GF or limit raise?

[Guest Jlall]

def disagree with 5332 13s passing, agree that several 5422 11s will pass.

[kenrexford]

If you pass a lot of 5332 13-counts, then you do not open like I open.

[jdonn]

If you pass a lot of 5332 13-counts, then you do not open like I open.

Agree with this :D

[whereagles]

Does this have a moral or is just an academic question? :unsure:

[firmit]

Does this have a moral or is just an academic question? :unsure:

If this post had been a poll in the B/I forum, I suspect the majority would show the hand as exactly what it is - a limit raise.

 

Given the light openings people use nowadays, I suspect the majority of the posters here also would make a limit raise. This is contradictory to the responses I guess, but the opening criteria in OP was somewhat constructive in nature. Thus a GF raise.

 

So, I guess the moral must be - it depends on opening style :)

[pclayton]

I really don't see what featherweight openings have to do with calling or not calling this a limit raise.

 

All it means is that there are less hands pard accepts with.

[CSGibson]

Does this have a moral or is just an academic question? :P

Academic. I was talking with another local bridge player, and we disagreed as to whether the hand was a GF, so I thought I'd post it. It seems clearly right to GF at imps, and probably right at MP as well, given Wayne's data and the comments of the bridge community here.

[awm]

My estimate would be that you should accept a four-card limit raise with a hand about two points heavier than the worst hand you'd open. Looking at Cascade's data, it seems like:

 

(1) If you normally need 12 hcp to open with 5332 shape, which is basically standard, then the hand given is a game force. Most people open a point or two lighter with 5422 or 5431; that is still fine.

 

(2) If you frequently open 5332 ten-counts (like some strong clubbers do) then you should be making a limit raise with this hand. Presumably this means that either you open 5431 9-counts also or that you would always accept a limit raise with an "opening" 5431 hand.

[Apollo81]

I think the OP hand is a GF, and that it's borderline. I think many but not all (5332) 13 counts will pass a 4-card limit raise. I think nearly all openable (5431) hands will bid game opposite a 4-card limit raise.

[Cascade]

I really don't see what featherweight openings have to do with calling or not calling this a limit raise.

 

All it means is that there are less hands pard accepts with.

If you make the same limit raises then you should accept with the same hands.

 

Doesn't that mean there are more hands that opener declines with rather than less that accept.

[Cascade]

Does this have a moral or is just an academic question? :)

If this post had been a poll in the B/I forum, I suspect the majority would show the hand as exactly what it is - a limit raise.

 

Given the light openings people use nowadays, I suspect the majority of the posters here also would make a limit raise. This is contradictory to the responses I guess, but the opening criteria in OP was somewhat constructive in nature. Thus a GF raise.

 

So, I guess the moral must be - it depends on opening style :)

I'd make a GF raise and I think we play a fairly light opening style our convention card says:

 

"Almost all 11 HCP hands, occasional 10 HCP; Most 10 HCP with 6+ card suit or two 5-card suits; Some distributional hands with 8-9 HCP; Can be light 3rd seat; can be lighter at favourable VUL"

 

We have probably pulled back slightly from when I wrote that. I certainly have I especially vulnerable. I passed a couple of 10 HCP with six-card suits last weekend.

[Cascade]

Does this have a moral or is just an academic question? :)

Academic. I was talking with another local bridge player, and we disagreed as to whether the hand was a GF, so I thought I'd post it. It seems clearly right to GF at imps, and probably right at MP as well, given Wayne's data and the comments of the bridge community here.

I am collecting some single-dummy data at the moment.

 

GIB is playing 1000 hands with five spades in a 5-3-3-2 12-count hand opposite this hand. This will hopefully give me a better handle on a how clear it is to GF. It will also give us an estimate albeit biased of how big the double dummy advantage or disadvantage is over single-dummy play.

 

It has played 384 hands in nearly 15 hrs so we are about 24 hrs away from getting the results. I might try and get some preliminary results after 500 hands this evening.

[awm]

I really don't see what featherweight openings have to do with calling or not calling this a limit raise.

 

All it means is that there are less hands pard accepts with.

