[whereagles]

han, on May 27 2008, 05:50 PM, said:

You got at least one too many!

oh.. now i see. it's a typo

[Echognome]

Alright. I'll play straw man. It just seems to me that double is flexible allowing partner to choose to bid or defend. Note that 2♦X isn't game, so that's not the end of the world. Also if partner returns to 2♥ we can't be unhappy.

Is the reason we don't like double:

1) Double is takeout, so partner will not be able to make the right decision based on our hand with defensive values.

2) Doubling will often endplay partner and we may end up in a bid fit in 3♣ going for a number.

3) Doubling gives the opponents a fielder's choice.

4) Some combination of above.

5) Some other reason that escapes gnome's grasp. (please explain)

[han]

If you double and partner passes 2DX then you'd be thrilled Matt. Thrilled.

[rogerclee]

Echognome, on May 27 2008, 11:03 AM, said:

Alright.  I'll play straw man.  It just seems to me that double is flexible allowing partner to choose to bid or defend.  Note that 2♦X isn't game, so that's not the end of the world.  Also if partner returns to 2♥ we can't be unhappy.

Is the reason we don't like double:

1) Double is takeout, so partner will not be able to make the right decision based on our hand with defensive values.

2) Doubling will often endplay partner and we may end up in a bid fit in 3♣ going for a number.

3) Doubling gives the opponents a fielder's choice.

4) Some combination of above.

5) Some other reason that escapes gnome's grasp. (please explain)

If partner bids 3♣ we are unhappy (unless partner actually has a good club suit, in which case he may balance with it anyway). Partner will never pass or bid 2NT. If partner bids 2♥, we have gotten basically nowhere. Not to mention that if partner does bid 3♣, and it is a good contract in a vacuum, it may start with three rounds of spades by the defense.

I think it is putting your neck with very few ways to win and a lot of ways to lose.

[rogerclee]

Jlall, on May 27 2008, 09:24 AM, said:

What did the panel do?

I think 2NT 100, Pass 90, 2♥ 80, Dbl 10. Kleinman is the director.

[Echognome]

Jlall, on May 27 2008, 10:16 AM, said:

X should have better clubs. You don't want partner to bid 3C over your X. Wouldn't you also X if your minors were reversed? To me those are 2 very different hands.

Ok. Understand it now. So basically double is more "takeout oriented" rather than just generic "I'm stronger than I have initially shown with 1♥ what do you think we should do about it."

I mean that's the context I was wondering.

To Roger's point, obviously partner's bid of 3♣ making us happy or unhappy depends entirely on how partner understands our double doesn't it?

[mike777]

Jlall, on May 27 2008, 11:24 AM, said:

655321, on May 27 2008, 01:35 AM, said:

Pass.

I know I would 'pass wtp' if this hand came up at the table, so it was interesting to see most of the panel bidding. (But, apparently I have learned nothing because I still pass  )

Yeah I would WTP pass as well, where are we going on this hand?

What did the panel do?

I will give forum posters another one or two days to post comments before I compare forum votes. I can give everyone a hint, majority did not pass. Great comments so far ty all.

[Halo]

2H

I've got a good hand and good hearts. This bid doesn't take us very high and isn't confusing.

Passing an nf 2D doesn't look like IMPs to me.

[MFA]

I like 2NT.
Sure, if the 1♠ overcall promises a near opener, and the 2♦ advance promises a good hand, we might be able to indulge ourselves and pass.
But I don't trust them that much, in real life we might have game.
We have tricks and stoppers, and this 2NT could easily be a good competitive move by itself, besides getting to some 3NT's.
-1100? No, they won't double very often even when we are potentially in big trouble.

2♥ is not so dangerous, but we don't get to any games and NT might very well play better if they have ruffs coming.

[NickRW]

According to the simulation I just ran
2N makes 8 or more tricks for South only 23% of the time and is 2 off or worse 60% of the time.
2H makes 8 or more tricks for South 46% of the time (and is 2 off or worse 19% of the time).
2D makes 8 or more tricks for East 67% of the time.

