[neilkaz]

Mark me down as favoring 3♠ as well with a goodish 5 card suit. I'll take my chances with a hand this strong but if I had somewhat less, I'd pass.

[gnasher]

I once opened 3♣ in 3rd seat on an 18-count, and later got to double 4♦ for 800. However, I don't think enough people do this often enough for us to worry about it.

Some people may preempt on a 12- or 13-count in third, but they hardly ever act again. Hence we're only rsking an undoubled penalty.

[FrancesHinden]

aguahombre, on 2013-September-11, 19:45, said:

I don't understand why people believe the choice is between 3S and DBL followed by 3S. The choice is between a takeout double and an overcall.

I don't think anyone has suggested dbl followed by 3S as an option. The side-tracked discussion has been about how much more that auction would show, not that anyone actually wants to follow it on this hand.

[gnasher]

FrancesHinden, on 2013-September-12, 14:30, said:

I don't think anyone has suggested dbl followed by 3S as an option. The side-tracked discussion has been about how much more that auction would show, not that anyone actually wants to follow it on this hand.

I did sort-of suggest it as an option, or at least I suggested that I'd like it to be an option. If it showed this hand, it would obviously be better than simply bidding 3♠.

[aguahombre]

FrancesHinden, on 2013-September-12, 14:30, said:

I don't think anyone has suggested dbl followed by 3S as an option. The side-tracked discussion has been about how much more that auction would show, not that anyone actually wants to follow it on this hand.

When multiple posters only said they would bid 3S and that Double followed by 3S shows more, that sounded like they were weighing an overcall versus a 2-step sequence ---and not considering a takeout double itself. But, that's just me reading words and trying to understand them.

It has occurred to me that in the poll referenced, maybe the contributers had the same problem and chose 3S for that same reason. Anyway, I put myself in that chair predicting the results of polled experts but not taking a position on 3S vs X.

At the table, I'd be thinking 3S wins or breaks even when Spades should be the strain, and that double would lose on a few of those; while, Double could gain when a red suit or 3CX should be the strain ---and would probably break even when we get to Spades anyway.

It is slightly more complicated than that. We must also consider whether ---if Spades do become trump---the X or the overcall would have helped us get to the correct level.

[JLOGIC]

Can't speak for anyone else but I was just replying to this:

Quote

Does double followed by 3♠ show this, or should we have more of something?

When I said X then 3S should show more. Obv shoulda quoted it to make it clear.

[jdeegan]

Interesting thread. It really separates the sheep from the goats. The real players all bid 3♠. Double is not insane, but it is clearly inferior - and for the same reason why you should bid 1♠ over a 1♣ opener by RHO. A pass is really timid.

Timo has his ranking system. Mine is even simpler. On the occasions where my bid disagrees with JLOGIC, I go back to the drawing board to see where I screwed up.

[Lorne50]

I ran a simulation with the first 2 hands having no bid, 3rd hand a normal 3C opener (7 cards or a good 6 card suit) and 4th hand fixed as shown. Results for number of tricks made over 1000 deals were:

. <=8 9 10+
♣ 528 283 189
♥ 613 190 197
♠ 376 254 370
NT 806 93 101

Difficult to be certain but it looks to me that 3♠ is a winner - it beats pass given that it makes 624 times (unless partner raises a lot when he shouldn't and 3♣ also makes quite often if you pass.

What I can't model is whether dble gets us a decent score since partner may pass for penalties or bid a major or bid 3N but gut feel looks like 3S is a winner.

[TylerE]

I don't think that's a valid aet of sims for determining this.

How do we do in ♥ when responder's ♥ are 2 or more cards longer than spades?

Or better yet, do this:

Generate 5000 random responder hands

If responder holds 4 good, or any 5 clubs defend 3♣ X
Otherwise we play in responders longest suit, with ♥ rather than ♦ when equal length.

That will come close to telling us how X does. I'm sure ♥ is worse than ♠ on random hands, but how about when responder actually has ♥?

[quiddity]

Thanks for the sim; I agree that passing looks pretty terrible based on those numbers.

[jdeegan]

TylerE, on 2013-September-13, 09:10, said:

I don't think that's a valid aet of sims for determining this.

How do we do in ♥ when responder's ♥ are 2 or more cards longer than spades?

