How to resolve bidding questions.

[EricK]

Is there a way to get a definitive answer to the sort of bidding questions which crop up repeatedly (eg light v. normal v. "sound" openings; tightly-defined v. loose pre-empts; rebidding 1S or 1NT on balanced hands after eg 1♣ 1♥; light openings in 3rd seat etc etc)

 

Simple argumentation and quoting selected hands doesn't work because both sides have good arguments and can find hands which show their methods to be "the best".

 

Double dummy analysis, where possible, doesn't help as it totally ignores real world factors (both that people are not capable of consistently coming up with the theoretically best approach and also that the theoretically best approach is not necessarily the same as the approach which happens to work double dummy).

 

Analysis of the results of hands with eg BridgeBrowser takes real world factors into account but has flaws of its own (eg you can never be sure the bid was chosen for the reason you are hoping so your sample may be flawed; it makes no allowance for the strength of players so you can't say whether results were due to good play or good bidding)

 

What may work, however, is a computer program which plays good bridge but has an easily programmable bidding interface so we can set up some structured tests and let it run for thousands of hands. One trouble is that for good results one would need to make sure the entire bidding system is very well programmed else the results may not be meaningful. Another is that one would need to try lots of different defensive structures out, or else there is the risk that the good results of one method are due to a poor choice of defensive structure.

 

So is there actually any hope that we can ever answer such questions and have valid evidence to back our answer up?

[cherdano]

What may work, however, is a computer program which plays good bridge but has an easily programmable bidding interface so we can set up some structured tests and let it run for thousands of hands. One trouble is that for good results one would need to make sure the entire bidding system is very well programmed else the results may not be meaningful.

Even then (and even considering just constructive bidding), you would just find out which bidding system is best suited for the bidding skills of this particular program.

 

Arend

[Free]

I guess at the moment you'll have to wait a loooooong time to find a definite answer. There is no better way, there's just "priority" and "personal preference" which fit in your system. All you can do is try to take an objective stand, and look at ALL angles.

 

For example, the 1♣-1♥-1♠/1NT problem: priority for some is to introduce the ♠ suit, for others it's showing the balanced nature of the hand with limited point count. The advantages aren't clear in this case, only if partner has a weak hand with 4♠ you'll get to the wrong contract (otherwise checkback - which in it's turn can be a disadvantage since you make artificial bids and showing more about your hand), and if he doesn't have ♠ you might wrongside the contract. What about when you're unbalanced with ♠: the first method just showed ♠, while the second method showed an unbalanced hand with ♠ - so the opening suit becomes natural. Here you have a clear advantage for the 2nd method. So when you look at the complete picture theoretically, the methods seem to balance eachother out on very different fields. All that is left is personal preference, and ofcourse the type of system you're playing...

 

For all of the posted discussions, you can make a similar analyse. What you probably can do is use GIB or so and give in the south hand you want to investigate. If you want to investigate very light preempts for example, give it 5332 with 3hcp, let it bid 3X and let GIB do the rest. Problems arise again, since the partner doesn't know what to do... So unless you can modify their systemic agreements to what YOU want them to play (in detail) it's impossible.

[jikl]

There is no answer yet.

 

People have proven mathematically which is better between 4 card majors and 5 card majors, not surprisingly, it has been proven both ways. :D

 

Same with reverse vs natural carding.

 

Sean

[blackshoe]

Bumblebees can't fly. :unsure:

 

The integral equations which define the behavior of neutrons in a nuclear reactor cannot be solved exactly.

 

There are so many possible variations in hands that trying to group them so as to allow hard and fast rules what to do with each type is fruitless. Adding in "table feel" and different opinions on hand evaluation just makes it worse. So we base "what to do" opinions on empirical evidence - most of which is at best poorly collected. Bottom line: you're not going to get definitive answers to those questions any time soon. :ph34r:

[willow23]

I class those bidding questions as issues...(No definite answer :unsure: )

 

The idea is to try out various methods and choose what works for the partnership/system etc.

