A Call to Arms - MoP Mechanics Testing

[Klaudandus]

they normalized HotR, right?

[Schroom]

We are changing Vengeance to increase AP equal to 5% of the damage you take for 20 sec. This buff will "roll" so that as it gets refreshed the unused part is added to the new buff, similar to how Ignite or Stagger work. The net result will be that Vengeance stacks by the amount of DPS being thrown at you over 20 sec. There is no cap.

WOAH WOAH WOAH

[theckhd]

they normalized HotR, right?

Yep. HotR now uses normalized weapon damage for both components.

HotR, Censure, SoT, HoW all match their tooltips very well.

Judgement matches the tooltip SP scaling, but still has the hidden AP scaling (which was also nerfed, from 0.397):

Code: Select all

j_fit =

     General model:
     j_fit(x) = a*x+b
     Coefficients (with 95% confidence bounds):
       a =       1.143  (1.137, 1.15)
       b =       572.9  (533.5, 612.4)

j_gof =

           sse: 3.5199e+003
       rsquare: 1.0000
           dfe: 6
    adjrsquare: 1.0000
          rmse: 24.2210

In other words, 562+0.3175*AP+0.508*SP

Something's going on with the EF data though. There's a distinct difference in the scaling for the three higher-AP data sets. I'm not sure what's going on, but the SP scaling for those last three data points is about 10% higher than the first three.

In any event, we're expecting something like (4030-4491)+0.377*SP for the base heal, and getting closer to 0.5*SP scaling. The HoT is similarly off, giving us ~6% SP scaling rather than 4.5%. And the base values are a little lower than the tooltip predicts:

Code: Select all

ef_fit =

     General model:
     ef_fit(x) = 3*(a*x+b)
     Coefficients (with 95% confidence bounds):
       a =      0.5303  (0.514, 0.5466)
       b =        3908  (3804, 4012)

ef_gof =

           sse: 1.6303e+005
       rsquare: 0.9993
           dfe: 5
    adjrsquare: 0.9991
          rmse: 180.5718


ef_fit2 =

     General model:
     ef_fit2(x) = 3*(a*x+b)
     Coefficients (with 95% confidence bounds):
       a =      0.5472  (0.5176, 0.5767)
       b =        4272  (4083, 4461)

ef_gof2 =

           sse: 5.3618e+005
       rsquare: 0.9978
           dfe: 5
    adjrsquare: 0.9974
          rmse: 327.4687


eft_fit =

     General model:
     eft_fit(x) = 3*(a*x+b)
     Coefficients (with 95% confidence bounds):
       a =     0.06158  (0.05956, 0.06359)
       b =       375.2  (362.3, 388.1)

eft_gof =

           sse: 2.4982e+003
       rsquare: 0.9992
           dfe: 5
    adjrsquare: 0.9990
          rmse: 22.3527

I think we're going to need to take a closer look at EF. Probably a bunch of 1-HP EF's (to increase sample size) with each gear set.

[Klaudandus]

OK, will get that once I get out from work

Any reason why cs is not?

[theckhd]

Any reason why cs is not?

Why CS is not what?

[Klaudandus]

Why cs damage is all over the place while hotr's isn't

[theckhd]

You didn't give me CS data, did you? It's not in the spreadsheet, at any rate.

[Klaudandus]

no, but what i meant why cs has a variable damage range rather than always hitting for the same amount like other abilities. specially considering hotr behaves now like this

[theckhd]

It won't hit for a range if you're using a Club. CS and HotR will both use the weapon's min and max damage values to determine the range. That's why I have you using the Club in the first place.

[Klaudandus]

was planning to do more logs and fill in the gaps... the servers have been offline for two hours now...

[theckhd]

No problem. I think the next couple we should focus on are EF and Holy Prism, since our data sets for those two abilities are ambiguous so far. I'd add WoG to that list, but we know it's getting buffed soon anyway, so we can probably just wait on that.

[theckhd]

Link

Can Sanctity of Battle affect Avenger Shield?

Yes.

+1 for the blog?

[Schroom]

+1 for the blog?

and Gold goes to: Theck! \o/ Hooray!

[theckhd]

OK, new build with changes to talents. I don't think we need to re-test anything we did in b15952, but we should test all the L90 talents and EF.

[Klaudandus]

Yeah, ahead of you in that matter, already prepped the spreadsheet for that.

Kinda glad I couldn't test anything at all last night.

OTOH, how many casts of ES do I need per AP level, 60?