Or better yet, do this:

Generate 5000 random responder hands

If responder holds 4 good, or any 5 clubs defend 3♣ X
Otherwise we play in responders longest suit, with ♥ rather than ♦ when equal length.

That will come close to telling us how X does. I'm sure ♥ is worse than ♠ on random hands, but how about when responder actually has ♥?

I agree with your sentiment, but I would broaden it to include sims in general (for the most part). For most bridge hands they get very tricky pretty quick when all four hands are in consideration. As a trained statistician, the only ones I trust usually involve a huge amount of labor. First, you have to cast a very wide net and run, say, 100 sims that never leave out a possible hand. Then you have to cull out (by hand) the ones that don't fit the conditions. That, typically, leaves, maybe, 50 hands. Then you have to analyze each one by hand taking into account possible bids by opponents and possible opening leads. Assuming double dummy defense by a robot that peeks is often a surefire way to get a wrong answer.

On this hand, you can very quickly use ordinary probability calculations to, at least, get the expected number of cards in each suit for each hand. With a little thought, you might be able to come up with a way to calculate the actual variances as well.

Come to think of it, if all you are after is suit distribution, Pavlicek's hand generator might do a quick and dirty job. Just don't forget to include East (with at least six clubs) and then cull out the occasional odd hand where he has too many clubs or a second long suit.

The expected values are:
East: 2-2-2-7
South: 5-3-3-2
West: 3-4-4-2
North: 3-4-4-2
This alone, should show you why 3♠ is the correct bid.

[PhilG007]

TylerE, on 2013-September-11, 08:43, said:

X looks clear to me.

A double in the direct seat would be for takeout....the last thing you want
with that holding in RHOs suit(!) Pass is the correct action on this hand.
If it's passed round to partner and HE doubles for takeout(as he should)
you can safely pass it for penalties...and prepare to enjoy the coming feast(!)

[gordontd]

PhilG007, on 2013-September-13, 16:39, said:

A double in the direct seat would be for takeout....the last thing you want
with that holding in RHOs suit(!) Pass is the correct action on this hand.
If it's passed round to partner and HE doubles for takeout(as he should)
you can safely pass it for penalties...and prepare to enjoy the coming feast(!)

I think you must be looking at some other hand.

[rhm]

jdeegan, on 2013-September-13, 11:14, said:

I agree with your sentiment, but I would broaden it to include sims in general (for the most part). For most bridge hands they get very tricky pretty quick when all four hands are in consideration. As a trained statistician, the only ones I trust usually involve a huge amount of labor. First, you have to cast a very wide net and run, say, 100 sims that never leave out a possible hand. Then you have to cull out (by hand) the ones that don't fit the conditions. That, typically, leaves, maybe, 50 hands. Then you have to analyze each one by hand taking into account possible bids by opponents and possible opening leads. Assuming double dummy defense by a robot that peeks is often a surefire way to get a wrong answer.

On this hand, you can very quickly use ordinary probability calculations to, at least, get the expected number of cards in each suit for each hand. With a little thought, you might be able to come up with a way to calculate the actual variances as well.

Come to think of it, if all you are after is suit distribution, Pavlicek's hand generator might do a quick and dirty job. Just don't forget to include East (with at least six clubs) and then cull out the occasional odd hand where he has too many clubs or a second long suit.

The expected values are:
East: 2-2-2-7
South: 5-3-3-2
West: 3-4-4-2
North: 3-4-4-2
This alone, should show you why 3♠ is the correct bid.

I wish I had your certainty what is right. I am not convinced at all.
Even if spades is your most likely successful strain, it does not follow that 3♠ is the right bid, mainly because this bid puts most of your eggs in one basket.
It is obvious that after 3♠ you can not play any more 3♣ doubled, 3 ♦ or 3♥ and you are unlikely to reach game in a red suit if that is right.
Is spades so much more likely the right strain that we should ignore all those alternatives combined?
If West has four or more spades we will likely reach spades anyway after DBL, that is DBL does not preclude playing ♠, if our spade fit is good, but overcalling 3♠ will preclude many other contracts.
If West has less than three cards in spades it is very likely we belong somewhere else, including 3♣ doubled or a red suit. Neither do I relish the thought that a vulnerable three level overcall might get raised on two cards,
Overcalling 3♠ is probably best if West has precisely 3 spades with less than 5 cards in hearts.
I do not need Pavlicek to know that partner is much more likely not to hold this precise combination in the majors
Also when we have only a 5-3 fit in spades, game in spades might not be so hot when partner is a passed hand and a preempt makes bad breaks more likely than usual.
It is not uncommon that 3NT plays better in such scenarios.
If partner is weak we might belong in a red suit even though he has three spades.
It is true that since LHO is passed the risk that we run into a penalty DBL diminishes, but a 3♠ overcall on a five card suit is still more dangerous than DBL.

