Am I analysing SimC results right?

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Am I analysing SimC results right?

Postby Akenia » Wed Aug 28, 2013 12:39 pm

Hello everyone :)

I'm new here and hope I can get some help.

Today I threw my prot pally und a prot warri into SimC and got a few questions.

1) what can you guys tell me about the two results in general?

2) why is there no scale factor for Crit in the Warri result despite the fact that it was selected in the options?

3) If I understand the TMI right, the TMI Distribution should always look like a f(x)=1/x function which is monotonically decreasing.
Upon my Pally this is the case but the Warri TMI Distribution looks a bit crappy although 50k iteration. Any thoughts?

Here are the results: Paladin: https://www.dropbox.com/s/e50gbxcrslelnyn/AkeniaT15H%2050k%20Iterationen.html
Warrior: https://www.dropbox.com/s/bcabx77i09oqirc/Falb%C3%A1T15N%2050k%20iteration.html

Nice greetings from Germany,
Akenia
Akenia
 
Posts: 5
Joined: Wed Aug 28, 2013 11:23 am

Re: Am I analysing SimC results right?

Postby Thels » Wed Aug 28, 2013 2:35 pm

A Simulation is as good as the data you give it. Theck spends a lot of time and effort making sure that the Prot Paladin module is up to date. I have no idea how much time and effort is going into the Prot Warrior module.
ImageImage
User avatar
Thels
 
Posts: 1295
Joined: Tue May 15, 2007 7:30 am
Location: The Netherlands

Re: Am I analysing SimC results right?

Postby theckhd » Thu Aug 29, 2013 8:08 am

Guten Tag!

I don't know why Crit doesn't show up on the scale factor plot, but as to it's value: does crit give any survivability benefit in 5.30? If so, maybe it isn't implemented in the warrior module? I don't maintain the warrior module, so I can't really say for sure. Best bet would be to go to http://code.google.com/p/simulationcraft/issues/list and open a new issue with a link to your warrior results.

The TMI distribution isn't actually a 1/x function in general. It's a highly skewed Gaussian, much like what you see in the warrior results. The paladin one probably looks very similar, but it's being hidden by the finite binning of the histogram. I'm not super-happy with the way that histogram plot works, to be honest; a single high-TMI iteration can cause it to look like the paladin results or worse, which makes it hard to interpret what's going on. I'm not sure if it's an easy change to make it a logarithmic plot though.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7950
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Am I analysing SimC results right?

Postby Akenia » Thu Aug 29, 2013 12:23 pm

Good evening

Thanks for your replies!

For Warriors, critrating has no practical usefulness in terms of mitigation at all. So that would be pretty much the answer.
And a logarithmic plot would actual help to interpret the TMI Distribution :)

But what about my first question (Paladin only)? Is a TMI of 3677 appropriate for hero content? Are e.g. values like DTPS good?
Maybe I have to compare the result with other results or raid logs?

My problem is, I don't know how to handle the result to get some clues on how to improve myself or to troubleshoot something.

Hmm... many questions^^

How do you guys get profit from your simulations?

Greetings, Akenia
Akenia
 
Posts: 5
Joined: Wed Aug 28, 2013 11:23 am

Re: Am I analysing SimC results right?

Postby Jackinthegreen » Thu Aug 29, 2013 2:35 pm

SimC is a free program maintained by people who have some free time and energy to devote to it. No one makes any money off it. THe "profit" from the simulations is the satisfaction of allowing people to make better choices in the game, which means improving the game experience as a whole.

The basic thing for TMI is the lower it is, the better you can survive bursts because your damage will be smoother. If you have it show stat values in terms of TMI, you'll have an idea on what to do as far as gear goes. At this point, though, it's mostly just boost haste and stamina to get the biggest gains, but there are some effects which will change that such as gaining tier bonuses and such.

I'm not entirely sure if a guide to TMI has been done for things like a certain value means you'll be fine against a certain level of content. Make sure you're simulating against the content you intend to do since telling it to set you against a T14H boss, for example, would give very different results than setting it to a T16H boss.

DTPS is Damage Taken Per Second. If you're basing your decisions on TMI, don't worry about DTPS because even if you take more damage overall, being able to smooth it out is very beneficial for you since dying to spikes is usually what kills tanks. Smoothing damage out also helps healers because they don't have to use their quick, expensive heals so often since you're not getting such large hits. Healers would prefer you have a constant, predictable steam of damage because they can use their more efficient heals.
Jackinthegreen
 
Posts: 49
Joined: Thu Aug 27, 2009 4:22 pm

Re: Am I analysing SimC results right?

