hands for 2♥, we avoided double because hearts were 4=3 and opener wouldn't, in my view, have dreamt of reopening with a 14 count with KJ9 in hearts.

FWIW, my constraints were dealer 12-21 hcp, 5-7 spades, and no side suit longer than spades. I could have said 10 or 11 as the low end but then expected that this would result in my having to choose which 10 or 11 counts were opening bids....I suspect that few would pass very many 12 counts with 5 spades...and my simulation didn't generate anything that resembled even an aggressive 2♣ opener.

BTW, in response to Josh's last comment....the reason that I felt and now feel even more strongly that pass is best is because while playing a 4-2 club fit will be rare and will even less frequently be catastrophic...not all bad outcomes are because we go for a number. Going -200 against air, or even against 110 or 140 is not a catastrophe, but making a habit of it will lead to the loser's bracket with consistency. My simulation suggests that this issue is as much a problem as the -800 or 1100 disasters.

[jdonn]

That seems like a very good method for you to convince yourself what is the best action, and a very bad method for you to convince anyone else what is the best action. Now my expected criticisms (even in the context of how unscientific the method is):

Sample size is too small.

Of course pass works better than it really would when your opponents never open with less than 12! I'm not sure why you are willing to analyze an entire made up hand in detail regarding how the bidding and play will go, but unwilling to decide whether someone is likely to open the bidding.

Does "winning call" mean it led to the best score on the hand? So if we run two hands and they go like this:
Pass = +150, Double = +100, 2♥ = -500
Pass = -500, Double = +100, 2♥ = +150
Then since pass was the 'winning call' once and double never was, pass is the winner? I don't think so, I just lost 2 imps to you then gained 12 imps! My point being that your yardstick for drawing conclusions is fatally flawed.

I think you should reconsider whether your study should even convince you of what is the best action...

[whereagles]

jdonn, on Jan 15 2010, 06:18 PM, said:

Sample size is too small.

Why?

n = 30 is the point where the student t distribution starts to get very close to a standardized normal, i.e. the point where small sample size effects start to dissipate.

[rogerclee]

whereagles, on Jan 15 2010, 11:23 AM, said:

jdonn, on Jan 15 2010, 06:18 PM, said:

Sample size is too small.

Why?

n = 30 is the point where the student t distribution starts to get very close to a standardized normal, i.e. the point where small sample size effects start to dissipate.

30 is also the atomic number of Zinc.

[mikeh]

1. I can assure you that in looking at the hands, I did not ignore the relative size of gaim/loss issue.... in fact, I find it strange that you would assume that I did. No post is going to contain a full recitation of all of the details. The hands on which pass lost (and there were 7 out of the 30) did not all result in game swings....5 of them were part-score pickups...1 was pushing the opps too high (I could have assumed that the opps would defend at the 3-level with 9 spades and go plus, making pass a winner, but it looked to me that the opps would take the push much of the time, and I wanted to avoid allowing my bias in favour of pass to persuade me to chalk this one up on my side of the issue) and only one of the hands resulted in a game pickup. This should, I think, make sense since we have a crap hand and game will usually require that partner have values and that LHO lacks them....which means, unless he bounces in spades, we are likely to get into the auction after a pass. And, to offset this one game swing, there were several hands on which a telephone number was either a certainty, against most opps, or at least a significant risk. In the meantime, many of the losses from action were 5-7 imps: not insignificant.

2. of course the sample size is small, and I conceded that. I gave the constraints so that anyone concerned with this issue could run their own simulation and render the results somewhat more meaningful. I wasn't running a study intended for review in a scientific journal

3. As for convincing myself....lol.... I have posted enough admissions that the opinions of others have persuaded me that my views were wrong that I would have hoped that you would recognize that I don't always twist facts to meet my biases. As it is, since I actually did some work...and you didn't....it seems to me that it is you who is letting your notions influence your perception of the results. I did not say and do not say that this simulation is conclusive...nor that it is objective....I even specified the reasons why these results should be viewed with caution! What part of that didn't you understand?

4. As for the 11 point idea....I told you why I didn't do it...it is the same reason I didn't analyze all 100 hands I generated....it was too time consuming relative to the importance of the issue to me. But I think you;d agree that for a substantial part of the bridge playing population, many 11 counts don't get opened in 1st chair..while many do. In a 30 deal simulation encompassing a 11 point hcp range for the opening bid, the reality is that one might expect maybe 1 or 2 additional hands where the consensus, even on this forum whuch is generally aggressive in terms of opening requirements, is that the hand is an opener. Assume that there were 2 such...assume that on both of them, action was demonstrably better than pass....the overall inferences to be drawn from the simulation would still be that pass was the best action.

[Vilgan]

Passing all 11 counts seems pretty strange. Seems safer to open them all imo for the simulation. Also, imps that change hands also seems to have relevance. Obviously gaining 2 imps for -50 vs -110 is not as important as -500 vs -140.

I think its hard to simulate pass vs double as there will be a lot of differences in how partnerships will work together after the X. I would personally suspect that pass is slightly superior but wouldn't be surprised to be wrong. I do think that just about any simulation will show 2♥ to be a crazy bid against decent opponents though. Also a bunch of partnership implications if you CAN overcall 2♥ on that hand that will make other auctions/defense/bidding sequences harder imo.

edit > oops, posted after you did. Sounds like imp implications were looked at

[jdonn]

mikeh, on Jan 15 2010, 01:40 PM, said:

1. I can assure you that in looking at the hands, I did not ignore the relative size of gaim/loss issue.... in fact, I find it strange that you would assume that I did.

I didn't assume anything. You talked about winning actions and losing actions, not how many imps were changing hands. What does that look like it means to you?

mikeh, on Jan 15 2010, 01:40 PM, said:

3. As for convincing myself....lol.... I have posted enough admissions that the opinions of others have persuaded me that my views were wrong that I would have hoped that you would recognize that I don't always twist facts to meet my biases.

I didn't mean that as a knock on you. It's a very general observation. If you generate hands using very broad specifications, analyze them in detail yourself, and don't show them to anyone else, then of course that will be very convincing to you. After all you just spent a long time examining them. But it will not be very convincing to anyone else, since it was totally unscientific and subjective and no one else even saw the hands.

[aguahombre]

I think Vilgan has coined a new word..."IMPlications".

[Jlall]

crack.

lol at simulating this

[MarkDean]

For me, this is an appropriate thread title. I find anything other than pass scary.