Rainer Herrmann

[benlessard]

I agree with Rainer, there is just too much thing that can happen for a simple sim to make any sense here. For example X may rightside a S contract or advancer may raise to 4S with only 2 spades etc..

[aguahombre]

rhm, on 2013-September-14, 08:40, said:

I wish I had your certainty what is right.

No, you don't.

[MrAce]

gordontd, on 2013-September-13, 16:56, said:

I think you must be looking at some other hand.

[aguahombre]

rhm, on 2013-September-14, 08:40, said:

If West has four or more spades we will likely reach spades anyway after DBL, that is DBL does not preclude playing ♠, if our spade fit is good, but overcalling 3♠ will preclude many other contracts.

Earlier someone advanced the contention that Pard with 4-4 in Majors would bid 3H over the double, and since we don't intend to bid 3S after doubling, the Spade suit would be lost. But I agree with you, instead. Right or wrong, I will bid 3S, not 3H with the 4-4 just to be consistent with what is right to do at the 1-level and for the same reasons.

And because of the above, I think if our Majors were reversed the 3H overcall rather than the double would be a lot more attractive.

[jdeegan]

rhm, on 2013-September-14, 08:40, said:

I wish I had your certainty what is right. I am not convinced at all.
Even if spades is your most likely successful strain, it does not follow that 3♠ is the right bid, mainly because this bid puts most of your eggs in one basket.
It is obvious that after 3♠ you can not play any more 3♣ doubled, 3 ♦ or 3♥ and you are unlikely to reach game in a red suit if that is right.
Is spades so much more likely the right strain that we should ignore all those alternatives combined?
If West has four or more spades we will likely reach spades anyway after DBL, that is DBL does not preclude playing ♠, if our spade fit is good, but overcalling 3♠ will preclude many other contracts.
If West has less than three cards in spades it is very likely we belong somewhere else, including 3♣ doubled or a red suit. Neither do I relish the thought that a vulnerable three level overcall might get raised on two cards,
Overcalling 3♠ is probably best if West has precisely 3 spades with less than 5 cards in hearts.
I do not need Pavlicek to know that partner is much more likely not to hold this precise combination in the majors
Also when we have only a 5-3 fit in spades, game in spades might not be so hot when partner is a passed hand and a preempt makes bad breaks more likely than usual.
It is not uncommon that 3NT plays better in such scenarios.
If partner is weak we might belong in a red suit even though he has three spades.
It is true that since LHO is passed the risk that we run into a penalty DBL diminishes, but a 3♠ overcall on a five card suit is still more dangerous than DBL.

Rainer Herrmann

All of your arguments are reasonable. At issue is whether bidding 3♠ is better percentagewise than double. Passing or doubling and then bidding 3♠ are both out of the picture according to ordinary bridge logic (not to say that either one or the other might not work best on any given hand). Most of the decision depends on the distribution of the various suits. We know for sure the expectancies. We know that the probability distributions around these expected values follow a binomial distribution, meaning they are roughly symmetrical in each case.

I was hoping someone listening might have Pavlicek's hand simulator up and running. I don't anymore. In fact, from where I am now, calculating all one gazillion of the binomial distribution outcomes might actually be easier for me than reviving Pavlicek. My point is that there is a clear probabilistic solution that resolves the issue of which is the better bid. Double gets killed (most of the time) when we play in a 4-3 red suit rather than a 5-3+ spade fit. It wins (on balance) when partner passes (rare). It wins when partner has 5 or 6 hearts. It breaks even (more or less) when partner has four or more spades and not four hearts or diamonds.

One last thing by far more important than all the probability B.S. JLOGIC bid 3♠.

[rhm]

jdeegan, on 2013-September-14, 16:44, said:

One last thing by far more important than all the probability B.S. JLOGIC bid 3♠.

I value JLOGIC views, but I have my own head and I dislike idolization.
Even top players have been known to disagree on bridge issues.

Rainer Herrmann