There are several ways how to randomly choose winners. RandomPicker uses several approaches:
GUID – Pseudorandom Numbers
According their name those numbers are not fully random. They are the result of a specific computer algorithm. It is easier to generate pseudorandom numbers than true random numbers so you can find them in many desktop or server applications by using the RAND() function: in spreadsheets, C++, PHP, ASP and other programming languages.
RandomPicker.com offers the GUID pseudorandom method – Globally Unique Identifier. GUID generates very long numbers so the probability of the same number in the set is extremely low (2128. Here is an example of one number presented in a hexadecimal format: 35AF1252-4BFE-29629-B8FA-20911A21490C. If you are interested in details, visit this article on Wikipedia.
GUID method is definitely sufficient to draw winners.
Thermal Noise – Genuine Random Numbers
There are several ways to generate real random numbers – radioactive decay, thermal noise, shot noise, clock drift, avalanche noise, radio noise and other. Sets of random numbers are subjects of statistical tests to “check” the randomness (Autocorrelation test, Kolmogorov-Smirnov test, Diehard tests etc.).
RandomPicker provides the thermal noise for those who want to use true random numbers to draw winners. Thermal noise random number generator is based on a physical process. The hardware generator supplies an absolutely constant voltage and measures the small variations in the resistor’s heat output. This entropy is converted into random numbers.
We will experiment with another nature sources to offer more ways to choose the winners. If you have any interesting idea, please let us know!