Pro Football Focus is imperfect. They're trying to bring sense to the nonsense that is pro football. It's impossible to tell the deviation between a good and a bad game, as most players have moments that are both. But other times the water carrying that PFF does is stupidly biased. Keep in mind that PFF's grades are not a standard view of "good vs. bad", it is their definition of "good vs. bad", so they measure and weigh things based on how they feel about specific aspects of the game.
It's why Myles Garrett constantly gets 90s for his games, even though he's bullied off the ball whenever the opposing team runs it. See, they believe that the "pass rush win rate", which isn't a thing, is somehow impactful. Because they believe that if a defensive player is within X-distance of a quarterback, then that quarterback must be affected by the player's presence. It doesn't matter if that defender didn't do anything tangible, he still made an impact.
Some how.
No other position gets the allowance of an 'almost-stat'. "Yeah, that Nico Collins had six catches, and two almost catches". "You mean drops?" See how silly that sounds? We saw how little impact Garrett made against the Cleveland Browns in the playoffs; that pass-rush win rate was not a factor. As it never is. Yet, Garrett was still praised, getting a rating of 72.3 on the day, despite the Browns' defense getting torn to shreds.
Yet, C.J. Stroud molly whops' the hell out of the Cleveland Browns, and he only gets a 77.8 from PFF, despite his massive playoff game. Make it make sense.
This is all said to say this; PFF is not the gospel so many people think it is. It's a website that only has the leverage it has because Chris Collinsworth owns it, and has gotten it integrated into his broadcasts. No, there's no conflict of interest there (*sarcasm intensifies*). But still, it's a good reference tool.
So we're going to use that tool today and find out which six of the Texans are having a good postseason so far and which six aren't. We're going to keep it to a 15-snap minimum this time around, so the lowest of the low or highest of the high may not be included if they didn't hit a certain threshold.