BBO Discussion Forums: Sources of variance in Bridge - BBO Discussion Forums

Jump to content

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

Sources of variance in Bridge

#21 User is offline   hrothgar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 15,372
  • Joined: 2003-February-13
  • Gender:Male
  • Location:Natick, MA
  • Interests:Travel
    Cooking
    Brewing
    Hiking

Posted 2023-January-31, 08:01

View Postthepossum, on 2023-January-22, 19:10, said:


I do appreciate the truth will never be known. Just endless speculation and debate.
And of course they have infinity on their side as a defence



This is simply not true.

The WBF adopted a program and an associated process for generating deals for its tournaments to forestall just this sort of discussion.

The system is called Square Deal.
It was developed by Hans Van Stavern.

It encompasses

1. An open source program to generate deals
2. A method of selecting seeds for hand generation
3. A process by which tournament organizers can pre-commit to using a salted version of a specific seed in advance of the tournament proper
4. A process to select the salt for the seed using public sources at the time of the event
5. Methods to validate that hands for the tournament were generated from the seed in question.

These types of methods are very common in fields like cryptography which often has to deal with very similar types of problems.

Applying them to bridge is not difficult
Alderaan delenda est
0

#22 User is offline   hrothgar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 15,372
  • Joined: 2003-February-13
  • Gender:Male
  • Location:Natick, MA
  • Interests:Travel
    Cooking
    Brewing
    Hiking

Posted 2023-January-31, 08:02

View Postthepossum, on 2023-January-22, 21:49, said:


One of the points I made was that there is no way to test what you or I say anyway



Incorrect
Alderaan delenda est
0

#23 User is online   thepossum 

  • PipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 2,362
  • Joined: 2018-July-04
  • Gender:Male
  • Location:Australia

Posted 2023-February-02, 03:31

View Posthrothgar, on 2023-January-31, 08:02, said:

Incorrect


I disagree. Within my bounds of certainty anyway

Has anyone ever written a good paper on the subject - or book. I am looking for a good project despite not feeling qualified

It was very unlikely for this sequence of hands to be totally random - before I am jumped on - yes we can show p<.00001 that this set of hands had x different from that, or this was not different from that, or we didn't engineer things to make it look not different from that. If the Bridge universe is so big how small does p have to be?

or I have a mate who helps organise the World Teams

There must be a place in such a huge "universe/s" as bridge where any weird sequence of hands is possible

What even constitutes a hand. Surely it is more than just an arrangement of cards???

And while we are here what makes a hand interesting is a complex matter too ( I risk causing offence so will go not further on the matter)

In case I am ever accused of disrespect I value your knowledge and opinion on such matters very highly and am grateful for you joining the thread Richard

I just find it one of the most interesting conceptual problems I have ever considered and its well beyond me. The likes(????) of John Horton Conway and John Nash come to mind. As I say I could only dream. I reckon I know enough to know it goes beyond high school probability theory

I don't need anymore fights so will just accept everything anyone says on the matter

As you can see despite my love of simplicity you can see why so many projects I worked on got out of hand very quickly and caused me problems
0

#24 User is offline   mycroft 

  • Secretary Bird
  • PipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 7,054
  • Joined: 2003-July-12
  • Gender:Male
  • Location:Calgary, D18; Chapala, D16

Posted 2023-February-02, 11:29

So, have you read the document that hrothgar pointed you to? (granted, because he typoed the name "SquareDeal", and "Square Deal" with a space means something different, it may have been difficult for you to do so.) The doc files, in particular the BigDeal simplified, the introductory parts of the BigDeal detailed and the SquareDeal documents, will give you the information you need.

Some choice quotes:

Quote

Before we can design new software it should be clear what it should achieve. Dealing software is supposed to simulate correct shuffling by the players themselves, and should deviate from this goal as little as possible. Therefore dealing software should conform to certain minimum requirements:
1. The software should be able to generate every possible bridge deal, since that is also possible with manual dealing
2. The software should generate every deal with the same probability, without being influenced by the board number, previous hands or any other circumstance
3. It should be impossible to predict deals, even after having seen all other deals in the session


Quote

To avoid all these flaws, the new software was designed according to a couple of principles:
1. Gather more than 96 bits of randomness to stand a chance of satisfying the requirement to be able to generate every possible bridge deal
2. Use a PRNG with good statistical properties to satisfy the requirement of generating all hands with equal probability
3. Use a PRNG with strong cryptographic properties to satisfy the requirement of not being able to predict deals
5. Make it as difficult as possible for the operator of the program to make the sort of mistake that would lead to a duplicated series of deals

The fifth principle will be taken to extremes: as far as possible the software should view the operator as an enemy in the struggle to generate unique sequences. Whatever the operator tries, he should not be able to get the program to generate the same sequence twice.
(my emphasis here. #4 is "allow multiple output formats to make the output easy to use". The software these two quotes reference has been available for use, and the source code available for review, and has been used for most tournaments worldwide, for over a decade)

And since there were issues with "potential manipulation of the output" brought up in the last 10 years, SquareDeal now:

Quote

SquareDeal ... uses the well-known program BigDeal for actual dealing, but adds two things:
1. An easy way to make multiple sets of deals for various sessions using consistent naming
2. A way for participants to check the hands were dealt honestly and without any tricks by the organizer
Using SquareDeal a specific procedure is needed, with various phases, by organizer and participants. There is no actual need for participants to do anything, but if they want to check the hands after the tournament, they need to do something even before the tournament.
Which involves "Using something I keep secret (until afterwards, for post-game checks), something I release publicly (so that I can't manipulate anything during dealing), and something that doesn't exist until well after I commit to the public information (but we can both find out trivially when it does happen)".