 

I for one support bids which give you the most information and keeps the bidding the lowest...

 

When I play with new partners and the bidding goes 1♣/1♦1♥ - 1NTand partner tables his hand and has a 4 card major I get dissapointed, especially when there is a major fit.

 

Everone has their own style/ preference...but I figure there is much pleasure in discussion and debate...so enjoy..

 

Nothing better that an interlectual or stimulating tug of war :ph34r:

[barmar]

If there were a "best" bidding system, wouldn't you expect most top bridge players to play it? The variety of bidding systems among players at the upper echelons is, I think, an indication that the choice of system is not really that significant.

[EricK]

If there were a "best" bidding system, wouldn't you expect most top bridge players to play it?  The variety of bidding systems among players at the upper echelons is, I think, an indication that the choice of system is not really that significant.

Just because the differences are not as great as the differences due to other effects doesn't mean that there is no "best" system. Nor does it mean that we should not look for the best system purely as an intellectualk exercise.

 

After all, I am never going to be a champion at bridge. Intellectual exercises are all I have left.

[barmar]

Well, if the experts can't agree on a best system, who can?

[EricK]

Well, if the experts can't agree on a best system, who can?

People who can spend their time systematically examining the question, rather than people who waste their time playing bridge :D

[kenrexford]

This is truly a difficult undertaking. A case in point, vaguely described, illustrates the problem.

 

Suppose on a given hand a slam appears to be slightly above average in chance of success. The likelihood of slam making depends upon two factors. First, how capable is the declarer. Second, how capable is the defense.

 

If you assume absolutely perfect declaring and defense, not double dummy but based upon each person's best action within the parameters of information available, the likelihood of the slam making tends to decrease as the information exchange necessary to determine the likelihood of success in that slam increases.

 

In other words, at one point in the auction, the slam may be 50-50, in the sense that 12 tricks will be available 50% of the time that I have X hand and Y information.

 

If I increase our partnership information, I might be able to weed out certain low-likelihood positions and only bid slams half the time, perhaps the best half of the time, from the original pool. Thus, my slam might move to 75%.

 

Tat information exchange might be costly. Of the 100 slam bids I might have made without information exchange, 50 would make. If I exchange info, I might bid slam on only 60 hands, but now perhaps only 45 slams actually make because of the defense having information. Thus, 75% likelihood of success.

 

What if, however, 60 of the slams would make without information exchange, but information exchange reduces the pool such that only 20 will make. This will cause my 60% to reduce to 33%, a bad move, when I bid 60 of the slams, not 100.

 

One might suggest, then, that you could tally up all the results and decide which system caters to all of this best. The problem is that the main "cure" to this problem is table feel. If LHO looks dull, or sleepy, you might leap to a slam, expecting him to miss the killing lead. If partner looks sleepy, perhaps 3NT is best at MP if the declarer switches. If I am brain dead, perhaps I let partner know what I have and decide.

 

Reviewing old results and decisions, and mathematics, misses all of this.

 

IMO, this variable does lead me to conclude that one critical factor should be present in any general approach to facilitate table feel moves, namely flexibility. A system that enables one to make calls as needed, and not as defined, seems best.

 

A simple example of this. I hear discussions about minor openings. Some like to open better minor. Some only open a three-card diamond suit when 4432. Some open a short club, but only when 4432.

 

For years, I have not understood this. I have played that a club could be short and diamonds are 3+. Then, sometimes I open a shorter minor for a lead, or because I want to make a takeout double difficult, or to facilitate a possible cuebid if a major is agreed, or to dissuade a lead, or for any number of reasons. Sure, usually you open beter minor. But, flexibility allows you to cater to real opponents, with unique predispositions, abilities, and the like. The system is generally defined, with the added partnership agreement to not hang partner for being creative or catering to table feel. I used to pre-alert that any call may be +1 or -1 in length or strength from that described, as this was a partnership understanding.

 

No analysis I can imagine addresses this variable.