ATB

[RSClyde]

Here are some possible explanations:

 

- If we have a doubleton, the doubleton suit will tend to be the best lead for the defence. If they lead their longest suit, they will tend to find the best lead.

If we have a 4333 shape, the best lead will be more dependent on honour holdings. If they lead their longest suit, it's less likely that this will be their best lead.

 

- When you're 4333 rather than 4432, you will have more two-way finesse positions. For example, compare A10 opposite KJxx with A10x opposite KJx. Both are worth three tricks double-dummy, but single-dummy the second one has a guess and the first does not.

 

- When you're 33 in two suits rather than 42, you're more likely to have a guess as to which one to play on.

Thank you for the productive dialog.

[rhm]

Here are some possible explanations:

 

- If we have a doubleton, the doubleton suit will tend to be the best lead for the defence. If they lead their longest suit, they will tend to find the best lead.

If we have a 4333 shape, the best lead will be more dependent on honour holdings. If they lead their longest suit, it's less likely that this will be their best lead.

 

- When you're 4333 rather than 4432, you will have more two-way finesse positions. For example, compare A10 opposite KJxx with A10x opposite KJx. Both are worth three tricks double-dummy, but single-dummy the second one has a guess and the first does not.

 

- When you're 33 in two suits rather than 42, you're more likely to have a guess as to which one to play on.

This is often said, but I am not sure it is very relevant except at the slam level, where declarer typically has more guesses than the defense.

The standard example given is always the two way finesse. But it seems to be finely balanced by options the defense gets right double dummy, for example opening leads.

Why everybody only concentrates on declarer play as an argument that double dummy results are skewed escapes me.

Fact is double dummy derives it tricks differently from single dummy logic, but it does not seem to favor declarer over defenders in general except what is known already, e.g declarer makes on average slightly more tricks in 3NT single dummy than double dummy, because the opening lead tends to be more crucial when there is no trump suit.

What double dummy analysis found out is that generally speaking distribution has on average only a small impact on trick taking capability when there is no trump suit, by which you can control your short suits.

Fit in long suits in notrump are nice but balanced by the fact that you must have short suits then as well.

 

But distribution does matter a lot and double dummy results agree, provided there is a trump suit for control.

 

Rainer Herrmann

[gnasher]

Fact is double dummy derives it tricks differently from single dummy logic, but it does not seem to favor declarer over defenders in general except what is known already

We're not talking about whether it favours anyone in general. We're talking about whether it favours anyone more in one specific category of hand than it does in a more general category of hand.

 

That is, we're comparing

(Single dummy tricks when 4333) - (Double dummy tricks when 4333)

with

(Single dummy tricks when balanced) - (Double dummy tricks when balanced)

Edited September 10, 2013 by gnasher

[mcphee]

Pairs and N elects to open, why is not clear to me but so it is. S apparently played a few hands with this partner before and ONLY tried for game with a reasonable opening bid, with a 3 card D fit to boot. To be fair it is not often that I find partner opening the bidding and I need strong NT values in order to have any positive feelings about sniffing for a game.

 

In pairs the idea is to be conservative, and the south hand was conservative, north was not. Even playing a weak NT the N hand is somewhat suspect holding no aces to open. Taking a second free call of 1NT was serving an active imagination.

 

N gets it all, from my view.

[JLOGIC]

Assuming they open and freely bid 1N with this kind of hand, then south gets all the blame since he didn't pass 2H which showed a minimum (presumably north has any 11-12 with 3 hearts so what is the point of bidding 3H?). Most likely south didn't get the memo that this is the style, in precision I open the north hand but I don't freely bid 1N. If south thought north had shown something more than a minimum to bid 1N freely, then 3H is certainly fine even after partner shows a minimum in that context...in fact I'd just drive it to game.

 

So I suspect the problem is with the expectations of the free 1N bid.

[beatrix45]

What is the world coming to? Some 80 postings over what may rank as, possibly, the most ordinary and pedestrian bridge hand in the long and storied history of the game. Theories abound! Computer simulations run rampant! Well, maybe not actually "rampant", but verging on it. Seriously, get in touch with your feelings. Play a little money bridge.

