Who Decided 2/1 as GIB system?

[arnprince]

I apologize if this has been addressed, but I have been on the site for years, and know that I have seen many more players, in total, who use SAYC as a bidding reference, whether or not it is their preferred system, than I have seen players who use 2/1. This said, why then is 2/1 the system that was chosen for the GIB robots to use? I feel that there should at least be a choice between 2/1 and SAYC, since, as above, many more players the world over know it. I understand there are other systems, and am not trying to begin a gripe session where everyone wants their own system as an option. However, it seems like an arbitrary choice made with little forethought and certainly little member input.

[Vampyr]

I think that a major reason must be that 2/1 is so much simpler, and therefore easier to program and to have understandable "agreements" with the human players.

[hrothgar]

> Who decided 2/1 as GIB system?

 

Willy the mail boy

 

On a more serious note: As I understand matters, BBO either purchased or assumed the GIB code base from Matt Ginsburg

At the time, GIB supported a number of different bidding systems.

 

From what I recall, GIB "best" system - the one it used for competitive play - was a MOSCITO variant called MOSCITO Byte

(Be glad it doesn't use this)

 

I assume that GIB's second best system was a 2/1 variant...

[lycier]

I think 2/1 GF is representative of the natural system and still is the basis of the study of the bidding system.

[1eyedjack]

I don't know about 2/1 being simpler. My perception is that most beginners are taught something similar to SA(YC) initially and after a while "progress" to 2/1 which they perceive to be an improvement not because it is simpler (or they would have learned that first) but because it is simply a more accurate system at arriving at optimum contracts. Note that on BBO help site there are two standard system templates suggested, "BBO Basic" which is akin to SA(YC) and "BBO Advanced" which is on the 2/1 model.

 

But if (as I believe) most players progress through SA(YC) to 2/1, players who know 2/1 will tend to have a passing familiarity with SA(YC), while the reverse is not so guaranteed. SA(YC) is therefore the lowest common denominator and it is not surprising that it prevails as a popular method for announcing familiarity in user profiles.

 

So the relevant question for the GIB programmers is: Should the system be dumbed down in order to maximise user familiarity, or should it pay lip-service to trying to improve its score by way of an improved system, at the possible expense of loss of some of its potential user-base? The question never entirely disappears. Having settled on 2/1, how about other add-ons, such as Drury (or reverse), Bergen, TWalsh, fit-jumps, kickback, and so on.

 

It is a balancing act. If 2/1 is a minority preference among the total population then I think it less so (if at all) among the more serious tournament players, and that balance at least I think to be appropriately struck.