Register

the proper way to judge dps on catus

Face-rippin fun.

Moderator: Forum Administrators

Honored
User avatar
Posts: 63
Joined: Sat Aug 17, 2013 11:27 am

the proper way to judge dps on catus

Postby xahn » Tue Dec 17, 2013 9:03 pm

Hello, I've been testing catus and trying to compare different sets/gear, Can I ask based on these results, which one will do the best dps.

Code: Select all
#1
[DPS] = 254.3K
    Mean:   254781 +/- 351.22 (11107)
  Median:   254309


Code: Select all
#2
[DPS] = 252.5K
    Mean:   253155 +/- 333.69 (10552)
  Median:   252482


Code: Select all
#3
[DPS] = 253.9K
    Mean:   254800 +/- 346.97 (10972)
  Median:   253898


The reason I'm asking, is that I am specially confused about #1 and #3, because in between the ()'s one is higher, but in the positive+ number, the other one is higher, which one do you take to consider which is the strongest? thanks.

Revered
User avatar
Posts: 325
Joined: Tue Oct 23, 2012 4:19 am

Re: the proper way to judge dps on catus

Postby Alpheus » Wed Dec 18, 2013 4:26 am

All three sims are pretty similar, you can use any of the them. If any of the sets include different set bonuses/trinkets then favor in the non-dps benefits before making your decision. Out of curiosity, what ilvl are you?
Image

Honored
User avatar
Posts: 63
Joined: Sat Aug 17, 2013 11:27 am

Re: the proper way to judge dps on catus

Postby xahn » Wed Dec 18, 2013 6:59 am

564 ilvl - but I am not changing my whole gear, I'm swapping basically the helm/shoulder and trying different ones with each sim, but I want to know which one is the better, so my question is what do I look for? is the DPS = ? or is it the "Mean" and do I look for the number in between the ()'s or the number which has the + and - ? as an example I gave you three results, if you were me, which one you will use as your helm/shoulder? #1 or #2 or #3 ?

Exalted
User avatar
Posts: 503
Joined: Wed Jun 29, 2011 4:49 am

Re: the proper way to judge dps on catus

Postby Stenhaldi » Wed Dec 18, 2013 7:56 am

To produce each of those results, Catus runs a number of iterations (1000 in your case) of the simulation. Each iteration yields its own DPS figure, so across all of these iterations, the program obtains a distribution of DPS figures.

The mean and median are two measures of the central value of this distribution. Typically these will be close together; the main practical difference is that the median tends to suppress outliers.

The number after the "+/-" is the standard error, i.e. the standard deviation of the mean. It's an estimate of how uncertain the given value for the mean is. So for example, if the mean is quoted as "254781 +/- 351.22", this means the true mean probably lies between 254781-351 and 254781+351.

The number in parentheses is the standard deviation of the sample. It's an measure of the dispersion of the distribution -- that is, how much the DPS figure varies between iterations. In a practical sense, it's an estimate for how much your DPS will vary from fight to fight (due to randomness alone).

---

If none of this made sense to you, then just look at the median and ignore everything else.

Honored
User avatar
Posts: 63
Joined: Sat Aug 17, 2013 11:27 am

Re: the proper way to judge dps on catus

Postby xahn » Wed Dec 18, 2013 8:11 am

Thanks sten, it made some kind of sense to me. but previously somebody told me to look at Mean, and never look at median. However now to my understand, that the gear that I swap in #1 example will be better by a slight, right? please tell me if i'm right :P

Exalted
User avatar
Posts: 503
Joined: Wed Jun 29, 2011 4:49 am

Re: the proper way to judge dps on catus

Postby Stenhaldi » Wed Dec 18, 2013 8:13 am

You can look at median or mean. It doesn't really matter as long as you pick one.

The answer to your question is that #1 and #3 are essentially equal -- they lie within uncertainty margin of each other.

Edit: Also, if you're simming 254k DPS at level 564 then you probably have some unusual settings. Level 564 should be doing closer to 350k DPS.

Honored
User avatar
Posts: 63
Joined: Sat Aug 17, 2013 11:27 am

Re: the proper way to judge dps on catus

Postby xahn » Wed Dec 18, 2013 8:27 am

I'm doing the 564 ilvl gear, basically I'm using import from armory, then changing few options like, Shrangle & then using my talent which is DoC. I dunno what possibly be wrong with my sim, maybe the raid buffs? ill check, btw my armory if you want to try yourself (Maybe if you have free time) and see what i'm doing wrong.

http://eu.battle.net/wow/en/character/o ... ahn/simple


EDIT: I just checked, I'm using group buffs and all. I have no idea why it's giving 254k DPS

Revered
Posts: 324
Joined: Wed Jul 31, 2013 10:50 pm

Re: the proper way to judge dps on catus

Postby Tremnen » Wed Dec 18, 2013 10:51 am

I ran your sim and I am getting


[DPS] = 352.6K
Mean: 353899 +/- 575.70 (18205)
Median: 352571
25%/75%: 340.3K - 365.1K
10%/90%: 331.9K - 378.9K
Min/Max: 302.4K - 416.6K

Revered
User avatar
Posts: 325
Joined: Tue Oct 23, 2012 4:19 am

Re: the proper way to judge dps on catus

Postby Alpheus » Wed Dec 18, 2013 1:35 pm

Stenhaldi wrote:Edit: Also, if you're simming 254k DPS at level 564 then you probably have some unusual settings. Level 564 should be doing closer to 350k DPS.


This is what I was thinking as well. Most likely running without raid buffs.
Image

Posts: 16
Joined: Sat Jul 06, 2013 12:26 am

Re: the proper way to judge dps on catus

Postby Ayleena » Wed Dec 18, 2013 5:15 pm

This last bit has me concerned. My catus, at ilevel 560 (I am now 561), had me at ~230 dps, and on a stand-up fight I pull about 200-210, so I figure I was pretty close. The only difference is that I don't have a Rune of any kind.

Honored
User avatar
Posts: 63
Joined: Sat Aug 17, 2013 11:27 am

Re: the proper way to judge dps on catus

Postby xahn » Wed Dec 18, 2013 5:24 pm

You were right, I hadn't had any buffs. (I thought I did) but no buffs, no food, no flask, after using those, I got to 340k with same gear. And I also don't have rune trinket.

Return to Kitty DPS

Who is online

Users browsing this forum: Google [Bot], Google Feedfetcher, ShmooDude and 9 guests