SAM out classing all other DPS

[Fyce]

*doesn't know that Ninja is actually the most busted DPS by a lot as they're only 200-300 DPS behind Samurai in most cases and bring WAY more raid DPS*

Yup. Checks out.

Trick Attack averages +1.67% group DPS.
On Susano, SAM is ~350 DPS higher than NIN -both without balance- (if you include Balance, the gap jumps from 350 DPS to 450. But I won't talk about that as the parses without Balance already prove my point).
For Trick Attack to fill this gap, you need 21k group DPS. This is, of course, assuming that the NIN doesn't miss any Trick Attack and assuming absolutly no downtime... Which, in reality, makes Trick Attack worse than what I said.

In short, for NIN to be at least at the level of SAM, you need a group that can do 21k DPS. It means that your whole group has to work quite hard to make up for the fact that they bring a NIN instead of a SAM which could've been fine by himself alone.

Susano needs 14k group DPS to be clearable. As of today, quite a lot of PF groups still hit enrage. So this 21k group DPS is far from being the norm.

I check my numbers. Do you?

[HolyTurtle]

Trick Attack averages +1.67% group DPS.
On Susano, SAM is ~350 DPS higher than NIN -both without balance- (if you include Balance, the gap jumps from 350 DPS to 450. But I won't talk about that as the parses without Balance already prove my point).
For Trick Attack to fill this gap, you need 21k group DPS. This is, of course, assuming that the NIN doesn't miss any Trick Attack and assuming absolutly no downtime... Which, in reality, makes Trick Attack worse than what I said.

In short, for NIN to be at least at the level of SAM, you need a group that can do 21k DPS. It means that your whole group has to work quite hard to make up for the fact that they bring a NIN instead of a SAM which could've been fine by himself alone.

Susano needs 14k group DPS to be clearable. As of today, quite a lot of PF groups still hit enrage. So this 21k group DPS is far from being the norm.

I check my numbers. Do you?

Data for the boss dps (Susano & Lakshmi) at the 80 percentile: https://www.fflogs.com/statistics/15...sdps&class=Any

Let denote SAM dps in term of percentage of raid dps as S.

On average, NIN does 0.925S, and Trick Attack accounts for 0.0167 (=0.1*10/60) of the raid dps. For Trick Attack dps increase to be no less than the dps difference between NIN and SAM, S needs to be no more than 0.2227 (=0.0167/(1-0.925)) of the raid dps.

Consider a standard composition in which SAM does the highest percentage of raid dps, i.e. other members have the lowest personal dps in their categories: DRK/WAR/SAM/NIN//MCH/SMN/AST/WHM.

At the 80 percentile, DRK does 0.59S; WAR 0.627S, SAM 1S, NIN 0.925S, SMN 0.867S, AST 0.254S, and WHM 0.326S.

Since the sum of all 8 members damage is 100%, S is 0.2170 (=1/(0.59+0.627+1+0.925+0.867+0.254+0.326)), which is less than the threshold 0.2227.

So at the 80 percentile, Trick Attack dps increase is always bigger than the dps difference between SAM and NIN.

Data for the boss dps (Susano & Lakshmi) at the 90 percentile: https://www.fflogs.com/statistics/15...sdps&class=Any

Doing the same calculations for the 90 percentile with the same aforementioned standard composition:
- S needs to be no more than 0.2012 (=0.0167/(1-0.917)).
- S is 0.1798 (=1/(0.62+0.651+1+0.917+0.851+0.866+0.289+0.369))

So at the 90 percentile in which other members do more dps in term of percentage of raid dps, Trick Attack dps increase further outscales the dps difference between SAM and NIN.

Conclusion: NIN raid dps contribution is higher than SAM raid dps contribution at 80 and 90 percentiles. The same calculations can be done for other percentiles.

[Fyce]

Conclusion: NIN raid dps contribution is higher than SAM raid dps contribution at 80 and 90 percentiles. The same calculations can be done for other percentiles.

Weird, I though I included where I took my data in my post. Oh well.
I took the average numbers pulled out from the Top100 parses regarding individual DPS of both NIN and SAM. The difference was about 350 DPS in favor of SAM. Which means that Trick Attack should do at least 350 DPS from 7 other party members for the NIN to be on par with SAM. For a 1.67% damage increase, this equals to 21k group DPS.
So, my numbers are definitly correct.

What differ is what data we use. And it's true that the difference in DPS between NIN and SAM when looking at specific percentiles doesn't align with what I found.

However, I will say that your assumption greatly depends on the fact that your group must share the same percentile between all players. I only went as far as assuming that SAM and NIN had the same skill level (which is what to assume from both Top100). On my side, I checked several real parses, only to see that player's skill is far from being homogeneous. In fact, quite a high number of parses from Top100 SAM and NIN had these players being atleast 96% perf, while the rest of the group was below 50%.
In these groups, a NIN provides way less overall DPS to the group than the SAM.

And this was my main conclusion, supported by your calculations: for NIN to be on par with SAM, the whole group needs to be top notch. This condition greatly devalues Trick Attack and often leads to the situation I described.

But still, it's hard to compare our results, as we both used very different data and, obviously, assumed that we took the right approach.

[HolyTurtle]

In short, for NIN to be at least at the level of SAM, you need a group that can do 21k DPS. It means that your whole group has to work quite hard to make up for the fact that they bring a NIN instead of a SAM which could've been fine by

Susano needs 14k group DPS to be clearable. As of today, quite a lot of PF groups still hit enrage. So this 21k group DPS is far from being the norm.

Choosing a phenomenon (350 dps difference between SAM and NIN) in the highest performance level (top 100 parses for each job) and a phenomenon in PF (21k raid dps is uncommon) to make an implication on choosing a raid composition in PF is irrational. In PF, the 350 dps difference between SAM and NIN is as uncommon as raid dps more than 21k is. In the top 100 parses, the 350 dps difference between SAM and NIN is as common as raid dps more than 21k is. Do not mix and match conditions from different performance levels to make an implication on a specific performance level.

Furthermore, using only 100 parses to represent the performance of any job is inadequate and misleading. Balancing should not be based on the top 100 parses (not even the 90 or 80 percentile), and choosing a raid composition should not revolve around the top 100 parses either since those parses do not reflect the performance of the majority of raiders.

What differ is what data we use. And it's true that the difference in DPS between NIN and SAM when looking at specific percentiles doesn't align with what I found.

Let me restate my conclusions clearly:
- If the performance of every party member is within 10 percentile of each other, then Trick Attack damage increase outscales the dps difference between SAM and NIN.
- If the performance of SAM and NIN are excessively high while the performance of all other members are low, then the dps difference between SAM and NIN outscales the Trick Attack damage increase.

And this was my main conclusion, supported by your calculations: for NIN to be on par with SAM, the whole group needs to be top notch. This condition greatly devalues Trick Attack and often leads to the situation I described.

This is wrong. At all percentile levels, Trick Attack damage increase outscales the dps difference between SAM and NIN, assuming all party members are within the same percentile. A group does not need to be top notch for NIN to provide more raid dps than SAM; the group members only need to be approximately equal in performance.

[AziraSyuren]

Susano needs 14k group DPS to be clearable. As of today, quite a lot of PF groups still hit enrage. So this 21k group DPS is far from being the norm.

...Uh.

That was high for Heavensward, sure. But Stormblood? No, not really.

PF randoms being garbage is nothing new and not an indicator as to whether or not something is imbalanced.

Being able to carry otherwise poor group DPS is exactly how a selfish DPS class should be designed. There's absolutely nothing wrong with something being the clear best in sub-optimal conditions.