Postby theckhd » Thu Aug 29, 2013 2:43 pm

Roughly speaking, I've tuned TMI such that you should get a value between 1k and 10k for a boss you're properly geared for. Below 1k means you significantly overgear that boss. Over 10k generally means you undergear it a little, and obviously the more undergeared you are the higher the number gets. It's not uncommon to sim a player in LFR gear against the T15H boss and get a TMI score in the millions.

That said, the 1k/10k boundaries are sort of arbitrary and fluid. The Vengeance changes coming next patch will move the goalposts slightly, so for example a score of 1k might jump up to 10k. So for now, a score of 3k is well within the "appropriately geared" range, but don't be surprised if it jumps up when you turn on the 'ptr=1' flag.

The absolute value of your TMI is a lot less important than using the metric to determine if, for example, a certain stat improves the value more than a different stat.

As far as I know, Simulationcraft doesn't get any profit in the monetary sense. It's all volunteer work.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7950
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: Am I analysing SimC results right?

Postby Worldie » Fri Aug 30, 2013 6:11 am

Can I be the noob here and ask what TMI stands for? ;)
theckhd wrote:Fuck no, we've seen what you do to guilds. Just imagine what you could do to an entire country. Just visiting the US might be enough to make the southern states try to secede again.

halabar wrote:Noo.. you don't realize the problem. Worldie was to negative guild breaking energy like Bolvar is to the Scourge. If Worldie is removed, than someone must pick up that mantle, otherwise that negative guild breaking energy will run rampant, destroying all the servers.
User avatar
Worldie
Global Mod
 
Posts: 13426
Joined: Sun Sep 02, 2007 1:49 pm
Location: Italy

Re: Am I analysing SimC results right?

Postby Sagara » Fri Aug 30, 2013 6:15 am

Theck Meloree Index.

You get one extra question to ask WTF that is. :-p
When that day comes, seek all the light and wonder of this world, and fight.

Worldie wrote:I used to like it [mean] back on Sylvanas.

Queldan - EU Stormrage (H) - Good night, sweet prince.
User avatar
Sagara
 
Posts: 3445
Joined: Mon Jan 12, 2009 1:04 am
Location: Belgium

Re: Am I analysing SimC results right?

Postby Worldie » Fri Aug 30, 2013 6:21 am

^ wtf that is? xd
theckhd wrote:Fuck no, we've seen what you do to guilds. Just imagine what you could do to an entire country. Just visiting the US might be enough to make the southern states try to secede again.

halabar wrote:Noo.. you don't realize the problem. Worldie was to negative guild breaking energy like Bolvar is to the Scourge. If Worldie is removed, than someone must pick up that mantle, otherwise that negative guild breaking energy will run rampant, destroying all the servers.
User avatar
Worldie
Global Mod
 
Posts: 13426
Joined: Sun Sep 02, 2007 1:49 pm
Location: Italy

Re: Am I analysing SimC results right?

Postby Sagara » Fri Aug 30, 2013 6:37 am

Theck recently devised a way to estimate a tank's anti-burstiness capability that was more practical than the gigantic tables he usually wielded on Sacred Duty - a simple number that DECREASED when a tank's ability to withstand bursts INCREASED.

He mainly used that because he wanted a simpler metric to plug around and needed a catchy name: Theck-Meloreee Index. TMI.

Since then it's been plugged in the prot paladin Simcraft module and I feel it's becoming a fast go-to value to estimate gear/talent choices.
When that day comes, seek all the light and wonder of this world, and fight.

Worldie wrote:I used to like it [mean] back on Sylvanas.

Queldan - EU Stormrage (H) - Good night, sweet prince.
User avatar
Sagara
 
Posts: 3445
Joined: Mon Jan 12, 2009 1:04 am
Location: Belgium

Re: Am I analysing SimC results right?

Postby Thels » Fri Aug 30, 2013 6:41 am

Sagara wrote:Since then it's been plugged in the prot paladin Simcraft module and I feel it's becoming a fast go-to value to estimate gear/talent choices.


It's not a prot-paladin specific option. All tanks have the option available in SimCraft.

Of course, their accuracy depends on the accuracy of the class/spec module that's being used. When released, several classes had a TMI value in the billions, simply because the simulator wasn't keeping their AM up correctly, like DKs spamming Death Strike whenever they had the 2 runes for it.
ImageImage
User avatar
Thels
 
Posts: 1295
Joined: Tue May 15, 2007 7:30 am
Location: The Netherlands


Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: No registered users and 1 guest

Who is online

In total there is 1 user online :: 0 registered, 0 hidden and 1 guest (based on users active over the past 5 minutes)
Most users ever online was 380 on Tue Oct 14, 2008 6:28 pm

Users browsing this forum: No registered users and 1 guest