From 2017-2019, he's a role to solid player model on your fangraph chart, but all but 15 pitchers a year (48 in the last 3 years) had an fWAR of over 1.75. I am not denying what that chart says. Only 38 RP'ers a year are called "solid" or better by the chart out of over 200 with 30+ IP. "Solid is the middle designation of their 7 categories, so only15 are above average, only 23 are average and a whopping 200 plus are below average, if we go by your reasoning.
My point is you are trying to say the chart says he is an average or below average RP'er.
The chart does not show that or even imply that.
If you issue is the English language and not understanding what the context surrounding the word average means, then I see how you got from point A to point B.
In English, when you say someone one is average, it means when compared to others in the group you named.
Had you named the group "all players," then a strong argument could be made that Barnes is below average in value.
You named the group RP'ers, which means you were comparing Barnes to other RP'ers ONLY. So an average RP'er would not be the same as an average player in value.
Here's how you chart breaks down by actual number of RP'ers in each group from 2017-2019 (divided each by 3 to get the average number of pitchers that fall into each category, each year):
(2) 3.5 or higher WAR MVP
(3) 3.5 to 3.3 WAR Super Stars
(10) 3.3 to 2.3 WAR All Star
(34) 2.3 to 1.75 WAR Good Player
(70) 1.75 to 1.1 WAR Solid Player
(180) 1.1 to 0.5 WAR Role Player
(200+) 0.1 to 0.5 WAR Scrub
(200)+ at 0 or below (30 IP+)
Getting the average for each season from 2017-2019:
1 super star
3 all star (we know more than 4 RP'er made the all star games)
11 good
23 solid
60 role
70 scrub
70+ worse than scrub
Barnes placed 21st in total fWAR in that time frame with a 3.5 fWAR. That comes to 1.17 per year or about a 1.2, but go ahead an take his career numbers up to 2021 but not including it, and call him a 3.6 over the 5 years for a 0.7 average year fWAR) making him a "role player" not a "solid player" had we used the 1.17 score.
As a role RP'er he would have over 140 RP'ers worse than him and 38 better and about 59 the same. No way is that "average."
As a solid RP'er, he'd have 15 better, 200+ worse and about 22 the same- way, way, way better than average.
He's not an average RP'er. He's far from "below average" as you wanted to call him.
You keep saying my method is flawed, but we are talking about AVERAGE RELIF PITCHERS, so we compare him to only RP'ers. That's how it works.
I'm not cherry-picking stats. I'm using the data that compares RP'ers to each others using your preferred stat, fWAR (not mine).