Here was the thread I was talking about. Just sort of died I guess. Too bad there's not more on macc.
http://www.bluegartr.com/threads/640...=1#post2258644
Here was the thread I was talking about. Just sort of died I guess. Too bad there's not more on macc.
http://www.bluegartr.com/threads/640...=1#post2258644
http://www.bluegartr.com/threads/674...curacy-Testing
Not sure what is there to pin down exactly. Enspell results may or may not be applicable to magic in general.
It would be interesting if there there was anything done on things other other than elemental magic. I think I recall something about paralyze on BG a while back but I was mainly interested in it only regarding how to model it (treating it like rate data for Poisson regression) rather than the results, and besides I think it was mainly regarding how paralyze potency is affected.
I'm not sure how much testing has been done on a potential magical hit rate floor, but I feel I can assure you that it's lower than 20%. If you look at the left graph on this page, it predicts the land rate at a given magical hit rate for spells with N states.
Violent Flourish caps out at a 99% stun rate or above (when it lands, which is based on melee accuracy like Head Butt). Based on this, I predicted it had more than 1 Stun state. If you don't understand what that means, read the page. Violent Flourish also parsed about a 99% stun rate when I was a level 75 fighting a level 77 Hpemde.
At the moment, I'm parsing myself Violent Flourishing a level 77 Hpemde with Sushi/Acc gear on as a level 61 Dancer. Grepping the log gives me:
86 attempts: 76 no procs, 6 procs, 4 misses
So I probably still had capped melee accuracy on Violent Flourish (no surprise, I was parsing a >60% hit rate and its acc bonus is at least 70). Proc rate looks like it was around 5%~10%, which could either be indicative of a 5% M.Hit Rate floor and 2 states or something else going on. I need a higher sample size before I can really say anything, but I feel confident that MHit Rate's floor is 5% at most.
For the level 3 resist, you have (1-MHR)^3 instead of MHR*(1-MHR)^3. Is that an error?
Nope.
If it resists three times, then it's a level 3 resist. It doesn't have to check against the hit rate again.
Those equations are only accurate for a 4 state spell, by the way.
To derive similar equations for 2/3 state spells, you drop the MHR* term from the front of the last state, and truncate the terms after that.
So a 2 state spell has hit rates of MHR and (1-MHR), not MHR and MHR*(1-MHR).
For a 5 state spell, the level 3 term would indeed be MHR*(1-MHR)^3.
Well, as far as I know all damage-dealing spells are 4-state. If you say that Sleep has two states (60 seconds and 30 seconds), and anything other than that is a resist, then you get:
60 seconds: MHR
30 seconds: MHR*(1-MHR)
Resist: 1 - MHR - MHR*(1-MHR) = 1 - 2*MHR + MHR^2 = (1 - MHR)^2
The other way to look at it would be:
Resist = MHR*(1-MHR)^2 + (1-MHR)^3 = MHR - 2*MHR^2 + MHR^3 + 1 - 3*MHR + 3*MHR^2 - MHR^3 = 1 - 2*MHR + MHR^2 = (1 - MHR)^2
So you're right, the resist rate is always going to be (1-MHR)^Number of States, but I guess we have a disconnect between the damage and enfeebling "number of states" definition.
I would say sleep has 3 states.
The chance of the first state (60s duration) is MHR, chance of the second state (30s duration) is MHR*(1-MHR) and chance of the third state (0s duration) is (1-MHR)^2.