From that I'd say 2N is an awful bid. The decision between pass and 2H is closer. At matchpoints the case for 2H is clear enough I think - 46% chance of a plus versus 33, not to mention the fact that opps may bid on where they are going to be more likely to go off. However this is IMP. I calculate that by passing you rate for an average score -25.1 on this board assuming no further bids, by bidding 2H you rate to be -21.

Trying to analyse what use double is as a call is next to impossible by this means - it depends on what partner thinks it means and how that fits with the particular hand held. However, if North thinks it means club tolerance, analysing cases where North holds at least 5 clubs and at most 1 heart shows that 3C makes 9 or more tricks 21% of the time and goes 2 or more off 56% of the time. Which is about as bad as bidding 2N.

And if partner thinks double means values, 2D rates to make without further looking at the North hand, so that may not be all too happy an outcome either.

Thus I think 2H is better than pass - just. 2N and double are poor.

Nick

[655321]

NickRW, on May 27 2008, 07:55 PM, said:

According to the simulation I just ran
2N makes 8 or more tricks for South only 23% of the time and is 2 off or worse 60% of the time.
2H makes 8 or more tricks for South 46% of the time (and is 2 off or worse 19% of the time).
2D makes 8 or more tricks for East 67% of the time.

As Hrothgar has responded in the past... this sort of post without more information is meaningless. What were the conditions you put on the 1♠ overcall, partner's pass, RHO's 2♦?

[Echognome]

NickRW, on May 27 2008, 04:55 PM, said:

According to the simulation I just ran
2N makes 8 or more tricks for South only 23% of the time and is 2 off or worse 60% of the time.
2H makes 8 or more tricks for South 46% of the time (and is 2 off or worse 19% of the time).
2D makes 8 or more tricks for East 67% of the time.

From that I'd say 2N is an awful bid. The decision between pass and 2H is closer. At matchpoints the case for 2H is clear enough I think - 46% chance of a plus versus 33, not to mention the fact that opps may bid on where they are going to be more likely to go off. However this is IMP. I calculate that by passing you rate for an average score -25.1 on this board assuming no further bids, by bidding 2H you rate to be -21.

Trying to analyse what use double is as a call is next to impossible by this means - it depends on what partner thinks it means and how that fits with the particular hand held. However, if North thinks it means club tolerance, analysing cases where North holds at least 5 clubs and at most 1 heart shows that 3C makes 9 or more tricks 21% of the time and goes 2 or more off 56% of the time. Which is about as bad as bidding 2N.

And if partner thinks double means values, 2D rates to make without further looking at the North hand, so that may not be all too happy an outcome either.

Thus I think 2H is better than pass - just. 2N and double are poor.

Nick

Obviously if double means values, then partner can choose to sit, bid 2♥, or bid 3♣. If double is more takeout oriented, then it is a misbid. That was my question. How you can from "...it depends on what partner thinks it means" (which I find fair) to "2N and double are poor," seems a strange leap to me. Supposing that partner knows what double means, he has more options than 2♥ which is a unilateral decision. So nearly ipso facto it's better if it's not a misbid. Agree?

[NickRW]

655321, on May 28 2008, 01:10 AM, said:

As Hrothgar has responded in the past... this sort of post without more information is meaningless.  What were the conditions you put on the 1♠ overcall, partner's pass, RHO's 2♦?

Well, the extra info that I didn't give is the assumptions I made for the dealer program.

H S:KJ2.KQJT3.AKJ.T4
WS>=4 & WP+WS>=15
ED>=5 & EP+ED>=12 & NH+NP<9
ES<4 & ED>EH & ED>EC & WS>WH & WS>=WD & WS>=WC & WC+WD<8

South has the quoted hand only
West has at least 4 spades and hcp+spades >= 15
East has 5 diamonds and hcp + diamonds >= 12
North hearts + hcp < 9 (North was silent after all)
East spades < 4 (didn't raise overcaller)
East diamonds longer than hearts or clubs
And the last bit should mean that West didn't overcall with long hearts or fail to make a takeout double on a 4144 shape or bid spades when he/she had a longer minor.

