![]() High-quality algorithmic randomness is similar. But in practice, for most real-world randomness, obtaining such an omniscient perspective is infeasible to the point of being impossible. In fact, research has shown that people can inject bias into their own coin flips to skew the result towards a desired outcome.Īrguably then, from a truly omniscient perspective, nothing, not even the physical world, is truly random-it is in the nature of causality that everything that happens has a chain of prior events that caused it. Likewise, for flips of a physical coin, we can suppose that with sufficient knowledge of the environment, we might accurately predict the outcome. For example, tools exist to predict the path of a roulette ball (using data gained after it has been released and before the croupier calls, “No more bets!”), at least according to people who make money selling such gizmos. In fact, many kinds of “natural” real-world randomness to us aren't truly random either. Some people consider this “natural” randomness to be superior to simulated randomness (sometimes quoting John von Neumann's statement, “Any one who considers arithmetical methods of producing random digits is, of course, in a state of sin,” out of context to bolster their arguments). In contrast, we often consider acts like “tossing a coin” (a real physical coin) or “seeing where a roulette ball lands” as examples of “true randomness”. ![]() Almost all random-number generation on computers is done using algorithms to produce a stream of numbers that (hopefully) match the expectations statisticians would have about random numbers. ![]()
0 Comments
Leave a Reply. |