A pseudorandom value is computed using a deterministic algorithm that is designed to have statistical properties similar to a true random value. There are various metrics to assess the randomness of a sequence of values; however the digits of π satisfy almost every such test despite being far from 'random'! The simplest methods make use of very large primes, as even simple arithmetic based them can yield hard to predict sequences.
As an example, look at the following sequence:
It appears to be pretty random, but in fact it is just the final three digits of a sequence starting 315739505 (the seed) and then generating new values by adding 37 and multiplying by 73 at each step. Real pseudorandom sequences use larger numbers, and more complicated functions at each step.
505, 902, 883, 496, 245, 922, 343, 76, 585, 742, 203, 856, 525, 362, 463, 836, 65, 782, 123, 16, ...
It appears to be pretty random, but in fact it is just the final three digits of a sequence starting 315739505 (the seed) and then generating new values by adding 37 and multiplying by 73 at each step. Real pseudorandom sequences use larger numbers, and more complicated functions at each step.