BBO Discussion Forums: Tiger Double - BBO Discussion Forums

Jump to content

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

Tiger Double The art of the speculative double

#21 User is offline   jdonn 

  • - - T98765432 AQT8
  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 15,085
  • Joined: 2005-June-23
  • Gender:Male
  • Location:Las Vegas, NV

Posted 2007-October-17, 16:58

Hannie, on Oct 17 2007, 04:55 PM, said:

This is a nonsensical argument. I'll let you figure out why.

I don't mean to say that the conclusion is wrong, just that the argument cuts no wood (literal translation of a Dutch expression especially for Josh).

As long as my name has been dropped, can you tell me why his argument is nonsensical, for my sake?
Please let me know about any questions or interest or bug reports about GIB.
0

#22 User is offline   awm 

  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 8,743
  • Joined: 2005-February-09
  • Gender:Male
  • Location:Zurich, Switzerland

Posted 2007-October-17, 17:13

It's matchpoints. If we assume that doubling doesn't effect whether declarer makes (only the score when he does or doesn't) and that no one is doubling the opponents in a partscore auction then our score against the pairs who defended a partial is already determined (if game makes, we lose to them regardless of whether we doubled game; if game fails, we beat them regardless of whether we doubled game).

So under the above assumptions, doubling (or not doubling) only effects our score relative to those who also defended a game contract. It doesn't matter where "the field" is getting on this hand.

Basically, if we think game is failing more than half the time, then we should double. Otherwise we should not double. It's that simple, unless doubling effects declarer's chance to make.

On this hand it seems like doubling will actually hurt declarer's chances of making, since he might assume the trumps are foul (and delay drawing trumps) or assume we have the trump honors (and finesse the wrong way). Obviously there are hands where declarer is just cold or just has no play, but it seems more likely that the double will induce declarer to go wrong, than that the double will induce declarer to go right.
Adam W. Meyerson
a.k.a. Appeal Without Merit
0

#23 User is offline   han 

  • Under bidder
  • PipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 11,797
  • Joined: 2004-July-25
  • Gender:Male
  • Location:Amsterdam, the Netherlands

Posted 2007-October-17, 17:24

If the rest of the field is in a making partscore then it really doesn't matter if you double. If they were going down then you would have had a top anyway, and if they make it then you would have had a bottom anyway.

Assuming that the double doesn't change the odds that they make it (not 100% clear I admit), you only have to compare the scores with the other pairs who are in game.

So if they make it more than 50% of the time then double will lower your expected MP score, and if they make it less than 50% of the time then double raises your expected MP score.

Of course, there may be a few other relevant contracts. For example, if other pairs are going down 800 in a partscore our way then double can cost but not gain. If other pairs misplay 3S for -1 then doubling can win but not lose. If some other pair is in 3NT+1 then doubling can only cost but if 3NT goes down 1 then doubling can only gain. Etc.

As Phil points out, 4S is likely either making or down 1. So at IMPs you would need better odds to double than at MPs, not the other way around.

edit: sorry Adam, I see you wrote the same. I'm a slow typer.
Please note: I am interested in boring, bog standard, 2/1.

- hrothgar
0

#24 User is offline   han 

  • Under bidder
  • PipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 11,797
  • Joined: 2004-July-25
  • Gender:Male
  • Location:Amsterdam, the Netherlands

Posted 2007-October-17, 17:29

awm, on Oct 17 2007, 06:13 PM, said:

On this hand it seems like doubling will actually hurt declarer's chances of making, since he might assume the trumps are foul (and delay drawing trumps) or assume we have the trump honors (and finesse the wrong way). Obviously there are hands where declarer is just cold or just has no play, but it seems more likely that the double will induce declarer to go wrong, than that the double will induce declarer to go right.

Well, one could argue that declarer may pick up the diamond queen if we double (not clear). Also (this agrees with your argument) partner may lead a heart if we double which could be good. Double may also make weaker declarers more nervous.

All together it isn't clear I think.
Please note: I am interested in boring, bog standard, 2/1.

