The real-time statistics are computed on fairly small blocks of data (8 Kb). The more data you pull from random.org, the higher the accumulated entropy gets. (A new statistics module will be online in a couple of months and will feature accumulated entropy.) I therefore doubt that taking larger amounts of data will make it any easier to compress it. Compare the statistics for the 8 Kb blocks with those for the 1 Mb file mentioned in my essay.
When you look at such small amounts of data, entropy isn't a very
good test for randomness. Even for large amounts of data, it is by no
means exhaustive. For example, a typical gzipped file has a high
level of entropy but is not very random. Other tests (such as chi
square) can be used to reveal this.
(Note that I'm not involved in this challenge. I just operate random.org.)
Slashdot Mirror
The real-time statistics are computed on fairly small blocks of data (8 Kb). The more data you pull from random.org, the higher the accumulated entropy gets. (A new statistics module will be online in a couple of months and will feature accumulated entropy.) I therefore doubt that taking larger amounts of data will make it any easier to compress it. Compare the statistics for the 8 Kb blocks with those for the 1 Mb file mentioned in my essay.
When you look at such small amounts of data, entropy isn't a very good test for randomness. Even for large amounts of data, it is by no means exhaustive. For example, a typical gzipped file has a high level of entropy but is not very random. Other tests (such as chi square) can be used to reveal this.
(Note that I'm not involved in this challenge. I just operate random.org.)