BBO Discussion Forums: Choose your poison - BBO Discussion Forums

Jump to content

  • 2 Pages +
  • 1
  • 2
  • You cannot start a new topic
  • You cannot reply to this topic

Choose your poison

#21 User is offline   nullve 

  • PipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 1,747
  • Joined: 2014-April-08
  • Gender:Male
  • Location:Norway
  • Interests:partscores

Posted 2021-May-12, 04:25

View Postmikeh, on 2021-May-11, 21:18, said:

Run a proper simulation, giving opener 7+ hearts...if 7 then not 7222 and always, if 7, very good hearts...at least AKJ10xxx and likely, then, either some 7330 or some 74 hand...even then there would be hands some would say fit and some would say don't. Make it too good, and one may have a 1H opener. Make it too weak and one might have a 3H opener.

Ok, here's another try:

Script:

Spoiler

One run:
avg. # of tricks if W is declarer in a D contract: 10.546
avg. # of tricks if E is declarer in a S contract: 10.311
Generated 437839 hands
Produced 1000 hands
Initial random seed 1620812489
Time needed   65.457 sec


View Postmikeh, on 2021-May-11, 21:18, said:

it is equally self-evident that xxxx is unlikely to make for a good contract...in spades or diamonds, although xxxx xx Qx AJ9xx makes a diamond contract pretty good...and I don't need a simulation to prove that, lol.

So, does 5 tend to be a better contract than 4 opposite xxxx (interpreted as 8xxx or worse)?

Slightly modified script:

Spoiler

One run:
avg. # of tricks if W is declarer in a D contract: 9.086
avg. # of tricks if E is declarer in a S contract: 8.1
Generated 6302538 hands
Produced 500 hands
Initial random seed 1620813618
Time needed   41.724 sec



View Postmikeh, on 2021-May-11, 21:18, said:

As I said before: if all I knew was that partner held 4+ spades, I'd bid x in a heartbeat. Lacking that information, I bid 5D

Fair enough.
0

#22 User is offline   mikeh 

  • PipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 11,446
  • Joined: 2005-June-15
  • Gender:Male
  • Location:Canada
  • Interests:Bridge, golf, wine (red), cooking, reading eclectically but insatiably, travelling, making bad posts.

Posted 2021-May-12, 05:38

View Postnullve, on 2021-May-12, 04:25, said:

Ok, here's another try:

Script:

Spoiler

One run:
avg. # of tricks if W is declarer in a D contract: 10.546
avg. # of tricks if E is declarer in a S contract: 10.311
Generated 437839 hands
Produced 1000 hands
Initial random seed 1620812489
Time needed   65.457 sec



So, does 5 tend to be a better contract than 4 opposite xxxx (interpreted as 8xxx or worse)?

Slightly modified script:

Spoiler

One run:
avg. # of tricks if W is declarer in a D contract: 9.086
avg. # of tricks if E is declarer in a S contract: 8.1
Generated 6302538 hands
Produced 500 hands
Initial random seed 1620813618
Time needed   41.724 sec




Fair enough.

One of us doesnít remotely understand the other. I point out that a simulation based on specifying that partner has 4 spades is of no relevance to our decision over 4H, since most of us donít know partnerís shape at that point, and your response is to run even more simulations specifying that partner has 4 spades. What, if anything, does that Ďproveí?

Plus I point out how running simulations, with no effort to look at the generated hands and try to estimate how the other players would, in real life, act is largely a waste of time.

Say you to run 1000 hands, randomizing partner and LHO.

Working out, say, that the partnership will make, on average, 8.3 tricks in spades ad 10.53 tricks in diamonds means nothing.

The number of tricks in spades is meaningless unless itís realistic to play in spades after double.

The number of tricks our way, in either strain, is meaningless if partner should be passing the double

The number of tricks our way is meaningless (in most cases) if LHO is raising hearts (only if partner has a hand on which he can bid over 5H is it relevant).

