Some Hypotheses about ARORA, the Financial Turing Test
by Patrick Burns.
Abstract: We suggest some hypotheses about why people are successful in ARORA. We offer ways of generatiing the random data to eliminate the use of those tricks. We also argue that success in ARORA is not necessarily an indication of useful knowledge.
This version: 2010 November 08 (pdf)
The blog post that led to the paper is “The ARORA guessing game”.
You can play the game at http://arora.ccs.neu.edu/