A game force should be a hand where game has decent odds opposite partner's minimum holding. If you game force with less than this, you will bid a lot of bad games. If you make non-forcing invites with hands this good, you will miss a lot of good games. So I doubt this statement is very controversial. However, this means that what is a game force depends rather a lot on what hands will partner open.

 

Obviously, this means that if game is really poor opposite most minimums for partner, you should not game force. You can define a limit raise as containing a certain number of points, or having a certain value using losing trick count or ZAR points or whatever of course. But it seems to make sense to state that a limit raise is the strongest possible raise that does not create a game force. Thus these are hands where game is typically not so good opposite a minimum holding from partner, but should have good chances if partner is just a little better than minimum (say about a queen better). Once again, it depends on what is a minimum opening for partner.

 

On the hand in question, it seems like:

 

(1) Partner's opening promises 12 hcp if 5332, or 11 hcp if 5422/5431. We should game force on the example hand.

 

(2) Partner's opening promises 10 hcp if 5332, or 9 hcp if 5422/5431. We should make a limit raise, since game will be lousy opposite partner's worst possible hand (so forcing game is crazy) but game will be good if partner has an extra queen (i.e. would open under criteria 1).

 

(3) Partner's opening could be any 8 hcp with four or more spades. Now we probably need to distinguish even further, between this hand (which makes game if partner has an ace extra) versus a slightly better hand (which makes game opposite a 5332 10-count). So you need multiple ways to make an invitational raise...

 

... which brings us to one of the big problems of systems which include both very light openings and a very wide range, that "invites" become much more common and you need ways to show different types of invite as well (distinguish the "mild invite" from the "serious invite").

[Cascade]

Here are the preliminary results of a single dummy simulation.

 

I dealt the hand in the opening post to one hand opposite a 12 count with some 5-3-3-2 distribution. GIB then played the hand single dummy. So far I have repeated this 522 times (out of a run of 1000 that is running).

 

This is a summary of the results:

 

Tricks  DD      SD
0       0       0
1       0       0
2       0       0
3       0       0
4       0       0
5       0       0
6       0       1
7       1       3
8      27      37
9     188     163
10    235     233
11     67      78
12      4       7
13      0       0

 

Double Dummy (rows) versus Single Dummy (columns)
     6   7   8   9  10  11  12	
 7   1   0   0   0   0   0   0   1
 8   0   0  12  13   2   0   0  27
 9   0   2  23 104  54   5   0 188
10   0   1   2  42 152  36   2 235
11   0   0   0   4  24  36   3  67
12   0   0   0   0   1   1   2   4
     1   3  37 163 233  78   7 522

 

Trick Diff  DD-SD
 -2          9
 -1        106
  0        306
  1         91
  2          9
  3          1

 

Single Dummy was on average 0.022988506 tricks better than Double Dummy. This was a total of 12 tricks over 522 hands.

 

The standard deviation was 0.73010942.

 

I think that means (my statistics are a bit rusty) that a test that Single Dummy has an advantage would fail. The test statistic would be:

 

0.022988506/(0.73010942/sqrt(522)) = 0.719379475

 

Which I think means there is a 0.47190714 chance (two tailed) of getting a higher statistic if single dummy tricks were in fact on average equal to double dummy tricks.

 

The single dummy simulations seem to concur with the double dummy simulations that this hand is worth a game force even opposite an unshapely (5-3-3-2) minimum (12 hcp) opening bid.

[mikestar]

The hand Frances gave is a good test:

 

Kxxxx

Kxx

Qxx

Kx

 

If this is an opening bid, the hand opposite is a limit raise; if this hand is a pass, a GF raise seems in order on responder's excellent hand.

 

 

I strongly suspect the OP would pass this mess. No one at all skilled with the LTC (for example, has read Klinger's book) would evaluate this hand as 7 losers. There are three negative factors:

• One queen, no aces
  
• Poor trumps
  
• Kx in ♣ is a poor 1 loser holding unless partner something in clubs
  

So this is an eight loser hand. If the x's are truly small cards and there are no 10's and 9's, then I would rate this nearer nine losers than seven.

[matmat]

I dealt the hand in the opening post to one hand opposite a 12 count with some 5-3-3-2 distribution. GIB then played the hand single dummy. So far I have repeated this 522 times (out of a run of 1000 that is running).

Heh.

