Timeline for Having a hard time with the law of the iterated logarithm
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 29, 2020 at 3:35 | comment | added | Vincent Granville | I added more details about my random number generator, see updated post, with new paragraph at the bottom. | |
| Mar 28, 2020 at 13:14 | comment | added | whuber♦ | I used an approximation that is far more accurate than the numerical imprecision in the double-precision arithmetic of the calculation: namely, that the increments in this random walk (after the first few) follow Normal distributions. One can therefore implement a simulation from the top down by generating an initial segment with Binomial increments and then obtaining the thinned values at larger times by accumulating Normal variates. To inspect the detail, construct Brownian bridges between those values. This is a kind of "just-in-time" simulation. | |
| Mar 27, 2020 at 22:23 | vote | accept | Vincent Granville | ||
| Mar 27, 2020 at 22:23 | comment | added | Vincent Granville | Thank you. Oh well, I will have to change quite a few things in two important papers that I have written, but glad to know the truth now (and most importantly, know why it works that way). It will actually make the papers more exciting after the correction. How did you produced these $2^{1000}$ deviates? Just curious. Some of the first random generators I used were faulty beyond $10^{8}$. | |
| Mar 27, 2020 at 22:04 | history | answered | whuber♦ | CC BY-SA 4.0 |