Quote Originally Posted by Dzian View Post
The issue is that a good portion of what makes a difference between one percentile and another isn't the player it's the party... You can take a take for example the 100th percentile player and put him in a group of randoms and his performance will be no where even close to that 100th percentile.
Firstly let me ask - do you use FFLogs? Do you know how to read it? All of its functions?

This is 100% inaccurate. It will be lower for sure, but it won't be "no where even close". I had plenty of purples where I had a team of 10th percentilers.

There's so many things that fflogs doesn't tell you that can have a significant impact on your performance. And a lot of them can be as simple as which direction is the boss facing or where is the boss being tanked?. A badly positioned boss might mean you have 10% less uptime. A boss thats constantly being turned might cost you a significant amount of positional bonuses.
This is also 100% incorrect. FFlogs can tell you all of that if you know where to look and how to read the logs.

A party member that's stood right next to you might then force you to move and lose cast time where if he was in a better position you wouldn't have to move... lost gcds..

Badly stacked ground aoes can make the difference between taking 2 steps to get out of or taking 7 or 8.. lost gcds.

That player with the stack marker forcing everyone to move away from the boss to stack on him instead of him running to the group on the boss.. more lost gcds

There are quite literally an almost endless number of things that can really impact performance none of which are ever represented in fflogs...
Every single one of these is captured in the logs and easily viewable.

Then you have things like other party members and how well they use there buffs. Is chain start being applies with foe and trick attack muliplying the damage boosts quite heavily or are they all being used sporadically at random intervals.
Easily identifiable in the logs.

All of these things might slash your performance in half and there's absolutely nothing you can do to control it.
This is also 100% inaccurate. It would not take a 6.5k DPS MNK and make him do 3.25k DPS. Please refrain from spreading misinformation.

Player percentiles are essentially worthless. Because it's generally the entire group that makes that number what it is. You could take a 10th percentile player put hi in a 100th percentile group and guarantee he'll do better even if he plays exactly the same....
Just because you don't know how to utilize a tool doesn't make it useless.

Quote Originally Posted by Ghishlain View Post
I do agree that composition and skill with affect the percentile a player can reach, but I do strongly disagree with the statement that looking at a percentile is completely worthless too.

It's up to the individuals looking at the logs to see where they can improve and where they're doing well. A percentile gives that player two things: (1) a range of where they are in terms of skill level with the rest of the player base in the same content and (2) gives them a metric / goalpost to aim for.

Also, a 80th percentile player playing in a group that is a 20th percentile skill won't suddenly make their DPS bottom out at the 20th percentile. It'll dip, yes, but I don't believe it'll dip further than a +/-10 percentile range. That's still a sizable gap but it doesn't make things worthless. A highly skilled player will be able to make non-optimal conditions work for them and can adapt to optimize for the conditions they've been given.
Bingo.