Howdy, I've had a gut feel the success rate on my GM tailor was lower than published for a while now, so I measured this afternoon. Studded Tunic Published rates (in the gump): 62% success, 20% exceptional Attempts: 76 Total Tunics: 40 (53%) Exceptional: 9 (22%) So, the percentage of tunics made which are exceptional seems around where it should be but the total tunics produced (and therefore total exceptional tunics produced) is distinctly low. Can anyone replicate this? Any idea what I'm doing wrong? Thanks, Wil
Given p=.62 and 76 trials, the odds of 40 or fewer tunics being created is around 6%. So you had an unlucky run, but it's far from proving things aren't working properly. Enter those first three numbers and see for yourself here: http://stattrek.com/online-calculator/binomial.aspx I guess technically "proving" just depends on how much uncertainty you can tolerate, but IMO I'd want to see like 300 more attempts and the results before thinking there's an issue here.
The 70 or 80 I did before I started keeping track "felt" low too. I'll keep track till I get to around 300 and follow up here.
In all honesty, even 300 isn't a significant enough sample size to declare the chances off. This is the sort of thing that Telamon could just do, and run the code directly about 100,000 times to get an more accurate statistic.
Another relatively hard one today so I collected data again: Studded Bustier: published rates - 84.2% success, 39.2% exceptional 85 trials 71 total (83.5%) 18 exceptional (25.4%)
And again: Studded Tunic gump: 62% chance success, 20% chance exceptional 72 trials 40 tunics (54.8%) 7 exceptional (17.5%)
Howdy, 149 trials on studded tunics, 80 successes. According to the web site you referenced, if the chance of producing one studded tunic is 62% as stated in the gump, the odds of getting 80 or fewer successes in 149 trials is 2.3%. How low do you want the odds to go before deciding that something is off in how the server computes tailoring success versus what it displays in the gump? Regards, Wil