- hrothgar
0

#25 User is offline   BillHiggin 

  • PipPipPipPip
  • Group: Full Members
  • Posts: 499
  • Joined: 2007-February-03

Posted 2007-October-17, 17:31

Hannie, on Oct 17 2007, 04:55 PM, said:

pclayton, on Oct 17 2007, 04:40 PM, said:

At MPs, the double is a lot worse IMO. There is very little for you to gain if they stretched to poor game. Even a plus 100 will be beating all of the -110's and -140's your way. Your 85% board just became a 95% board.

If they happened to be in the field contract and wrap it, you've converted a 40% board to a zero.

This is a nonsensical argument. I'll let you figure out why.

I don't mean to say that the conclusion is wrong, just that the argument cuts no wood (literal translation of a Dutch expression especially for Josh).

Assume that double does not alter the chances of the contract succeeding.
Let a be the number of other tables that stop short of game.
Let b be the number of other tables that bid game and do not get doubled.
Let c be the number of other tables that bid game and do get doubled.

If they score 9 tricks then your score is a+b/2 if you do not double or a+b+c/2 if you do. Your gain is (b+c)/2 by doubling.
If they score 10 tricks then your score is b/2+c if you do not double or c/2 if you do. Your risk is (b+c)/2 by doubling.
Therefore, your risk to gain is equal - double when you estimate the odds of their success at less than 50%.

When your double leads declarer to a line of play that works when normal lines fail, you lose big by doubling.

When your double guides partner to a winning defense (or leads declarer to a losing line), you win big by doubling.

Does that "cut wood"?
You must know the rules well - so that you may break them wisely!
0

#26 User is offline   jdonn 

  • - - T98765432 AQT8
  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 15,085
  • Joined: 2005-June-23
  • Gender:Male
  • Location:Las Vegas, NV

Posted 2007-October-17, 17:38

I think the above arguments involve the usually unrealistic assumption that the amount of tricks the opponents will make is predetermined to be equal at all tables (ignoring the impact doubling might have on that.) Why can't our opponents be in 4♠-1 when others are in 3♠-1? Every argument you guys are making begins assume this, assume that, if we assume the following, when frankly I don't particularly agree with any of the assumptions being made.

(You also ignore that doubling might make them take FEWER tricks. After all we have xx in trumps, this could easily induce a misguess.)

I think Phil's argument is fine. It's perfectly logical to deduce that when double would have been right we already had a good score, but when double would have been wrong our score wasn't going to be that bad. He just makes different assumptions than you (yours = people all make the same tricks, his = if this contract is normal then double is risky, and if its abnormal then double is unnecessary)
Please let me know about any questions or interest or bug reports about GIB.
0

#27 User is offline   pclayton 

  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 9,151
  • Joined: 2003-June-11
  • Location:Southern California

Posted 2007-October-17, 17:41

Hannie, on Oct 17 2007, 03:24 PM, said:

If the rest of the field is in a making partscore then it really doesn't matter if you double. If they were going down then you would have had a top anyway, and if they make it then you would have had a bottom anyway.

Assuming that the double doesn't change the odds that they make it (not 100% clear I admit), you only have to compare the scores with the other pairs who are in game.

So if they make it more than 50% of the time then double will lower your expected MP score, and if they make it less than 50% of the time then double raises your expected MP score.

Of course, there may be a few other relevant contracts. For example, if other pairs are going down 800 in a partscore our way then double can cost but not gain. If other pairs misplay 3S for -1 then doubling can win but not lose. If some other pair is in 3NT+1 then doubling can only cost but if 3NT goes down 1 then doubling can only gain. Etc.

As Phil points out, 4S is likely either making or down 1. So at IMPs you would need better odds to double than at MPs, not the other way around.

edit: sorry Adam, I see you wrote the same. I'm a slow typer.