And so on.
'one of the great markers of the advance of human kindness is the howls you will hear from the Men of God' Johann Hari
0

#23 User is offline   lamford 

  • PipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 6,235
  • Joined: 2007-October-15

Posted 2021-May-12, 07:23

I agree with mikeh that double is poor. 5D for me, and second choice pass. If partner has his striped-tailed jacket on he will double 5H and pull the redouble to 6D.
I prefer to give the lawmakers credit for stating things for a reason - barmar
0

#24 User is offline   nige1 

  • 5-level belongs to me
  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 8,840
  • Joined: 2004-August-30
  • Gender:Male
  • Location:Glasgow Scotland
  • Interests:Poems Computers

Posted 2021-May-12, 07:51

Nullve's simulations seem to show that 4 is more likely to make than 5, whenever East has 4 or more spades (which is quite likely).
  • If you double, EW might still play in 5 but
  • If you bid 5, you can't get back to 4.
  • A drawback of double is that when partner has fewer than 4 s and chooses to pass -- 4X seems quite likely to make.
Admittedly, double-dummy analysis is unrealistic but less tedious and controversial than human analysis of 1000 deals. NS agreement about a 4 pre-empt might be a factor, especially if South's hearts need to be headed by top honours. In my circles, I'm unaware of any such requirements.

Below is a rerun of the simulation without restraints. Please note that 5 is still poor and 4 often makes.




predeal west SKT62, DAKJ9642, CQT
condition hearts(south) > 7 
or (hearts (south) == 7 and (clubs (south) > 3 or diamonds (south) > 3)) 
produce 1000
action 
frequency "W makes 5D" (tricks (west, diamonds) > 11, 0, 1),
frequency "E makes 4S" (tricks (east, spades) > 9, 0, 1),
frequency "S makes 4H" (tricks (south, hearts) > 9, 0, 1)

Frequency W makes 5D:
    0         827
    1         173
Frequency E makes 4S:
    0         513
    1         487
Frequency S makes 4H:
    0         367
    1         633
Generated 47501 hands
Produced 1000 hands
Initial random seed 1620829163
Time needed  120.043 sec

0

#25 User is offline   nullve 

  • PipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 1,747
  • Joined: 2014-April-08
  • Gender:Male
  • Location:Norway
  • Interests:partscores

Posted 2021-May-12, 08:01

View Postmikeh, on 2021-May-12, 05:38, said:

One of us doesnít remotely understand the other. I point out that a simulation based on specifying that partner has 4 spades is of no relevance to our decision over 4H, since most of us donít know partnerís shape at that point, and your response is to run even more simulations specifying that partner has 4 spades. What, if anything, does that Ďproveí?

Well, I thought you made such a simulation relevant with your first reply:

View Postmikeh, on 2021-May-09, 16:43, said:

I was trying to post earlier, but my phone died. Anyway, I think double of 4H is nuts, and not because Cyber was laying a trap, lol. We have negative defence and double gets passed far too often, and when it doesnít it often leads to a poor 4-4 spade fit. AKJxxxx spells trump. 5D is clear, imo

0

#26 User is offline   mikeh 

  • PipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 11,446
  • Joined: 2005-June-15
  • Gender:Male
  • Location:Canada
  • Interests:Bridge, golf, wine (red), cooking, reading eclectically but insatiably, travelling, making bad posts.

Posted 2021-May-12, 10:18

View Postnullve, on 2021-May-12, 08:01, said:

Well, I thought you made such a simulation relevant with your first reply:

I suggest, when carrying on a back and forth with another poster, that you pay attention to all of that posterís posts. Fixating on one early post is nuts. Perhaps the fact that I expressly stated that if I knew partner had 4+ spades, Iíd double ought to have had some impact on you....but obviously it didnít.

