View Single Post

dipstik's Avatar


dipstik
01.17.2013 , 11:56 AM | #11
Quote: Originally Posted by Kitru View Post
I still want to know how you somehow managed to determine that 4% shield chance is worth more than 4% K/E DR (which is blatantly false to pretty much anyone that knows anything about mitigation in TOR). Even *ignoring* Blade Barrier, the hybrid should have better mitigation because that's the only difference in passive mitigation: 4% K/E DR compared to 4% shield chance.
I never managed that. the pre blade barrier numbers for hybrid are less (hence better). those numbers were me trying to work backwards, I have updated numbers.

here is the methodology:

for reference, look at the tanking stat weights for the variable definitions.

> Juggernaut Tanking 32-7-2;

> d90(dr):=1/(100)*(5 +12+10+5+ 30 * ( 1 - ( 1 - ( 0.01 / 0.3 ) )^( ( (dr+46) / 50 ) / 0.55 ) )): ;
> d100(dr):=1/(100)*(5 +12+5+ 30 * ( 1 - ( 1 - ( 0.01 / 0.3 ) )^( ( (dr+46) / 50 ) / 0.55 ) )):
> sh(shr):=1/(100)*(5+19+ 50 * ( 1 - ( 1 - ( 0.01 / 0.5 ) )^( ( shr / 50 ) / 0.32 ) )):
> ab(abr):=1/(100)*(20+50 * ( 1 - ( 1 - ( 0.01 / 0.5 ) )^( ( abr / 50 ) / 0.18 ) )):
> ac90 := .5;
> ac100 := .5;
> ac90*(1-d90(dr))*(1-sh(shr)*ab(abr))+ac100*(1-d100(dr))*(1-sh(shr)*ab(abr));

> HYBRID Juggernaut Tanking 17-22-2;

> d90(dr):=1/(100)*(5 +12+10+5+ 30 * ( 1 - ( 1 - ( 0.01 / 0.3 ) )^( ( (dr+46) / 50 ) / 0.55 ) )): ;
> d100(dr):=1/(100)*(5 +12+5+ 30 * ( 1 - ( 1 - ( 0.01 / 0.3 ) )^( ( (dr+46) / 50 ) / 0.55 ) )):
> sh(shr):=1/(100)*(5+15+ 50 * ( 1 - ( 1 - ( 0.01 / 0.5 ) )^( ( shr / 50 ) / 0.32 ) )):
> ab(abr):=1/(100)*(20+50 * ( 1 - ( 1 - ( 0.01 / 0.5 ) )^( ( abr / 50 ) / 0.18 ) )):
> ac90 := .5;
> ac100 := .5;
> mit(dr,shr,abr):=ac90*(1-d90(dr))*(1-sh(shr)*ab(abr))+ac100*(1-d100(dr))*(1-sh(shr)*ab(abr));


where mit(dr,shr,abr) gives the mitigation from the mitigation stats. you can see that full jugg has 4% more shield.

to make my life easier i took the balanced numbers for 1400 to 2000 and created a linear regression that gives mit(N), where N=dr+shr+abr. this gives:

M[FJ](N):=-4.608E-5*N+0.346: for full jugg (FJ) and M[J](N):=-4.068E-5*N+0.3267: for hybrid jugg (HJ)

now i take the kinetic vs elemental into account by MA(N,DR[iK],DR[iE]):=0.9*M*DR[iK]+0.1*DR[iE]
where MA is mitigation with armor/DR and i is used as a placeholder for HJ or FJ

where DR[iK] is the damage reduction for kinetic (which is 1-0.5311 for hybrid and 1-0.4911 for full jugg) and elemental is 1-0.2 for both DR[FJE]=DR[HJE]=1-0.2.

so the equation for taken dps looks something like:

DT=DPS*MA(N,DR[K],DR[E])-DA-SH-HPS = DT[i](DPS,N)-DA[i],

where the i can use a FJ or HJ for the DR numbers to be referenced. for FJ we have
DA[FJ]=990/12 and DA[HJ]=990/9,
where DA is damage absorbed (from blade barrier for instance), SH is self heals (from combat technique for shadows) and HPS is heals from the healer (typically assumed to be zero for time to kill calc), where all are in units of DPS.
the final step is to include the DA in the DPS function, where this is taken as 990/12 for full spec and 990/9 for hybrid spec so we get:
DT[FJ](DPS,N):=DPS*(0.9*M[FJ](N)*DR[FJK]+0.1*DR[FJE])-990/(12):
DT[FJ](3000,1800)=519

and
> DT[HJ](DPS,N):=DPS*(0.9*M[HJ](N)*DR[HJK]+0.1*DR[HJE])-990/(9):
> DT[HJ](3000, 1800)= 451.0


to find the amount of dps required to make mitigation increase the time to kill we say that DT*t=H(N)
where H(N) is the health pool size as a function of N, which i have also set to a linear function to make my life easier:

H[J](N):=-10.5*N+44911:
H[FJ](N):=-10.5*N+44911:

same subscript uses apply for full and hybrid juggs.

so DT*t=H gives the t which kills the tank, but if we solve for t, then differentiate t with respect to N and set that to zero, we can find a relation between DPS, N and HPS for when the slope goes positive with increasing N. but thats not being discussed here.

time to kill is just H/DT