[rhm]

What is the world coming to? Some 80 postings over what may rank as, possibly, the most ordinary and pedestrian bridge hand in the long and storied history of the game. Theories abound! Computer simulations run rampant! Well, maybe not actually "rampant", but verging on it. Seriously, get in touch with your feelings. Play a little money bridge.

I happen to be interested in ordinary hands, because they crop up every day and how you handle them has a lot to do whether you win at money bridge or any other form of scoring.

The ones you seem to prefer, which crop up every other leap year, I do not care, not even for money bridge.

 

Rainer Herrmann

[gwnn]

- If we have a doubleton, the doubleton suit will tend to be the best lead for the defence. If they lead their longest suit, they will tend to find the best lead.

If we have a 4333 shape, the best lead will be more dependent on honour holdings. If they lead their longest suit, it's less likely that this will be their best lead.

Isn't this saying that defenders are more likely to find a good lead in double-dummy play than in single-dummy play? So this factor will make 4333 take more tricks single-dummy than double-dummy. I also note that cherdano made a similar point but with the opposite conclusion upthread (i.e. double dummy, defenders have a better chance of finding your doubleton) so I am confused.

[gnasher]

Isn't this saying that defenders are more likely to find a good lead in double-dummy play than in single-dummy play? So this factor will make 4333 take more tricks single-dummy than double-dummy.

That's true. Would you find it convincing if I now dismissed this factor as insignificant?

 

I also note that cherdano made a similar point but with the opposite conclusion upthread (i.e. double dummy, defenders have a better chance of finding your doubleton)

Yes, so we have two conflicting effects:

- Against 4333, a single-dummy lead is less likely to hit declarer's weakest suit.

- Against 4333, missing declarer's weakest suit is less likely to matter.

 

so I am confused.

Me too. These thought experiments never seem to get us very far. Maybe we can get Phil King to do some analysis instead?

[helene_t]

It wouldn't surprise me too much if 4333 turned out to be slightly better than 4432 for notrump purposes. Short suits suck at NT. If partner has a 5-card suit I strongly prefer to have 3 cards in it. 3 may be better than 2 regardless of partner's length, but if partner has 4 in the suit it makes almost no difference if I have 3 or 4 in that suit.

 

If the reason for downgrading 4333 is that we don't always end up in NT, (32)44 should be downgraded as well. So maybe conventional wisdom really is flawed. Interesting.

 

Edit: it may also depend on the strength of the two hands. If I have 18-19 and p has 6-7 then maybe 4333 is better since with a doubleton opposite partner's long suit I may not have entries to cash his long suit after it is established, while a doubleton opposite partner three-card suit may be vulnerable as partner often doesn't have a stopper in that suit. OTOH with 12 HCPs in my hand and 12 HCPs in dummy I probably would prefer 4432.

[gwnn]

Me too. These thought experiments never seem to get us very far. Maybe we can get Phil King to do some analysis instead?

Perhaps we should wait until third-world countries get better at bridge, then we could get large-scale, reliable single-dummy simulations at a reasonable price (mods: feel free to move this to the Water Cooler!).

[rhm]

Perhaps we should wait until third-world countries get better at bridge, then we could get large-scale, reliable single-dummy simulations at a reasonable price (mods: feel free to move this to the Water Cooler!).

Yes indeed why accept: "Fact is double dummy derives it tricks differently from single dummy logic, but it does not seem to favor declarer over defenders"

 

when we can stay confused instead and nurture our prejudice.

Let us wait for some obscure evidence, which will not arise in the near future.

 

Rainer Herrmann

[CSGibson]

It wouldn't surprise me too much if 4333 turned out to be slightly better than 4432 for notrump purposes. Short suits suck at NT. If partner has a 5-card suit I strongly prefer to have 3 cards in it. 3 may be better than 2 regardless of partner's length, but if partner has 4 in the suit it makes almost no difference if I have 3 or 4 in that suit.