Nigeís last attempt is a little more persuasive than any of yours, but his constraints are, imo, poor. Assuming that opener will have any 7 card suit is as silly as your assuming that partner always has 4 spades, plus he still...like you...pays zero attention to what the generated hands look like, and whether they all fit the auction. Nor does he consider how other players will act.

I know...itís way too much work. But thatís the point, without that work, the simulation will be based on a number of hands where opener would bid 3H (or take some other call) and even if one weeded those out, we still have to contend with filtering out hands on which LHO would bid, which makes our declaring less likely.

And, of course, we have to evaluate more than how many tricks our side can make. We have to look at how often partner passes a double and whether that is usually good or bad compared to our 5D result.

We also have to consider which action leads to the best result when partner has a good hand...and some hands will seem good to partner opposite a double while others will seem good opposite 5D...and on all such good hands we have to decide which action partner will take opposite either double or 5D.

And I defy anyone to do all of this work objectively, given that we know what our hand looks like.

Finally, of course, even if one were prepared to do the amount of work required to render a simulation even remotely useful, the reality is that someone else, working with the same hands one had generated, would filter the results differently...and be just as Ďrightí as one could claim to be.

Iow, simulations here are of extremely limited value, and overly simplistic ones of none at all, imo. I must admit I laughed out loud when I read Nige conceding that his simulation is unrealistic but he still uses it. Not exactly a logical approach, lol.
'one of the great markers of the advance of human kindness is the howls you will hear from the Men of God' Johann Hari
0

#27 User is online   Winstonm 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 15,659
  • Joined: 2005-January-08
  • Gender:Male
  • Location:Tulsa, Oklahoma
  • Interests:Art, music

Posted 2021-May-12, 11:23

View Postmikeh, on 2021-May-12, 00:06, said:

If I had to pass at any form of scoring, it would be mps, but I'd not pass at any form of scoring.

Imps risks a double game swing. Picture partner with Qx xxx Qxxx KJxx. We're cold for 5D while they're cold for 4H if opener has 8 hearts...8 hearts and two black aces. Mps, it's only a board. Btw, if we double and catch partner with that hand, we're minus 790. Into our 600....obviously we can equally generate hands where double wins big, so I'm not pretending this is likely.

I'm not a fan of Bobby Wolff....despite his being a great player....but I do like a saying he used in the BW MSC, on numerous occasions....hands like these are 'too dangerous to pass'

Remember...LHO doesn't know your hand, especially when you bid 5D. For all he knows you have 8 diamonds. He's more likely to take the push to 5H over 5D than over double, especially if he's loaded in spades (when passing the double might give him a shot at 4S)

Admittedly that's a tiny, maybe non-existent, edge against good players who know that they bid immediately or not at all.


I played quite a few times in Dallas against Bobby and I can't remember ever getting a good result so I should probably listen to him - and you. Thanks for the response.
"Injustice anywhere is a threat to justice everywhere." Black Lives Matter.
0

#28 User is offline   nullve 

  • PipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 1,747
  • Joined: 2014-April-08
  • Gender:Male
  • Location:Norway
  • Interests:partscores

Posted 2021-May-12, 12:18

View Postnige1, on 2021-May-12, 07:51, said:

predeal west SKT62, DAKJ9642, CQT
condition hearts(south) > 7 
or (hearts (south) == 7 and (clubs (south) > 3 or diamonds (south) > 3)) 
produce 1000
action 
frequency "W makes 5D" (tricks (west, diamonds) > 11, 0, 1),
frequency "E makes 4S" (tricks (east, spades) > 9, 0, 1),
frequency "S makes 4H" (tricks (south, hearts) > 9, 0, 1)


Found an error: "> 11"

Rerun using "> 10" instead:

Frequency W makes 5D:
    0	     552
    1	     448
Frequency E makes 4S:
    0	     537
    1	     463
Frequency S makes 4H:
    0	     412
    1	     588
Generated 45475 hands
Produced 1000 hands
Initial random seed 1620844908
Time needed  110.964 sec

