Timeline for Convergence concept for MonteCarlo(?) simulation of a tournament
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 9, 2017 at 23:17 | vote | accept | cd98 | ||
| Sep 8, 2017 at 17:54 | comment | added | eric_kernfeld | Yes, that's correct: it happens event by event, and if you specify any reasonable idea of total error, it will converge because of the pointwise (event-wise) convergence of the estimates. Theoretically, it helps that your state space is finite. My reference of choice is to the multivariate law of large numbers: stat.umn.edu/geyer/5101/notes/n2.pdf | |
| Sep 8, 2017 at 17:16 | history | bounty ended | CommunityBot | ||
| Sep 5, 2017 at 16:51 | comment | added | cd98 | Thanks for your answer. If I understand you correctly, the convergence happens event by event? I thought there would be some convergence in distribution idea. Since the random variable here is discrete, I guess approximating by Event is the same as converging to the full probability distribution, right? (If you can add a reference for this, I'll accept your question) I thought about calculating things exactly, but the sample space was so huge that it wasn't possible. | |
| Sep 2, 2017 at 11:26 | history | answered | eric_kernfeld | CC BY-SA 3.0 |