[Cascade]

hrothgar, on Nov 22 2004, 01:41 AM, said:

As usual, I think that a mixed strategy is best

Pass = 20% of the time
3H = 20% of the time
3N = 25% of the time
4H = 15% of the time
5H = 20%

The relatively uniform distribution suggests a very tough hand with no clear favorite bid.

Are these percentages calculated or guesses?

[hrothgar]

Cascade, on Nov 21 2004, 10:33 PM, said:

hrothgar, on Nov 22 2004, 01:41 AM, said:

As usual, I think that a mixed strategy is best

Pass = 20% of the time
3H = 20% of the time
3N = 25% of the time
4H = 15% of the time
5H = 20%

The relatively uniform distribution suggests a very tough hand with no clear favorite bid.

Are these percentages calculated or guesses?

Sorry to say that I pulled the numbers out of my butt...

Solving these types of problems is VERY difficult.

1. As I noted earlier, bridge is a multi-stage game. A bidding decision during one round of the auction impacts every other round of the auction. Worse yet, the "bidding game" also impacts the play of the hand... (As an example, a number of game theorists have studied Poker, however, they focus on 5 card draw or Hold-Em rather than 7 card stud)

2. The payoff matrix for the bidding game depends on how effective the opponent's are at countering the strategy in question.

I'd be very surprised to ever seen a clean analytical proof that is able to provide definitive percentages.

[flytoox]

hrothgar, on Nov 21 2004, 07:16 PM, said:

flytoox, on Nov 21 2004, 08:25 PM, said:

RIght. still, here how can you have a mixed strategy? Strictly speaking, this is a decision makig problem. There is no strategic interaction here. Cos so far pd and opp's decicision didnt rely on your bid of, from pass to 5H.

Bridge bidding can can/should be modeled as a multi-stage game. The set of available bids and definition of the available bids is contingent on earlier stages of the game.

You are completely correct that the specific meaningss of bids during earliier stages of the game do not change based on my current bid. However, my choice of bids this round will very much impact future bids by the opponents.

As a simple illustration, the entire problem can be considered as equivilent to output choices in a Stackelberg Equilibrium...

Not quite same as Stackleberg problem. Here you really dont know if you will be the leader or follower. Also, in the stackleberg problem, there is credible committment. And you are not punished by being aggressive. However, in competitive bidding, you are very likely to be punished by overbid.


THe area I think mixed strategy applies most is falsecard. Say throw 9 from T9x. The reason why this kind false card is always effective is that declarer cannt be sure if you are falsecard or playing honestly.

I definitely agree with you that game theory concept can be used in bridge. But I dont think this is the right occasion.

[the hog]

"I can't afford to pass, it's possible my rho is completely busted and partner is the beast at the table"

Ok this is a hand from the weekend's gnot final, and Jimmy hit the nail right on the head. You CAN'T afford to pass.

I held
void
AQxxxx
AKxx
Kxx
Now don't say "that is a X". I disagree with that with a good 6 card H suit. Partner passed, which I feel is a HUGE position. Opener bid 3S, I x again, lefty made a totally idiotic bid of 4S and I doubled again all pass.

This went for 500, but we are cold for 6H on the 1-1 break, and "yes" had pd bid 4H. I was going to look for slam.

Note what partner has done. By not supporting H he has:
* caused us to miss a slam
* caused m to totally overvalue the H suit in a defensive capacity. I cashed 2 Ds, the A of H and tried another H and fell off my seat when this gave the opps a ruff and sluff, not that it cost!

I think this hand is worth 4H - Nickf, (Nick Fahrer form Australian Bridge Mag by the way), is on the ball - you have 5 card support, you have 2 doubletons and a probable working Q; pd has made a vul overcall. Seriously, what more do you want?

[flytoox]

The_Hog, on Nov 21 2004, 10:21 PM, said:

"I can't afford to pass, it's possible my rho is completely busted and partner is the beast at the table"

Ok this is a hand from the weekend's gnot final, and Jimmy hit the nail right on the head. You CAN'T afford to pass.

I held
void
AQxxxx
AKxx
Kxx
Now don't say "that is a X". I disagree with that with a good 6 card H suit. Partner passed, which I feel is a HUGE position. Opener bid 3S, I x again, lefty made a totally idiotic bid of 4S and I doubled again all pass.

This went for 500, but we are cold for 6H on the 1-1 break, and "yes" had pd bid 4H. I was going to look for slam.

Note what partner has done. By not supporting H he has:
* caused us to miss a slam
* caused m to totally overvalue the H suit in a defensive capacity. I cashed 2 Ds, the A of H and tried another H and fell off my seat when this gave the opps a ruff and sluff, not that it cost!

I think this hand is worth 4H - Nickf, (Nick Fahrer form Australian Bridge Mag by the way), is on the ball - you have 5 card support, you have 2 doubletons and a probable working Q; pd has made a vul overcall. Seriously, what more do you want?

You know pd is likely to pass, and still you just overcall 2H. So you want to blame for his pass. What is wrong with dbl? Why cannt you dbl with 6Hearts?

[luke warm]

whether or not ron's 2h is correct vs. double, i still think his partner must show support on this auction... what are the reasons for rho to pass? 1) he has a trump stack and wants his p to reopen... 2) he's got a hand he should pass with

he obviously hasn't a trump stack (unless ron psyched), therefore #2... whether 3h or 4h should be bid is moot, support should be shown... but that's just my opinion

[nikos59]

I now answer with the huge benefit of hindsight, but I
wouldn't bid 4H, merely 3. I usually bid 4 with five
trumps but this hand seemed unattractive, also
the lack of spade raise.

nikos

[the hog]

Double is not a good call on this hand.

