Timeline for A Markov consensus
Current License: CC BY-SA 3.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 17, 2018 at 23:15 | comment | added | Mateusz Kwaśnicki | Oh, this seems to be much harder! Unless perhaps $m = 1$ (a relatively non-huge number), which is a somewhat natural model in genetics, I believe. | |
| Jan 17, 2018 at 18:17 | comment | added | Hauke Reddmann | Very fascinating! I wonder what happens when you replace "each node draws a random number having the majority" by "each node looks a huge number m of times on a random opinion from the array and counts which opinion came most often in this set". Now also a minority report could survive, although with very low probability. | |
| Jan 16, 2018 at 20:08 | history | edited | Mateusz Kwaśnicki | CC BY-SA 3.0 |
fixed typo
|
| Jan 16, 2018 at 13:30 | comment | added | Mateusz Kwaśnicki | @HaukeReddmann: I just added some numerical evidence. | |
| Jan 16, 2018 at 13:29 | history | edited | Mateusz Kwaśnicki | CC BY-SA 3.0 |
added numerical evidence
|
| Jan 15, 2018 at 14:27 | comment | added | Hauke Reddmann | My university computer is faster :-) Looks like your estimation is correct, but your n too low. Until 10^7, E still rises to the above 3-1/e (2.6-2.65). But for n=10^8, I found 2.44. My stats module says my values are trustable up to +-0.1, so the effect is significant. (I get numerical and time problems if I try to still go higher. Number theoretic aspects seem not so relevant - 10^7+-1 has no effect vs. 10^7.) | |
| Jan 15, 2018 at 6:44 | history | edited | Mateusz Kwaśnicki | CC BY-SA 3.0 |
self-correction
|
| Jan 15, 2018 at 0:03 | history | answered | Mateusz Kwaśnicki | CC BY-SA 3.0 |