Theck's MATLAB TPS analysis - WotLK/3.x

[econ21]

EDIT : The question about AS (+its hit table, +pvp "anomalies") seems to have vanished into thin air. Still, I'm leaving this here for anyone else potentially interested in nitpicks.

Sorry about that, I must have deleted the post while you were replying. The post was, as you say, about the hit cap for AS and based on PVP anomalies. The reason I deleted it was because subsequent testing AS against a dummy - hit capped, there were no misses. I know one of the misses I experienced in PVP was against a rogue, so he probably had the +4% ranged miss talent or something. The other, I can't recall clearly - I think it was a horde hunter who I guess must have put a scorpid sting on me or something. It's just them happening within a minute or so of each other made me falsely wonder if AS had a high miss rate like our taunts.

[theckhd]

This thread is now officially for the WotLK (i.e. current) code base. I will leave this thread open and stickied for now. I'll un-stickify it and possibly close it once 4.0 is released and the code is no longer relevant.

I am dangerously close to having a very, very rough version of the Cataclysm code structure ready for public review. When i say "very rough," I mean it - lots of stuff won't be implemented right away, and none of the calculation files are even begun. But it will be easier to catch bugs and discuss implementations with more eyes looking at the code.

Which means it's time to reiterate the call to arms: I'm looking for people willing to help build and maintain the code base. The project is hosted on Google Groups with a proper SVN implementation. Even if you just want to spot errors and suggest corrections, your help will be appreciated.

[Awyndel]

Keep up the good work Theck. We are really gonna need this project once blizzard finally decides how we are gonna turn out.

Unfortunately I can't code myself, but I will keep bugging you on the forum .

[Gil]

Hey,

I know this is not so important anymore as cata is knocking on the door, but i am going crazy with the Glyph comparision section. (http://maintankadin.failsafedesign.com/forum/index.php?p=529034&rb_v=viewtopic#p529034)

Section "Stats from Gear" states that 10 expertise rating adds 40 TPS. (below soft cap). Well, SoV glyph gives us 10 expertise, that is 80 expertise rating. So the glyph should give 40x8= 320 TPS , 160 after soft cap.

Still, on the graph, both the yellow and the blue bar is the double of this value, around 640/320. ALso the text after the graph explicitly says that the glyph is better tps even after the soft cap, then the AS glyph that is around 200.

[theckhd]

Hm... that does look fishy.

[Leuthas]

I've got to see this... I'm intrigued.

[theckhd]

Well, it's a pretty simple factor-of-two error. After looking at the code, I figured out what went wrong. When I initially coded it, the default gear set must have had over 26 expertise natively. As a result, I just doubled the result for the plots.

In later gear sets we dropped below 26 expertise, but I never fixed the error. I've fixed it now (for what good that does) by plotting using a scaling factor that depends on your base expertise without the glyph. The new plot should be correct, I'll go back and fix the text to reflect the changes.

[theckhd]

Goodbye cruel world.

[Epimer]

Before this disappears off the front page in a flood of 4.0 posts, I'd like to say thank you again for the amount of work you've put into this thread, Theck. Easily my most referenced page on the site (on the internet...?) and a massive, massive help to loads of people.

/salute