Seemed reasonable to me, perhaps not to you. If you don't like the assumptions do your own simulation, or go by gut instinct if you think you have more experience than a 100 quite specific example deals.

As for Hrothgar's opinions, I have huge respect for those posts. However, I have probably learnt more from doing simulations than I have from reading books that quote "holding this, do that" based on what is usually no evidence at all other than the writer's experience and beliefs.

Zar Petkov wrote:

"I have virtually all books on bidding (and have actually read them) and I can say that almost without exception all of them revolve around HEAVY sets of examples to illustrate the point made in one or another aspect of the bidding process, basically saying: “With this hand you bid this, with that hand you bid that. Get it?”
No, I don’t get it. Chances for me to hold this or that hand in my lifetime are ... very slim, to put it mildly. And if it is some other hand, then it will be something else that matters, so I can never “catch up”."

Which is roughly my thoughts on the matter too.

Nick

[NickRW]

Echognome, on May 28 2008, 01:13 AM, said:

Obviously if double means values, then partner can choose to sit, bid 2♥, or bid 3♣. If double is more takeout oriented, then it is a misbid. That was my question. How you can from "...it depends on what partner thinks it means" (which I find fair) to "2N and double are poor," seems a strange leap to me. Supposing that partner knows what double means, he has more options than 2♥ which is a unilateral decision. So nearly ipso facto it's better if it's not a misbid. Agree?

Some of what you say is certainly fair. If double would be interpretted by partner as takeout, then clearly it is a misbid and the quite likely 3C contract looks to be poor.

If double is values, then double would be better on this particular hand as you clearly have values. What I am querying is just what sort of diamond holding you expect partner to have to leave the double in. He isn't likely to have that many, nor any much quality either. But nor did partner raise your opening either - so what sort of bid do you really expect partner to make? Perhaps I am confessing to no real experience of playing double as values in this situation....

However, though I analysed some cases where North might remove a take out double type of double, I admit I was rather dismissive of looking at cases where North might leave the double in. So I admit I short changed you there.

Mind you, other people can do these analyses as well as me - and some clearly don't really appreciate them - so frankly I don't feel like doing anymore tonight.

Nick

[655321]

NickRW, on May 27 2008, 08:45 PM, said:

655321, on May 28 2008, 01:10 AM, said:

As Hrothgar has responded in the past... this sort of post without more information is meaningless.  What were the conditions you put on the 1♠ overcall, partner's pass, RHO's 2♦?

Well, the extra info that I didn't give is the assumptions I made for the dealer program.

H S:KJ2.KQJT3.AKJ.T4
WS>=4 & WP+WS>=15
ED>=5 & EP+ED>=12 & NH+NP<9
ES<4 & ED>EH & ED>EC & WS>WH & WS>=WD & WS>=WC & WC+WD<8

South has the quoted hand only
West has at least 4 spades and hcp+spades >= 15
East has 5 diamonds and hcp + diamonds >= 12
North hearts + hcp < 9 (North was silent after all)
East spades < 4 (didn't raise overcaller)
East diamonds longer than hearts or clubs
And the last bit should mean that West didn't overcall with long hearts or fail to make a takeout double on a 4144 shape or bid spades when he/she had a longer minor.

Seemed reasonable to me, perhaps not to you. If you don't like the assumptions do your own simulation, or go by gut instinct if you think you have more experience than a 100 quite specific example deals.

As for Hrothgar's opinions, I have huge respect for those posts. However, I have probably learnt more from doing simulations than I have from reading books that quote "holding this, do that" based on what is usually no evidence at all other than the writer's experience and beliefs.

Zar Petkov wrote:

"I have virtually all books on bidding (and have actually read them) and I can say that almost without exception all of them revolve around HEAVY sets of examples to illustrate the point made in one or another aspect of the bidding process, basically saying: “With this hand you bid this, with that hand you bid that. Get it?”
No, I don’t get it. Chances for me to hold this or that hand in my lifetime are ... very slim, to put it mildly. And if it is some other hand, then it will be something else that matters, so I can never “catch up”."