Han, if you can demonstrate with any degree of confidence that the opponents are playing or not playing in the field contract, I would agree with you. If you want to make spec doubles against contracts most of the field is in, go ahead, but I tend not to play 'tops and bottoms'.
"Phil" on BBO
0

#28 User is offline   BillHiggin 

  • PipPipPipPip
  • Group: Full Members
  • Posts: 499
  • Joined: 2007-February-03

Posted 2007-October-17, 17:48

I covered 3 scenarios - double does not affect play, double helps declarer and double helps the defense. Yes, there was a bit of simplification for the first case - but the overall result is the same under that assumption. The other two assumptions are either big win or big loss situations and the math is rather trivial.
You must know the rules well - so that you may break them wisely!
0

#29 User is offline   awm 

  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 8,743
  • Joined: 2005-February-09
  • Gender:Male
  • Location:Zurich, Switzerland

Posted 2007-October-17, 17:59

The point I'm making is just this:

At matchpoints, you should double more aggressively if your opponents seem to be in a non-field contract.

Another way to put this is:

If, assuming everyone is in the same contract your opponents just bid, you think that doubling rates to be a winner, you should definitely double. In fact, sometimes it's right to double a non-field contract even if you wouldn't double if you thought everyone was there. The opposite, however (where you double if it's the normal spot but don't double if you think it's a non-field spot) should not happen.
Adam W. Meyerson
a.k.a. Appeal Without Merit
0

#30 User is offline   pclayton 

  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 9,151
  • Joined: 2003-June-11
  • Location:Southern California

Posted 2007-October-17, 18:25

awm, on Oct 17 2007, 03:59 PM, said:

The point I'm making is just this:

At matchpoints, you should double more aggressively if your opponents seem to be in a non-field contract.

Another way to put this is:

If, assuming everyone is in the same contract your opponents just bid, you think that doubling rates to be a winner, you should definitely double. In fact, sometimes it's right to double a non-field contract even if you wouldn't double if you thought everyone was there. The opposite, however (where you double if it's the normal spot but don't double if you think it's a non-field spot) should not happen.

Adam what are you saying?

1 has nothing to do with 2, as far as I can see, although I tend to agree with 2.

1 seems to contradict what Han is saying.
"Phil" on BBO
0

#31 User is offline   cherdano 

  • 5555
  • PipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 9,520
  • Joined: 2003-September-04
  • Gender:Male

Posted 2007-October-17, 18:44

Hannie, on Oct 17 2007, 05:24 PM, said:

If the rest of the field is in a making partscore then it really doesn't matter if you double. If they were going down then you would have had a top anyway, and if they make it then you would have had a bottom anyway.

Assuming that the double doesn't change the odds that they make it (not 100% clear I admit), you only have to compare the scores with the other pairs who are in game.

So if they make it more than 50% of the time then double will lower your expected MP score, and if they make it less than 50% of the time then double raises your expected MP score.

This isn't quite true, because the likelyhood that game makes and the likelyhood that the other pairs are in game may be correlated.

Let me make a simple toy example: Let's suppose the rest of the field is completely uniform, will be playing in spades, and either all of them play in 3S or in 4S, and that the contract is either taking 9 or 10 tricks at all tables. Let's say the chance that the rest of the field is in game is 50%, and that we are
- 70% likely to beat it when the other pairs are not in game, and
- 40% likely to beat it when the other pairs are in game.

Then we are obviously more than 50% likely to beat it, but it is also obviously losing tactics to double.

Of course this is pretty artificial, but I think there is some truth to this model - if the other pairs are in game as well, there is a better chance that our opponents have judged correctly, and that game will make.
The easiest way to count losers is to line up the people who talk about loser count, and count them. -Kieran Dyke
0

#32 User is offline   han 

  • Under bidder
  • PipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 11,797
  • Joined: 2004-July-25
  • Gender:Male
  • Location:Amsterdam, the Netherlands

Posted 2007-October-17, 19:06

That is a good point Arend, but I think on this kind of auction (responder invites, opener reinvites, responder accepts) that effect is relatively small.
Please note: I am interested in boring, bog standard, 2/1.

- hrothgar
0

  • 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