NIN/MNK difficulty?

[Jkei]

Here are the results -
SMN = 5,209.6 - 4288.1 = 921.5
BLM = 5,214.7 - 4,336.8 = 877.9
MNK = 5,220.2 - 4,417.7 = 802.5
DRG = 5,067.2 - 4,300.5 = 766.7
SAM = 5,412.5 - 4,647.2 = 765.3
MCH = 5,164.2 - 4,399.0 = 765.2
BRD = 4,924.9 - 4183.1 = 741.8
RDM = 4,882.2 - 4,186.0 = 696.2
NIN = 5,040.8 - 4,351.0 = 689.8

This is incorrect. Calculating the difference between 95th and 60th% leaves the confounding factor that not all jobs have equal dps potential; this calculation is biased, making jobs with high personal dps appear more difficult when this is not necessarily true. You'll want to use ratios instead (95th% / 60th%) to remove the difference in dps potential from the equation, giving a relative difference rather than an absolute one.

Now what the statistics do not show you, is the how landing or missing TA effects you and the group. Landing your TA increases the raids dps so there is no way to measure the highest dps - lowest dps for specifically ninjas in a fight. The only honest way to measure a classes' difficulty is perhaps fflogs cast per minute (CPM) measurement. You'll have to go through each fight and see the average CPM between NIN and other classes.

One of the top end fights for Neo-Exdeath for one group shows:
SMN at 5k dps with 36.1 CPM
NIN at 4.5k dps with 44 CPM

Found similar numbers for NIN/MNK comparison.

CPM is a terrible measure of job difficulty. It should be glaringly obvious that jobs with inherent haste buffs (GL, Huton, Shifu, etc) will always end up with greater CPM values than ones that do not, or even have extra-long GCDs in the case of BLM -- which, by the way, has the highest 95th - 50th difference ratio across all fights, suggesting it is the most difficult job to optimize, rather than the least difficult one as this CPM ranking would show it to be. Furthermore, an analysis with a sample size of 1 is never statistically significant.

When the biggest dps disparity probably comes from people unable to use true north correctly and doing something as simple as "positionnals" (because that's what monk is about).

> Taking into account that on the actual tier there's a lot of situations where true north is godlike resulting in a bigger disadvantage for monk players not using it than any other melees.

So, carefully planning cooldown usage results in great gains for monk? And that takes skill? It almost seems like good cooldown mapping is a universal part of all dps jobs' optimization.

[...] your numbers are based on personal DPS so I explained you that you can't judge a class difficulty only by its personal DPS.
Again, NIN optimization is mostly about rDPS, [...] you wont feel it [lower TA uptime] that much on its personal DPS.

First off, we're judging the difficulty not by personal dps, but the ratio of personal dps at different percentiles. As for TA and raid dps:
The dps gained from TA is 10% * uptime%. This is equally true for every party member's dps, including the NIN, as for the raid as a whole. Greater TA uptime is an equal relative gain for the NIN and the raid. You insist that NIN's ability to buff raid dps somehow does not correlate with increased personal dps. This is factually false. Learn to math.

So yeah, if you really want FFlogs to start getting relevant please at least start at the 70 - 75%, if you take the lowest % where players still can't do positionnals properly you're gonna end up with a bigger gap between classes with a lot of positionnals and classes with almost none. And you mostly want to look at people that understand how their class works to get an idea of their difficulty, knowing that while doing everything wrong you get more dps out of a class than another one doesn't mean anything in terms of overall difficulty.

I'm assuming you're referring to 50th % with "the lowest % where players still can't do positionnals properly". Whether or not this is true, or where the threshold lies where a player's % of succesful positionals is proper or not, is not something I'm going to address here because that is a subjective matter. What is a fact, however, is that the 50th % is the median player. 50th % performance is what a person with median knowledge and skill of a job can accomplish, good or bad. Their ability to get positionals right is the median, whether they're good or not. Comparing the average joe to the expert is a very valid method to quantify skill ceilings.

Of course, this post shouldn't be necessary at all. The burden of proof for your claim that NIN is more difficult than MNK lies with you.
Great job on the no-comment sassy .gif, though.