 

If the reason for downgrading 4333 is that we don't always end up in NT, (32)44 should be downgraded as well. So maybe conventional wisdom really is flawed. Interesting.

 

 

It would surprise the heck out of me. 4+ card suits have the ability to take tricks without using extra HCPs - ie, the length makes them a threat, not just the honor cards, though, of course, you need honor cards in the suits to win a tempo war. They also restrict defender's choices, and sometimes become essential in the endgame via squeezes and throw-ins, much more so than 3 card suits.

[gnasher]

Rainer, if it's "fact", I assume you have some evidence to offer? Can you show us evidence that the difference between double-dummy and single-dummy results is the same when declarer is 4333 as when he is 4432?

[CSGibson]

Yes indeed why accept: "Fact is double dummy derives it tricks differently from single dummy logic, but it does not seem to favor declarer over defenders"

 

when we can stay confused instead and nurture our prejudice.

Let us wait for some obscure evidence, which will not arise in the near future.

 

Rainer Herrmann

 

 

Maybe not overall, but when declarer has more guesses than defense, then double-dummy will favor declarer, and vice versa. when declarer has KJ combinations, double dummy will more likely favor declarer than single dummy play. When declarer has his choices restricted, double dummy play will probably favor the defense more. It will tend to average out, but it does not mean that it cannot favor declarer over defenders for specific hand types.

[rhm]

Rainer, if it's "fact", I assume you have some evidence to offer? Can you show us evidence that the difference between double-dummy and single-dummy results is the same when declarer is 4333 as when he is 4432?

No I can not.

The statement only claimed as fact that double dummy derives it tricks differently from single dummy. (who doubts that?)

I have not seen comparisons of double dummy and single dummy results grouped according to distribution, only grouped according to contract level and whether the contract was suit or notrump.

But I admit I personally would be quite surprised if somebody would show that the difference in trick outcome is significantly affected by distribution, except possibly for freak distributions.

 

Rainer Herrmann

[rhm]

Maybe not overall, but when declarer has more guesses than defense, then double-dummy will favor declarer, and vice versa. when declarer has KJ combinations, double dummy will more likely favor declarer than single dummy play. When declarer has his choices restricted, double dummy play will probably favor the defense more. It will tend to average out, but it does not mean that it cannot favor declarer over defenders for specific hand types.

True, that's why low level notrump contracts favor slightly declarer and slams favor slightly the defense compared to double dummy play.

If we were discussing slam potential of this hand you might convince me.

Overall the difference is remarkably small. What is easy, is to construct carefully complete deals where this is not so. They are not representative.

What I am less convinced of is that you can predict from a single hand that declarer will have many more guesses to take than the defense just because the hand under consideration has broken honor combinations and/or is 4333.

 

Rainer Herrmann

[gwnn]

rhm it's fine to believe "I trust that SD is not far from DD," but we were just wondering whether this difference relatively bigger or smaller in case of 4432 or 4333. These subtle effects were what I was confused about. Please don't just quote me out of context and make it sound like I'm/we're confused about everything. Thanks!

 

To clarify: do you think that based on the numbers we were presented with Kxxx KJx KJx xxx is indeed an average 11 count for notrump purposes and people who disagree are only doing so because they are nurturing their prejudices?

[gnasher]

No I can not.

The statement only claimed as fact that double dummy derives it tricks differently from single dummy. (who doubts that?)

I have not seen comparisons of double dummy and single dummy results grouped according to distribution, only grouped according to contract level and whether the contract was suit or notrump.

But I admit I personally would be quite surprised if somebody would show that the difference in trick outcome is significantly affected by distribution, except possibly for freak distributions.

I see. So the "fact" that you asserted is irrelevant to the matter we're discussing. And when you suggested that some of us were basing our opinions on prejudice, that wasn't based on any evidence, but merely on your own prejudice.

[cherdano]

Why are we just discussing 4333 vs other shapes? The hand in question is Kxxx KJx KJx xxx. Lots of potential for KJ guesses or two-way finesses. Lots of potential for tricky guesses on which suit to develop first. Rainer would say that there are also a lot of opportunities to make the wrong lead, or to guess wrong which suit to set up on defense.

