Difference between revisions of "Pseudo random number generators"

Latest revision as of 19:24, 22 June 2016

Pseudo random number generators can hide most of reality from us.

Just take some "freely" chosen 2D block based computer game with a few dozen or so types of blocks (e.g. solid walk-through enemies goodies ...) and a grid of at least few thousand cells.

• Calculating the number of possible play-field board configurations is easily possible.
• Listing them all is beyond the capability of our universe.

To get at least a little chance for really every possible board configuration to show up before our eyes we try to use randomness.

Usually on computers pseudo random number generators (PRNGs) are used for randomness. (TODO: add details or link) The portion of board configurations that these PRNG numbers are capable of revealing to you in relation to all the possible board configurations is just an infinitesimally small speck. This is because the number of possible random seeds (usually system time) times the number of following computing steps is much much much smaller than the full number of all the combinatoric possibilities.

They PRNG numbers basically never hit a truly random number. Since this applies not only to games but also to art and research usage of PRNGs is leaving almost all of the truely hidden secrets of our world never to be revealed.

Why PRNGs are used so much

In spite of this horrid shortcoming of PRNG numbers they are still used because these numbers may be empirically well distributed. That is they can be convincingly random for a game or sufficiently well distributed to not produce detrimental artifacts in scientific endeavors.

Computers where and in most cases still are notoriously bad at accessing good random numbers.

Maybe surprisingly for some humans too are very bad at producing random numbers right from their head without tools (no coin flipping allowed) (TODO: add test yourself link). What certainly has separated humans form computers is that humans (with their non computer tools) are a more open information system with boundaries of perception that are more gentle less obviously present and obviously expandable to unknowable extent. Computers from our perspective where closed information systems. Looking from third person view computers always were open systems through us humans as intermediate step. As artificial independent advanced sensors wielding systems become more prevalent (automated research into the limits of the current physical knowledge) humans will loose this special position of finding / suggesting and testing new axioms for practicability.

True random number generators TRNGs

Interestingly even if we use coin-flips or quantum random to decide the layout of the example grid no can proof that every in principle possible configuration had truly a chance to come up - This is just our gut feeling that seems to make sense.