Which is roughly my thoughts on the matter too.

Nick

Hmm, by the tone of your response it seems you have taken some offence. Not sure why.

I have no problem with simulations - they are a fine idea. The only point I was trying to make is that often posters glibly report that such and such an action is correct because they have 'done a simulation'. And maybe they are right. Alternatively, maybe the assumptions they have coded are such as to produce an incorrect simulation.

See, for example, this thread.

OK, thank you for posting your code. Yes, I would make some different assumptions.

I think East, missing the AKJ of diamonds, is likely to have at least 6. Also, I would be very confident East has fewer than 3 spades.
I expect North to have fewer than 4 hearts. Also North, with 3244, or 31(45), might make a negative double with 6/7 HCP, (but these hands will be very rare on the bidding anyway)
West could easily have 5 spades and fewer than 10 HCP to overcall. Also, some spade hands that would overcall 2, 3 or 4♠ are probably caught under your definition.

How much difference, if any, this would make to the results I don't know.

[NickRW]

655321, on May 28 2008, 02:20 AM, said:

Hmm, by the tone of your response it seems you have taken some offence.  Not sure why.

Because you called my contribution to the thread "meaningless".

Quote

I have no problem with simulations - they are a fine idea.  The only point I was trying to make is that often posters glibly report that such and such an action is correct because they have 'done a simulation'.  And maybe they are right.  Alternatively, maybe the assumptions they have coded are such as to produce an incorrect simulation.

See, for example, this thread.

OK, thank you for posting your code.  Yes, I would make some different assumptions.

I think East, missing the AKJ of diamonds, is likely to have at least 6.  Also, I would be very confident East has fewer than 3 spades.
I expect North to have fewer than 4 hearts.  Also North, with 3244, or 31(45), might make a negative double with 6/7 HCP, (but these hands will be very rare on the bidding anyway)
West could easily have 5 spades and fewer than 10 HCP to overcall.  Also, some spade hands that would overcall 2, 3 or 4♠ are probably caught under your definition.

How much difference, if any, this would make to the results I don't know.

I agree I made assumptions and others would not make exactly the same ones - which is kind of the same situation one actually faces at the table - just exactly how does this opponent bid? Stating things on a convention card only covers part of the story.

For example, I'll sometimes overcall at the one level on a four card suit. For some this is a no-no. And yes, East is quite likely to have 6 diamonds and is more likely to have less than 3 spades rather than exactly 3 and so on - depending in part of his/her partner's propensity for overcalling at the one level with 4 card suits or not. Also vulnerability and so on plays a part. For me, how many of the opps bids I take away with my bid plays a part - perhaps not so much for others.

Never the less, I find actually looking at real cases instructive. And looking at a largish number of real cases which you can do with a dealer program, a DD analyser and a spreadsheet really very instructive and, within the boundaries of the cases you looked at, of more value than one's own or anyone else's opinions, no matter how expert.

A keen bridge player plays say 4 nights a week, during which they play about a 100 hands. In the course of a year that amounts to 5000 boards. My laptop can do that overnight while I sleep and, furthermore, will have played all those hands without error 20 times each - once for each strain and for all 4 seats. Human experience can't match that.

Nick

[han]

The point that made by 655321 was that others cannot benefit much from the results of a simulation without knowing the conditions. Certainly most of us would like to look at those critically to see if they correspond with what we think is reasonable.

I agree with you that simulations can really add something to our understanding of these problems. However, what you posted initially, the double dummy percentages, is not the most useful information.

