
Brian Bachman |

OK. I took the path of least resistance through the math and science requirements in college 25 years ago, but even as a politics major with extremely atrophied mathematical skills I can eyeball this and guess that the answer will be about 50% for standard array and perhaps 20% for elite array. Is there a particular reasion other than idle curiosity or strange desire to practice your math skills to want to be more precise?

Sizik |

For a sufficiently large number sequence though is not there always going to be some sort of pattern? This sounds like an interesting topic to read up on.
Also ‘R’ is Ruby on Rails?
http://en.wikipedia.org/wiki/R_(programming_language)
The (default) RNG used by R is the Mersenne twister, which has a period of 2^19937 - 1, although I've heard that given 624 consecutive output values, you can predict all future values.

ntin |

OK. I took the path of least resistance through the math and science requirements in college 25 years ago, but even as a politics major with extremely atrophied mathematical skills I can eyeball this and guess that the answer will be about 50% for standard array and perhaps 20% for elite array. Is there a particular reasion other than idle curiosity or strange desire to practice your math skills to want to be more precise?
Computer Science is all about precision and efficiency, cannot really just eye ball it and say good enough.
This thread is pretty interesting have some new reading material.

nicklas Læssøe |

Very interesting stuff about the RNG used by R. I can honestly say that i didnt know that, and i guess one of my professors must have simplified his answer when i asked him a year ago.
I do know for a fact that "most" danish statisticians use R to calculate stuff, but i begin to wonder why they dont employ one of the better algorithms for RNG instead.

Malaclypse |

Very interesting stuff about the RNG used by R. I can honestly say that i didnt know that, and i guess one of my professors must have simplified his answer when i asked him a year ago.
I do know for a fact that "most" danish statisticians use R to calculate stuff, but i begin to wonder why they dont employ one of the better algorithms for RNG instead.
Mersenne twister is most definitely 'one of the better algorithms for RNG'.