Tweet
Did another run of 1000 distillate of divine healing and had 54 failures this time for a 5.4% failure rate which seems more in line to me than my first run of 1000.
-
Kekenewah Silverfire: 82 High Elf Wizard: Drinal Server salvage 3
300 Jeweler (plus 15%) & JM3, 300 Smith w/ SM3 (plus 15%), 248 brewer (plus 5%), 300 tailor (plus 15% & TM3), 200 baker, 217 Pottery (plus 5%), 200 Fisherman, 300 spell research w/ AT3 (plus 15%)
Kekema: 81 gnome magician 250 fletcher (plus 5%), 215 tinkerer (plus 5%)
Kekero: 65 dwarf rogue 300 make poison (plus 12%)
Keketrol: 74 troll shaman 300 alchemist w/ AM3 (plus 15%) salvage 3
Kekecha: 75 Erudite enchanter 122 smith (plus 5%)
Kekenebog: 59 Ogre warrior 203 blacksmith (plus 5%)
Kekera: 62 halfling ranger: 189 blacksmith (plus 5%)
Keneken: 75 dark elf necro: 112 tailor (plus 5%)
Keke 70 High Elf wizard on Mayong 300 plus 12% research with AT3 -
I don't agree with this idea that the RNG has anything other than a uniform distribution. All those experiencing good or bad luck with the RNG don't have it to themselves. Between each player's call to the server's RNG routine there are probably hundreds of calls by other players and mobs.
What I remember from a few years ago is that a developer new to the team was also convinced something was wrong with the RNG, so he ran his own tests and found nothing wrong with it. Although, even if the distribution looks uniform, a la a bar chart, there are other tests that need to be run. I can't remember the exact test or the mathematics behind it, but it's something like this: a good RNG should not produce consecutive numbers or or triplets (or higher as much as you want or can test) correlated to each other. In our game's case, a player's call to the RNG might consist of server call #1, #83, #197, #213 etc. and the idea that these calls are correlated just don't make sense to me.Comment
Comment