[gnasher]

NickRW, on May 28 2008, 02:45 AM, said:

H S:KJ2.KQJT3.AKJ.T4
WS>=4 & WP+WS>=15
ED>=5 & EP+ED>=12 & NH+NP<9
ES<4 & ED>EH & ED>EC & WS>WH & WS>=WD & WS>=WC & WC+WD<8

One of the things one should always do as part of this sort of simulation is to examine a sample of the hands generated and confirm that they match the real-life conditions. I reproduced your conditions using Thomas Andrew's excellent Deal program, and looked at the first 20 of the resulting deals. I noticed these problems [edit: or, at least, I think they're problems]:

Deal 5: East has Q3 42 987643 A75
Deal 6: East has Q A82 T8432 QJ76
Deal 7: East has --- A52 986543 Q853
Deal 8: East has 3 A54 Q8764 Q987
Deal 11: North has A 9752 T65 98765; East has 965 8 Q9432 AK32
Deal 12: North has 43 A752 987 9632; East has 976 84 QT653 AJ7
Deal 13: East has Q85 A64 98432 K7
Deal 14: North has 873 864 63 AJ763
Deal 15: East has T6 865 Q8653 AQ5
Deal 16: East has T6 A42 T9763 K75
Deal 17: West has A765 A97 T54 K72; East has 843 2 Q9762 AQJ6
Deal 19: North has 74 87642 8 K9732

If I were doing the simulation, these would persuade me to revisit the criteria.


The code I used was:

south is "KJ2 KQJT3 AKJ T4"

main {
if {[spades west] < 4 ||
[spades west] + [hcp west] < 15 ||
[diamonds east] < 5 ||
[diamonds east] + [hcp east] < 12 ||
[hearts north ] + [hcp north] >= 9 ||
[spades east] >= 4 ||
[diamonds east] <= [hearts east] ||
[diamonds east] <= [clubs east] ||
[spades west] <= [hearts west] ||
[spades west] < [diamonds west] ||
[spades west] < [clubs west] ||
[clubs west] + [diamonds west] >= 8
} {reject}
accept
}

This post has been edited by gnasher: 2008-May-28, 03:39

[NickRW]

gnasher, on May 28 2008, 08:36 AM, said:

One of the things one should always do as part of this sort of simulation is to examine a sample of the hands generated and confirm that they match the real-life conditions.

I do do that. I have a utility that will reformat the file in html so it is easy to see what rubbish you've generated. This is useful. In my first attempt I placed no restriction on the North hand and a lot of them were coming out with things that most people would have raised in competition at the very least.

However, it is also, to some extent a thankless task - one starts to place more and more restrictions on what some unknown opponent can and can't hold, when real life experience shows that actual opponents and even partners do things that you wouldn't have (or fail to do things you would).

If one's criteria are at all reasonable (and you might debate to what extent mine were, but it is a judgement call), one also has to realise that quite a lot of hands are reasonable and the "unreasonable" only a percentage of the whole - which is the sort of thing one actually encounters at the table when playing against people.

Quote

I reproduced your conditions using Thomas Andrew's excellent Deal program,

I'm currently using the freebie download dealer that is on Richard Pavlicek's site - it is moderately sophisticated, but really tedious to make truly complex filtering criteria. Perhaps I'll look at alternatives.

Quote

Deal 12: North has 43 A752 987 9632

This is one difference between Brits and Americans. A lot of people on the west side of the pond, I think, would raise with this. SAYC, with only 2C and 2N as strong bids, has a quite a high ceiling (understatement!) on the upper range for a 1 bid - which means that some 4 and 5 counts look attractive to respond with. Plus, playing 5 card majors means that you have an extra trump in supprt. Playing Acol, you only have normal support and partner isn't going to be holding a 21 or so count, so less reason to respond with 4, 5 and even some 10 loser 6 counts would hardly be worth bothering with.

But this is where one starts to impose what may (or may not be) unreasonable restrictions on the output of the dealer program.

Quote

Deal 5: East has Q3 42 987643 A75

Well, yeah - perhaps many would have said nothing with this. Or maybe they shouldn't???? If East can make 2♦ 2/3rds of the time when chosing to respond with an ace, a queen, 6 diamond shaped things and 5 other assorted bus tickets perhaps East knows something....

One has to understand that placing further restriction on the East hand, either a point more and/or one more diamond - which will make East more likely to respond in some people's eyes is, of course, only likely to raise the % of the time when 2♦ will be a make.

Anway, if I'm going to get the third degree doing unpaid analyses for you lot, frankly I don't feel much disposed to sharing information with you - you clearly don't want it.

Nick