0

#29 User is offline   nige1 

  • 5-level belongs to me
  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 8,840
  • Joined: 2004-August-30
  • Gender:Male
  • Location:Glasgow Scotland
  • Interests:Poems Computers

Posted 2021-May-13, 02:32

View Postnullve, on 2021-May-12, 12:18, said:

Found an error: "> 11"

Rerun using "> 10" instead:

Frequency W makes 5D:
    0         552
    1         448
Frequency E makes 4S:
    0         537
    1         463
Frequency S makes 4H:
    0         412
    1         588
Generated 45475 hands
Produced 1000 hands
Initial random seed 1620844908
Time needed  110.964 sec


Thank you Nullve :) Yesterday wasn't a good day :( Apologies to anybody still reading this thread :(



0

#30 User is offline   nige1 

  • 5-level belongs to me
  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 8,840
  • Joined: 2004-August-30
  • Gender:Male
  • Location:Glasgow Scotland
  • Interests:Poems Computers

Posted 2021-May-13, 09:49

View Postmikeh, on 2021-May-12, 10:18, said:

I suggest, when carrying on a back and forth with another poster, that you pay attention to all of that poster's posts. Fixating on one early post is nuts. Perhaps the fact that I expressly stated that if I knew partner had 4+ spades, I'd double ought to have had some impact on you....but obviously it didn't. Nige's last attempt is a little more persuasive than any of yours, but his constraints are, imo, poor. Assuming that opener will have any 7 card suit is as silly as your assuming that partner always has 4 spades, plus he still...like you...pays zero attention to what the generated hands look like, and whether they all fit the auction. Nor does he consider how other players will act.
It's hard to guess how a random set of players will act. For example, on suit quality, many UK players open 3N or 4 with near solid , so a 4 opener tends to show a weaker suit. I'm an advocate of computer-simulations because I prefer to leave such debates to others.

View Postmikeh, on 2021-May-12, 10:18, said:

I know...it's way too much work. But that's the point, without that work, the simulation will be based on a number of hands where opener would bid 3H (or take some other call) and even if one weeded those out, we still have to contend with filtering out hands on which LHO would bid, which makes our declaring less likely. And, of course, we have to evaluate more than how many tricks our side can make. We have to look at how often partner passes a double and whether that is usually good or bad compared to our 5D result.
One of Terence Reese's mantras was "put your faith in the long suit". Here, IMO the best argument for overcalling 5 is that if you double and partner has fewer than 4, he might pass: and 4X is likely to make. Partner is likely to hold at least 4 s, however, and then 4 seems to have good prospects; better than 5, at double-dummy, anyway.

View Postmikeh, on 2021-May-12, 10:18, said:

We also have to consider which action leads to the best result when partner has a good hand...and some hands will seem good to partner opposite a double while others will seem good opposite 5D...and on all such good hands we have to decide which action partner will take opposite either double or 5D.And I defy anyone to do all of this work objectively, given that we know what our hand looks like.
Agree. Good luck with that.

View Postmikeh, on 2021-May-12, 10:18, said:

Finally, of course, even if one were prepared to do the amount of work required to render a simulation even remotely useful, the reality is that someone else, working with the same hands one had generated, would filter the results differently...and be just as 'right' as one could claim to beIow,
Agree; but, in the end, we must rely on our own judgement.

