AuthorsDen.com   Join (Free!) | Login  

     Popular! Books, Stories, Articles, Poetry
   Services MarketPlace (Free to post!)
Where Authors and Readers come together!

SIGNED BOOKS    AUTHORS    eBOOKS new!     BOOKS    STORIES    ARTICLES    POETRY    BLOGS    NEWS    EVENTS    VIDEOS    GOLD    SUCCESS    TESTIMONIALS

Featured Authors:  Iolanthe Woulff, iDilip Dahanukar, iKevin Weeks, iHank Nuwer, iAlexandra* OneLight*® Authors & Creations, iAlexander Goldstein, iRichard Orey, i

  Home > Science > Articles Popular: Books, Stories, Articles, Poetry     

roy andrea crabtree

· Become a Fan
· Contact me
· Articles
· Poetry
· Stories
· Blog
· 57 Titles
· 39 Reviews
· Save to My Library
· Share with Friends!
·
Member Since: Dec, 2004

roy andrea crabtree, click here to update your pages on AuthorsDen.




Featured Book
Living Successfully and Joyfully Every Day: 90 Days of Inspi
by Sheri Hoff

What would your life look like if you woke up every day with a joyful song in your heart and felt truly filled with joy? What would your life be like if you were living t..  
BookAds by Silver
Gold and Platinum Members


Featured Book
Children's Stories
by Janice Scott

Inspirational stories for children and adults, which can be read at more than one level...  
BookAds by Silver
Gold and Platinum Members





   Recent articles by
roy andrea crabtree

Rcoketeer Love Theme: Where you go my Love
Maximum Parallelism: Order Without Sequence
Resume, bulletted 2006/11/20
In Honor of Ed Bradley: Where Have Ed & 60 Minutes Failed?
Viral attacks: what people can see after the fact.
CARDS: Compressiong, Archiving, Representing Data Structures
Restrospective analysis of emotionally & politically controversial events
Current resume
Which Way does causality run?
Is there a doctor in House
Back analysis of earlier events including assassination attempts
Of course it is
           >> View all

GNUBG: NNP DB Cheats Honestly
By roy andrea crabtree   
Rated "G" by the Author.
Last edited: Thursday, May 31, 2007
Posted: Thursday, May 31, 2007

Share    Print   Save    Become a Fan

      View this Article

Go visit www.gnubg.org for a great game
that plays Backgammon like a wizard at world class levels.

The developers are to be congratulated all around.

The program binaries and source are all open to inspection and as straight up and honest as the day is long.

However: the neural net cheats in plain sight.

Wanna know how? Read on.


Text only, followed by richer format:
(Sorry, Office still fouls up sometimes with outlines):