Don't just say "I can see all these problems, therefore there's no way to disprove my conclusion, therefore somebody must have done it." Yes, there are all those problems. People have been trying to come up with solutions to these problems (in and out of bridge) since Shannon's work in Bell Labs in 1943. There are (complicated) solutions to a lot of the trivial problems, at least.
When I go to sea, don't fear for me, Fear For The Storm -- Birdie and the Swansong (tSCoSI)
0

#25 User is offline   hrothgar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 15,372
  • Joined: 2003-February-13
  • Gender:Male
  • Location:Natick, MA
  • Interests:Travel
    Cooking
    Brewing
    Hiking

Posted 2023-February-02, 12:22

View Postthepossum, on 2023-February-02, 03:31, said:


There must be a place in such a huge "universe/s" as bridge where any weird sequence of hands is possible



The following might be a helpful way to frame some of this

1. Start by enumerating every possible bridge hand in existence
2. Sort these hands however you want
3. Assign a unique number to each and every bridge hand

Next create a pseudo random number generator that is capable of creating a stream of numbers between one and the number of possible bridge hands
Seed the PRNG and then start walking though the database of hands until you have all the hands that you need for your event

It is absolutely possible for this type of technique to create some biased set of hands.

Imagine a world in which the way in which you sorted the set of bridge hands had some period to it. (All hands that correspond to a number than ends in a zero contain a singleton or something like that). And then, assume that the your stream of numbers is coming from a Linear Congruent Generator or some such where the number stream also has a period to it. And these two systems have the same frequency

Well, in this case the hands that you're generating would very likely be biased.

However, so long as you're living in a world in which the PRNG that you're using and the method that you used to assign numbers to hands didn't collide like this you're golden. (And, its possible to chose the sorting method and the design of the PRNG such that they shouldn't share this sort of ugly property)
Alderaan delenda est
0

#26 User is offline   hrothgar 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 15,372
  • Joined: 2003-February-13
  • Gender:Male
  • Location:Natick, MA
  • Interests:Travel
    Cooking
    Brewing
    Hiking

Posted 2023-February-02, 12:25

You might also find it useful to take a look at two dimensional space filling designs

Not the same as a PRNG, however, a bunch of the same issues are involved and its easy to visualize what is taking place
Alderaan delenda est
0

#27 User is offline   blackshoe 

  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 17,562
  • Joined: 2006-April-17
  • Gender:Male
  • Location:Rochester, NY

Posted 2023-February-02, 21:20

View Postmikeh, on 2023-January-22, 23:46, said:

I’m not a fast player…I used to be, then I learned (from a very good player) to slow down…and to THINK.

"Slow is smooth, smooth is fast". :-)
--------------------
As for tv, screw it. You aren't missing anything. -- Ken Berg
I have come to realise it is futile to expect or hope a regular club game will be run in accordance with the laws. -- Jillybean
0

#28 User is offline   helene_t 

  • The Abbess
  • PipPipPipPipPipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 17,068
  • Joined: 2004-April-22
  • Gender:Female
  • Location:UK

Posted 2023-February-03, 03:44

I just pulled some variance components out of my sleeve:

Total points for NS, single board across a somewhat heterogenous field (say a typical club night):
- Strength of the NS cards: 60%
- Randomness and skills: 40%

Mathchpoints for NS, single board across a somewhat heterogenous field (say a typical club night):
- Skills of NS relative to EW: 30%
- Skills of the NS field relative to the EW field: 5% (but this will depend on the size of the field)
- Which random mistakes happen to get punished (or awarded!) on this board: 45%
- Luck in toss-up situations: 13%
- Systems issues (for example some boards are good for the weak NT pairs): 5%
- Ethics (use of UI, incomplete disclosure): 2%

Maybe the black/white distinction between the 3rd and the 4th point is problematic. If a declarer gets lucky with a 49% guess for finding the queen it is technically an award for a mistake but not distinguishable from a toss-up situation, as the probabilities are usually a bit subjective.
The world would be such a happy place, if only everyone played Acol :) --- TramTicket
0

#29 User is offline   pilowsky 

  • PipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 3,620
  • Joined: 2019-October-04
  • Gender:Male
  • Location:Israel

Posted 2023-February-03, 04:27

Possibly the problem posed by the possum doesn't relate to variance associated with dealing the cards (an issue 'dealt' with many times here).
Suppose you sit down in an average tournament with an average partner on an average day.
What is it in Bridge that causes such enormous variance that, say compared to chess or go or football, the pair that is known to be excellent doesn't necessarily defeat a weak pair?

Clearly, many factors are in play.
Tiredness of a person could be major source of variance.
Unexpected gifts another.

As noted above randomness doesn't mean blandness.
So one pair could get lucky and find a sequence of deals that particularly suit their style.
In this way Bridge seems to have many more non-skill uncontrollable elements than many other games.

One advantage (there are quite a few) of the robot game is that many of these sources of variance are stabilised.
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