View Postmikeh, on 2021-May-12, 10:18, said:

Simulations here are of extremely limited value, and overly simplistic ones of none at all, imo.I must admit I laughed out loud when I read Nige conceding that his simulation is unrealistic but he still uses it. Not exactly a logical approach, lol.
:) Glad to entertain :) Computer simulations seem more objective and convincing than subjective evaluations based on gut feelings. Computer simulations require realistic assumptions; about which players can argue; although. anybody is free to code their own assumptions. Also, many players regard double-dummy analysis to be a flawed method of determining probable outcomes. Nevertheless, for example, David Bird and Taf Anthias have convinced many experts to revise their leading methods with their best-selling books
  • Winning Notrump Leads and
  • Winning Suit Contract Leads

0

#31 User is offline   nullve 

  • PipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 1,747
  • Joined: 2014-April-08
  • Gender:Male
  • Location:Norway
  • Interests:partscores

Posted 2021-May-13, 14:16

View Postnige1, on 2021-May-13, 09:49, said:

Here, IMO the best argument for overcalling 5 is that if you double and partner has fewer than 4, he might pass: and 4X is likely to make. Partner is likely to hold at least 4 s, however, and then 4 seems to have good prospects; better than 5, at double-dummy, anyway.

I'm afraid I would apply LoTT and pass the double also on the vast majority of hands with 4S3H.
0

#32 User is offline   nige1 

  • 5-level belongs to me
  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 8,840
  • Joined: 2004-August-30
  • Gender:Male
  • Location:Glasgow Scotland
  • Interests:Poems Computers

Posted 2021-May-14, 05:12

View Postnullve, on 2021-May-13, 14:16, said:

I'm afraid I would apply LoTT and pass the double also on the vast majority of hands with 4S3H.

IMO, as MikeH says, what you do, depends on judgement and player habits. I would tend to bid 4 because my partner's double indicates primary support unless he has a strong hand. The likely soundness of South's 4 pre-empt is also relevant. IMO, dealerGib simulations are quite revealing. Below is another variation, defining South's hand in terms of HCP and losers.
predeal west SKT62, DAKJ9642, CQT
condition hearts (south) > 6 and 17 > hcp (south) and 6 > loser (south) and loser (south) > 3
produce 1000
action 
frequency "W makes 5D" (tricks (west, diamonds) > 10, 0, 1),
frequency "E makes 4S" (tricks (east, spades) > 9, 0, 1),
frequency "S makes 4H" (tricks (south, hearts) > 9, 0, 1)

Frequency W makes 5D:
    0         614
    1         386
Frequency E makes 4S:
    0         592
    1         408
Frequency S makes 4H:
    0         353
    1         647
Generated 58680 hands
Produced 1000 hands
Initial random seed 1620989701
Time needed  139.255 sec

0

#33 User is offline   Cyberyeti 

  • PipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 12,177
  • Joined: 2009-July-13
  • Location:England

Posted 2021-May-14, 05:56

View Postnige1, on 2021-May-14, 05:12, said:

IMO, as MikeH says, what you do, depends on judgement and player habits. I would tend to bid 4 because my partner's double indicates primary support unless he has a strong hand. The likely soundness of South's 4 pre-empt is also relevant. IMO, dealerGib simulations are quite revealing. Below is another variation, defining South's hand in terms of HCP and losers.
predeal west SKT62, DAKJ9642, CQT
condition hearts (south) > 6 and 17 > hcp (south) and 6 > loser (south) and loser (south) > 3
produce 1000
action 
frequency "W makes 5D" (tricks (west, diamonds) > 10, 0, 1),
frequency "E makes 4S" (tricks (east, spades) > 9, 0, 1),
frequency "S makes 4H" (tricks (south, hearts) > 9, 0, 1)

Frequency W makes 5D:
    0         614
    1         386
Frequency E makes 4S:
    0         592
    1         408
Frequency S makes 4H:
    0         353
    1         647
Generated 58680 hands
Produced 1000 hands
Initial random seed 1620989701
Time needed  139.255 sec



I think you're allowing the S hand to be too good (hands that open 1) and precluding some hands that do open 4 (who doesn't open a badish 9 card suit and out or 8 to the KQJ)
0

Share this topic:


  • 2 Pages +
  • 1
  • 2
  • You cannot start a new topic
  • You cannot reply to this topic

1 User(s) are reading this topic
0 members, 1 guests, 0 anonymous users