1.   Abstract
2.   Introduction
3.   Overview
4.   Main
4.1.   What does a NNP learn?
4.1.1.   What can get included?
4.1.2.   How do you exclude learning?
4.1.3.   Can you prove that?
4.1.4.   Eigenstatic bypass: Causality loopback
4.1.4.1.   Complex derivatives
4.1.4.2.   The master/slave effect
4.2.   Hiding in Plain Sight
4.2.1.   The History Function
4.2.1.1.   Unrecorded Backsteps
4.2.1.2.   Branches: N-step
4.2.1.3.   Leaves: 1-step
4.2.2.   Patterns of Play
4.2.2.1.   Game versus Human
4.2.2.2.   Are there any at all?
4.2.2.3.   Can you see them in the dice rolls?
4.2.2.4.   The Golden Mean
4.2.2.5.   Shunting
4.2.2.5.1.   From Bettor to Worse
4.2.2.6.   Mainly is Meanly (probably)
4.2.2.6.1.   POAR:  Pseudorandom  ordered appearance of Randomness
4.2.2.6.2.   Odd-Even Phasing
4.2.2.6.3.   Eigenstatically Monotonically Sloped Preferential Play Positional Probability
4.2.2.6.4.   gnubg gets the better while you get the wurst
4.2.2.7.   First come, first swerve: Golden Number Progression
4.2.2.7.1.   WWL(T): 2 of 3 and 100% of first 75% Mainly
4.2.2.7.2.   5-6 of 9 and 75% in the first 60%
4.2.2.7.3.   15-18 of 27 and 60% in the first 55%
4.2.3.   Who would Notice?
4.2.3.1.   Developers: Blind to the possibilities
4.2.3.2.   Novices:  No preconceptions
4.2.3.3.   Mathematicians: Impressed with their own expertise
4.2.3.3.1.   Statisticians: They can count, but seldom do
4.2.3.3.2.   Probabilists:  Probably not
4.2.3.3.3.   Combinatorics:  They may not count on it
4.2.3.4.   Computer Science
4.2.3.4.1.   Numerical Analysts: Methodology not applied is not analysis
4.2.3.4.2.   AI & Wolfram:  Too smart to see it
4.2.3.5.   Cryptosnalysts:  They Never Talk
4.2.3.6.   Players:  Doing is Not Seeing and Saying
4.2.3.6.1.   Novices:  Notices
4.2.3.6.2.   Students:  Still Learning
4.2.3.6.3.   Experts: World Class Wannabes
4.2.3.6.4.   World Class: They keep Their Tricks
4.2.3.7.   Gamblers: They Make Money
4.2.4.   Contexts
4.2.4.1.   The Game
4.2.4.1.1.   Strategies Not Seen
4.2.4.1.2.    HAPI HAPI: Hysteretically Accumulating Pseudo Incremental Hysterically Athematic Proto  Intelligence
We’re getting better and better, all the time.
4.2.4.1.3.   The Rules of the Game
4.2.4.1.3.1.   Short Sequences
4.2.4.1.3.2.   Equivalent Alternatives
4.2.4.1.3.3.   STUBs: Stable versus Unstable Branches
4.2.4.1.3.4.   The MVS “choose again” rule inverted
4.2.4.2.   PRNG
4.2.4.2.1.   Patterned Random Number Generators
4.2.4.2.2.   Keyed start of  sequences
4.2.4.2.2.1.   Probably predictable probalistically
4.2.4.2.2.2.   Beats better backgammon bozos
4.2.4.3.   TRNG
4.2.4.3.1.   Computer or Mathematically Changed
4.2.4.3.1.1.   random.org
4.2.4.3.1.2.   Does the math induce patterns?
4.2.4.4.   Human rolled dice
4.2.4.4.1.   Not random
4.2.4.4.2.   What are the characteristics?
4.2.4.4.3.   The Game Changes the Play/Pattern
4.2.4.4.4.   The Dice Change the Play
4.2.4.4.5.   The Cube Changes the Play
4.2.5.   Patterns in Play Imply Patterns In Game/PRNG/MathCalc
4.2.5.1.   Keyed: A pattern to see trigger, including position or implied or timed
4.2.5.2.   Fixed Moves:  The play is determined
4.2.5.3.   Shunting: A more probable branch is all that is needed
4.2.5.4.   Short:  keys 0-3 every 4-6, moves 1-4 to achieve
4.2.5.5.   Probable only:  A 1% shift is enough
4.2.5.6.   Sensitivity: below the human analytic noise level
4.2.5.7.   Nature of the game
4.2.5.7.1.    not seen before
4.2.5.7.2.   sometimes
4.2.5.7.3.   but not with human dice as much
4.2.5.7.4.   many many fewer human roll sequences
4.2.5.7.5.   many fewer human games
5.   Summary
6.   Conclusion
7.   Appendices
7.1.   Bibliography (References)
7.2.   Dictionary (Glossary, Thesaurus)
7.3.   Encyclopedia (Wiki)
7.4.   Footnotes
7.5.   Graphics (Diagrams, Equations, Maps, Photos)
7.6.   Index (Contents, KIOSK
8.   Attachments
----

