Originally Posted by purefury
if i could ask something on top of that for future.
stats scale differently, and it would be very useful information to see how much dps 1point gives and (for example) 200points would give (per each point)
im very curious to see how extra 200haste translates to dps compared to 1haste*200, especially when we can swap gear gaining 120haste or so on single item, and i feel stat value for current bis might be very 'temporary'
ill try geting simcraft too, but last time gui overwhelmed me
|
As Calidius surmises, you have bumped into the largest flaw in any dps simulation architecture. Formulation provides precise numbers that are more-or-less correct. Simulation provides correct numbers that are more-or-less precise.
The accuracy of simulation scale factor generate degrades in super-linear fashion with decreasing stat deltas.
The following formula is NOT truly correct:
Scale_Factor = ( DPS_StatDelta - DPS_Baseline ) / StatDelta
Multiple runs could produce DPS_StatDelta values in the range DPS_StatDelta_Min to DPS_StatDelta_Max
Similarly, DPS_Baseline is really a range from DPS_Baseline_Min to DPS_Baseline_Max.
That means Scale_Factor is also a range:
Scale_Factor_Min = ( DPS_StatDelta_Min - DPS_Baseline_Max ) / StatDelta
Scale_Factor_Max = ( DPS_StatDelta_Max - DPS_Baseline_Min ) / StatDelta
Ideally, you want Scale_Factor_Min and Scale_Factor_Max to be as close as possible. (I do not actually report these values now, but perhaps I should.)
We are squeezed between wanting to create trustworthy scale factors and the fact that DPS is not a perfectly linear graph. The closer DPS is to linear the safer it is to use large stat deltas.