Closed navv1234 closed 9 years ago
Reported by philoptik
on 2013-08-27 19:14:12
Started
I'm not sure what the concern is. Here's the stat weights from the attached html file:
Stat Scale Factor
Str -0.69
Sta -1.55
Exp -1.39
Crit -0.45
Haste -0.03
Mastery -1.30
Armor -0.80
Dodge -0.80
Parry -0.92
From this, strength is actually less effective than dodge or parry, not more. And
remember that every point of strength gives approximately 1 parry rating worth of parry,
so they should be pretty close.
More importantly, the uncertainties in that report are +/- 0.45, which means you can't
really say anything significant about the relative worth of strength, dodge, and parry.
Within experimental uncertainty, they're all identical.
Reported by theckhd
on 2013-08-27 21:18:34
Ok, I had it my head that the Str didn't convert into Parry at 1:1 so that makes sense.
Can you give any pointers to reduce the margin of error? Just throw more iterations
at it?
Reported by maynard.matt.a
on 2013-08-27 22:02:03
Pretty much. 10k is usually pretty good for paladins, but it may take more iterations
for warriors.
Reported by theckhd
on 2013-08-27 22:09:30
I just ran the profile with 500k iterations. Results are pretty reasonable (error +-
0.05_
Str Agi Sta AP Exp InvExp Hit Crit Haste Mastery Wdps Armor Dodge Parry BlockR
Scale Factors -0.46 0.00 -1.35 0.00 -1.87 0.00 0.00 -0.26 -0.00 -0.95 0.00 -0.70 -0.67 -0.68 0.00
Reported by dr.max.moellers
on 2013-08-28 13:06:10
I just ran the profile with 50k iterations making just these one change to drop some
hit (but still over cap)
reforge hit to dodge on the waist
everything looked good except Hit which had values all over the place and ended up
with a value 15x larger than any other stat with a margin of error twice as big as
the bar itself. Does this sound like normal behavior for a stat with a hardcap?
Reported by Shat.Eonar
on 2013-08-28 13:50:41
Str Agi Sta AP Exp InvExp Hit Crit Haste Mastery Wdps Armor Dodge Parry BlockR
Scale Factors -0.37 0.00 -1.30 0.00 -1.88 0.00 -19.21 -0.27 -0.02 -0.84 0.00 -0.69 -0.64 -0.67 0.00
Reported by Shat.Eonar
on 2013-08-28 13:51:43
yea. hit calc can be buggy depending on how far it is over cap.
Reported by dr.max.moellers
on 2013-08-28 14:51:13
Fixed
Reported by dr.max.moellers
on 2013-08-30 22:55:42
Fixed
Originally reported on Google Code with ID 1859
Reported by
maynard.matt.a
on 2013-08-27 15:31:26