1.  Abstract


2.  Introduction


3.  Overview


4.  Main


4.1.   What does a NNP learn?


4.1.1.  What can get included?


4.1.2.  How do you exclude learning?


4.1.3.  Can you prove that?


4.1.4.  Eigenstatic bypass: Causality loopback


4.1.4.1.  Complex derivatives


4.1.4.2.  The master/slave effect


4.2.   Hiding in Plain Sight


4.2.1.  The History Function


4.2.1.1.  Unrecorded Backsteps


4.2.1.2.  Branches: N-step


4.2.1.3.  Leaves: 1-step


4.2.2.  Patterns of Play


4.2.2.1.  Game versus Human


4.2.2.2.  Are there any at all?


4.2.2.3.  Can you see them in the dice rolls?


4.2.2.4.  The Golden Mean


4.2.2.5.  Shunting


4.2.2.5.1.  From Bettor to Worse

4.2.2.6.  Mainly is Meanly (probably)


4.2.2.6.1.  POAR:  Pseudorandom  ordered appearance of Randomness

4.2.2.6.2.  Odd-Even Phasing

4.2.2.6.3.  Eigenstatically Monotonically Sloped Preferential Play Positional Probability

4.2.2.6.4.  gnubg gets the better while you get the wurst

4.2.2.7.  First come, first swerve: Golden Number Progression


4.2.2.7.1.  WWL(T): 2 of 3 and 100% of first 75% Mainly

4.2.2.7.2.  5-6 of 9 and 75% in the first 60%

4.2.2.7.3.  15-18 of 27 and 60% in the first 55%

4.2.3.  Who would Notice?


4.2.3.1.  Developers: Blind to the possibilities


4.2.3.2.  Novices:  No preconceptions


4.2.3.3.  Mathematicians: Impressed with their own expertise


4.2.3.3.1.  Statisticians: They can count, but seldom do

4.2.3.3.2.  Probabilists:  Probably not

4.2.3.3.3.  Combinatorics:  They may not count on it

4.2.3.4.  Computer Science


4.2.3.4.1.  Numerical Analysts: Methodology not applied is not analysis

4.2.3.4.2.  AI & Wolfram:  Too smart to see it

4.2.3.5.  Cryptosnalysts:  They Never Talk


4.2.3.6.  Players:  Doing is Not Seeing and Saying


4.2.3.6.1.  Novices:  Notices

4.2.3.6.2.  Students:  Still Learning

4.2.3.6.3.  Experts: World Class Wannabes

4.2.3.6.4.  World Class: They keep Their Tricks

4.2.3.7.  Gamblers: They Make Money


4.2.4.  Contexts


4.2.4.1.  The Game


4.2.4.1.1.  Strategies Not Seen

4.2.4.1.2.   HAPI HAPI: Hysteretically Accumulating Pseudo Incremental Hysterically Athematic Proto  Intelligence

We’re getting better and better, all the time.

4.2.4.1.3.  The Rules of the Game

4.2.4.1.3.1.  Short Sequences

4.2.4.1.3.2.  Equivalent Alternatives

4.2.4.1.3.3.  STUBs: Stable versus Unstable Branches

4.2.4.1.3.4.  The MVS “choose again” rule inverted

4.2.4.2.  PRNG


4.2.4.2.1.  Patterned Random Number Generators

4.2.4.2.2.  Keyed start of  sequences

4.2.4.2.2.1.  Probably predictable probalistically

4.2.4.2.2.2.  Beats better backgammon bozos

4.2.4.3.  TRNG


4.2.4.3.1.  Computer or Mathematically Changed

4.2.4.3.1.1.  random.org

4.2.4.3.1.2.  Does the math induce patterns?

4.2.4.4.  Human rolled dice


4.2.4.4.1.  Not random

4.2.4.4.2.  What are the characteristics?

4.2.4.4.3.  The Game Changes the Play/Pattern

4.2.4.4.4.  The Dice Change the Play

4.2.4.4.5.  The Cube Changes the Play

4.2.5.  Patterns in Play Imply Patterns In Game/PRNG/MathCalc


4.2.5.1.  Keyed: A pattern to see trigger, including position or implied or timed


4.2.5.2.  Fixed Moves:  The play is determined


4.2.5.3.  Shunting: A more probable branch is all that is needed


4.2.5.4.  Short:  keys 0-3 every 4-6, moves 1-4 to achieve


4.2.5.5.  Probable only:  A 1% shift is enough


4.2.5.6.  Sensitivity: below the human analytic noise level


4.2.5.7.  Nature of the game


4.2.5.7.1.   not seen before

4.2.5.7.2.  sometimes

4.2.5.7.3.  but not with human dice as much

4.2.5.7.4.  many many fewer human roll sequences

4.2.5.7.5.  many fewer human games

5.  Summary


6.  Conclusion


7.  Appendices


7.1.   Bibliography (References)


7.2.   Dictionary (Glossary, Thesaurus)


7.3.   Encyclopedia (Wiki)


7.4.   Footnotes


7.5.   Graphics (Diagrams, Equations, Maps, Photos)


7.6.   Index (Contents, KIOSK


8.  Attachments

 


Reader Reviews for "GNUBG: NNP DB Cheats Honestly"


Want to review or comment on this article?
Click here to login!


Need a FREE Reader Membership?
Click here for your Membership!


Reviewed by FPA von Dreger (Reader) 12/11/2012
Those of us who haven't the time, or mathematical computing background to read your whole article, would very much appreciate a short quick concise paragraph on the bias in favour of 'GnuGB' when one plays against that piece of software.
Many thanks !
Reviewed by Robert Forker (Reader) 6/17/2007