Sounds like another argument about judgement for which there is no hard evidence available either way.

[rhm]

To clarify: do you think that based on the numbers we were presented with Kxxx KJx KJx xxx is indeed an average 11 count for notrump purposes and people who disagree are only doing so because they are nurturing their prejudices?

In my estimate Kxxx KJx KJx xxx is an average 11 count for low level notrump contracts up to and including 3NT. (The jacks look useful to me)

However, I also think if you end up in a trump contract, this hand is on average worth less, but particularly at game or higher levels.

When you hold a distributional hand with cards in the majors it is likely that you will end up in a suit contract.

But when you open a balanced hand you do not know where you will end up. So a compromise seems prudent.

 

Rainer Herrmann

[RSClyde]

Why are we just discussing 4333 vs other shapes? The hand in question is Kxxx KJx KJx xxx. Lots of potential for KJ guesses or two-way finesses. Lots of potential for tricky guesses on which suit to develop first. Rainer would say that there are also a lot of opportunities to make the wrong lead, or to guess wrong which suit to set up on defense.

Sounds like another argument about judgement for which there is no hard evidence available either way.

Because the simulation was repeated for a general 4333 and the results were similar: 61%

[gwnn]

Yes, that is consistent with the scenario that Kxxx KJx KJx xxx is weaker than an average 11-count but it is artificially made better by double-dummy artefacts.

[VM1973]

This is probably true playing vanilla 2/1 with a pick-up partner and no inferences. With a rich enough system, I think the value of describing your hand accurately can be more important than masterminding the auction before you have any idea what's happening. Yes, 1nt might not rate to be the best contract if it passes out, but 1♣ doesn't rate to be the right contract if this hand passes out: AKQx, AQxx, Kx, xxx.

Again I think you're missing the point.

 

The point is that you should make bids that have multiple ways of working out well and only one or no ways of working out badly. Imagine, for example, that you are fond of opening 2-2-4-5 with 1NT if it falls into your NT range, which we'll assume for now is 15-17. Obviously there is a way that this bid works out badly otherwise this would be common practice. That being the case, you occasionally do it, which I don't have a problem with. The point is, however, that if this bid has extra ways of going wrong then you should not take that action.

 

Holding:

 

♠K2

♥K2

♦KT32

♣AQJ32

 

a 1NT bid might work out quite well. On the other hand, holding:

 

♠32

♥K2

♦KQJT

♣AQJ32

 

you have a second flaw in that you don't have all suits stopped. Holding:

 

♠32

♥AQJT2

♦K2

♣KQJ2

 

you have a third flaw in the form of a 5-card heart suit. Holding:

 

♠KQJ2

♥AQJT2

♦K2

♣32

 

you have a 4th flaw. God willing no one would open this hand 1NT because it is just too flawed for that action.

Nor will any double-dummy Dealmaster simulations opposite 3-2-4-4 9 counts convince me that opening 1NT with the 4th hand is a good action.

Sorry.

[the hog]

It wouldn't surprise me too much if 4333 turned out to be slightly better than 4432 for notrump purposes. Short suits suck at NT. If partner has a 5-card suit I strongly prefer to have 3 cards in it. 3 may be better than 2 regardless of partner's length, but if partner has 4 in the suit it makes almost no difference if I have 3 or 4 in that suit.

 

If the reason for downgrading 4333 is that we don't always end up in NT, (32)44 should be downgraded as well. So maybe conventional wisdom really is flawed. Interesting.

 

Edit: it may also depend on the strength of the two hands. If I have 18-19 and p has 6-7 then maybe 4333 is better since with a doubleton opposite partner's long suit I may not have entries to cash his long suit after it is established, while a doubleton opposite partner three-card suit may be vulnerable as partner often doesn't have a stopper in that suit. OTOH with 12 HCPs in my hand and 12 HCPs in dummy I probably would prefer 4432.

 

It would certainly surprise me.