"You know pd is likely to pass".

Really? Why? partner's usually support when they have it. The problem with X is threefold :
it does not emphasise the H
doubling with a S void is not pleasant if partner leaves it in
If I x and next hand bids 3S and pd passes, now what? Do you want to guess 4H?

[luke warm]

sounds right to me

[inquiry]

2♥ is right, and I think 3♥ raise is right.

[Fluffy]

hrothgar, on Nov 21 2004, 01:41 PM, said:

As usual, I think that a mixed strategy is best

Pass = 20% of the time
3H = 20% of the time
3N = 25% of the time
4H = 15% of the time
5H = 20%

The relatively uniform distribution suggests a very tough hand with no clear favorite bid.

I was gonna say that you forgot to put 6♥ into the table Richard, but now I read the solution and it is not funny any more :-(.

Quote

you have 5 card support, you have 2 doubletons and a probable working Q; pd has made a vul overcall. Seriously, what more do you want?

A 5 card suit implies at leas 1 doubleton, 2 doubletons imply a 5 card suit directly, arguing with both longness and shortness is kinda redundant, isn´t it?


BTW I really can´t see any problem with overcalling 2♥.

[cf_John0]

At least 2H:suppporting pd while bidding against opp's S!

[Free]

2♠, invitational in ♥. I don't like my rubish, but I have a 4th ♥...

Let partner in on the decision.

[Flame]

Id bid 3H. and sometimes might pass or bid 2H as Richard advice, but wouldnt bid 4H or 5H.

[Chamaco]

3H.
I owe pard a raise, but I do not think the LOTT says we should bid 4H.
The LOTT does NOT mean "bid at the level of your combined trump length", it says that we should evaluate ou trump lemgth and the likely opps trump length and base our cost-benefit analysis based on the assumption that total tricks = total trumps.

At equal vuln, often simply bidding to the level of our trumps works, but at unequal vulnerability LOTT analysis is very tricky and moe complicated.

Here we know of a 10 cards trump fit in H for our side, but we ignore opps trump length.
Furtermore it is likely that with such a poor hand we do not have 4H, but it is a sacrifice, and even then, a doubtful one since RHO did not raise (neither side may make any game).

So let's try a LOTT cost benefit analysis:

they are likely to have 8 spades. So there are 18 tricks.
If we go down 1 doubled, (-200) they can make only 3 spades (140), when we lose 60. So we may as well let them go d four spaes for down 1 (please note that I do not buy the reasoning "I bid anyways to push them up" against good opps; at this vulnerability at IMPS they are going to sit for a double).

If we go down 2 doubled (500) or worse it won't be a bargain against any contract they can make.

[whereagles]

There is no risk of pushing them into a making game on marginal values. You hold 4 spades and a precious little club queen. No way they're going to make a game here. If something, it's PARD, not opps, who's got the power, as it was later confirmed.

[Chamaco]

whereagles, on Nov 22 2004, 02:34 PM, said:

If something, it's PARD, not opps, who's got the power, as it was later confirmed.

Bidding 3H vuln vs NV will leave the bidding open if pard has a battleship (at it happens here).
After 3H at unfavourable vuln the overcaller can pretty much expect the hand held by advancer and can procede to place the contract.

[flytoox]

The_Hog, on Nov 21 2004, 11:27 PM, said:

Double is not a good call on this hand.

"You know pd is likely to pass".

Really? Why? partner's usually support when they have it. The problem with X is threefold :
it does not emphasise the H
doubling with a S void is not pleasant if partner leaves it in
If I x and next hand bids 3S and pd passes, now what? Do you want to guess 4H?

THere are just too many hands you want to play in minors or in 4H but pd will pass if you bid a simple 2H.

If pd has:

QXXX
KX
XX
QJXXX

you want to be in 4H, or 6C. But pd will pass your 2H. Or shouldnt he?

If pd has

QXXX
X
QXXXXX
AX

you want to be in 6/7D, but pd will pass your 2H.

If pd has
QXXX
X
XX
AQXXXX

you want to be in 6/7C, but pd will again pass your 2H.

Even pd has
QXXX
XXXX
XX
XXX
you still want to be in game.


There are so many hands pd will pass your 2H.

[whereagles]

Actually, this one is similar to a hand I held this week-end:

Imps, all vuln, good opps. You hold:
♠ x
♥ 9xx
♦ JT8x
♣ Kxxxx

RHO you LHO pard
1♠ pass 1NT 2♥
pass ??

Well, do you bid 3♥ or do you fear that might lure them into a making 4♠ on marginal values? I reasoned I should bid 3♥ for two reasons:
1. I have reasonable defense (club king and diamond sequence)
2. Pard made a vulnerable non-exclusive overcall on a possible misfit auction. He's got to have them, and 4 might be on.

Right I was: he had a battleship and 4 is cold. LHO had QJx of spades and out, and was psyching. LOL

[Chamaco]

whereagles, on Nov 22 2004, 02:52 PM, said:

Actually, this one is similar to a hand I held this week-end:

Imps, all vuln, good opps. You hold:
♠ x
♥ 9xx
♦ JT8x
♣ Kxxxx

RHO you LHO pard
1♠ pass 1NT 2♥
pass ??

Well, do you bid 3♥ or do you fear that might lure them into a making 4♠ on marginal values? I reasoned I should bid 3♥ for two reasons:

Sure, this hand is even clearer, pard streatched to bid 2H vulnerable *in the sandwhich seat*.

When pard bids a suit in the sandwhich seat, I will bend over backwards to show him any kind of support I can offer :-)