I am going to ask a stupid question, as that is the only thing i have talent for. I assume that you are playing the exact same hands single dummy and double dummy? and I also assume that the defence is also SD?

[whereagles]

Cascade, I must say you are the greatest :) You gotta tell me one of these days how to run those simulations. After I finish a book and a couple other things, that is. I'll call you back in 2 years, ok? :P

[Cascade]

I dealt the hand in the opening post to one hand opposite a 12 count with some 5-3-3-2 distribution.  GIB then played the hand single dummy.  So far I have repeated this 522 times (out of a run of 1000 that is running).

Heh.

I am going to ask a stupid question, as that is the only thing i have talent for. I assume that you are playing the exact same hands single dummy and double dummy? and I also assume that the defence is also SD?

Yes the exact same hands.

 

Yes the defense is also single dummy - GIB is playing all four hands and I assume it doesn't cheat :)

[Apollo81]

When you make a mistake is it called a Cascade failure? =)

[inquiry]

Here are the preliminary results of a single dummy simulation.

 

I dealt the hand in the opening post to one hand opposite a 12 count with some 5-3-3-2 distribution. GIB then played the hand single dummy. So far I have repeated this 522 times (out of a run of 1000 that is running).

 

This is a summary of the results:

 

Tricks  DD      SD
0       0       0
1       0       0
2       0       0
3       0       0
4       0       0
5       0       0
6       0       1
7       1       3
8      27      37
9     188     163
10    235     233
11     67      78
12      4       7
13      0       0

 

Double Dummy (rows) versus Single Dummy (columns)
     6   7   8   9  10  11  12	
 7   1   0   0   0   0   0   0   1
 8   0   0  12  13   2   0   0  27
 9   0   2  23 104  54   5   0 188
10   0   1   2  42 152  36   2 235
11   0   0   0   4  24  36   3  67
12   0   0   0   0   1   1   2   4
     1   3  37 163 233  78   7 522

 

Trick Diff  DD-SD
 -2          9
 -1        106
  0        306
  1         91
  2          9
  3          1

 

Single Dummy was on average 0.022988506 tricks better than Double Dummy. This was a total of 12 tricks over 522 hands.

 

The standard deviation was 0.73010942.

 

I think that means (my statistics are a bit rusty) that a test that Single Dummy has an advantage would fail. The test statistic would be:

 

0.022988506/(0.73010942/sqrt(522)) = 0.719379475

 

Which I think means there is a 0.47190714 chance (two tailed) of getting a higher statistic if single dummy tricks were in fact on average equal to double dummy tricks.

 

The single dummy simulations seem to concur with the double dummy simulations that this hand is worth a game force even opposite an unshapely (5-3-3-2) minimum (12 hcp) opening bid.

If I understand your data/results to date...

 

Out of 522 hands that would clearly pass a limit raise (that is your 11 point 5332 criteria i take it), game would make on between 307 (dd) or 318 (sd). In either case, game made on more than 50% of the hands (58.8 or 60.9%).

 

With those percentages (assuming simulated hands are constructed correctly and the results are, well, correct), bidding game instead of issuing a limit raise is a clear winner at MP and IMPs.

[matmat]

continuing the stupid question route --

 

I don't remember whether GIB takes the auction into account. Obviously this is irrelevant for the double dummy analysis, but does it matter for SD?

[Cascade]

When you make a mistake is it called a Cascade failure? =)

Ok I will show my ignorance ...

 

That is over my head :)

[Cascade]

Out of 522 hands that would clearly pass a limit raise (that is your 11 point 5332 criteria i take it), game would make on between 307 (dd) or 318 (sd). In either case, game made on more than 50% of the hands (58.8 or 60.9%).

12 count not 11 count.

[Apollo81]

When you make a mistake is it called a Cascade failure? =)

it's a series of related failures each triggered by the previous failure

 

http://en.wikipedia.org/wiki/Cascading_failure

 

they call it a cascading failure but i've always heard it called a cascade failure

[Cascade]

When you make a mistake is it called a Cascade failure? =)

it's a series of related failures each triggered by the previous failure

 

http://en.wikipedia.org/wiki/Cascading_failure

 

they call it a cascading failure but i've always heard it called a cascade failure

Only when I am juggling.