Tuesday, November 12, 2013

Pro Football Focus Ranking Numbers Reveal No Insight

Peyton nor Demaryius not listed as top performers for Week 10

Had to post some remarks about this article write up and these pointless rankings/numbers that they attach to players that reveal nothing about how a team played in a team game and the caliber of competition they played, let alone a solid win or loss from the week. Look at some of the teams these players are on and their results from week 10. They have some great aspects to how they get to these numbers and analyze and player by every single play they have, but on the flipside, they are not insightful when it relates to the broad spectrum of how the league's teams, and its players are executing as a whole. Particularly, in relation, to wins and losses for a team game.

We are not really saying that we or anyone should not be a big fan of +Pro Football Focus (PFF) articles that are based on these ratings. In fact, PFF has some very insightful stuff on their site that we love to read. Their ratings system though, we are not buying. For one, how do they come to give a rating and in reality, we feel it is just to put a number on each position just for the sake of it. Instead of doing the old school way of describing a player's caliber of play as extremely great, good, so-so, not so good, and/or pretty bad. 

Plenty aware that a lot of work goes behind reaching these numbers and that they are probably pretty solid. The performance of others in football relies on the 10 other guys next to you. We prefer to look at the teams wins & losses and regular stats with a bit of modernization of some of the stats. Like yards per play, for example. With the up-tempo teams that are featured nowadays, coaches often look at how much yards per play they are giving up in comparison to total yards allowed. One week, for example the +Denver Broncos face the +Chargers, who average these amount of plays per game. The next week, just for arguments sake, they face the Patriots who average plays/game. Based on getting off that many more plays per game, the Patriots should obviously average more yards of offense per game and more than likely more yards given up on defense since they are more plays overall in their games. 

It is very similar to +ESPN constant attempt to change the QB ratings/stats to TotalQBR, but that is flawed right off the bat when looking into it because they give a higher rating to QBs that are able to come back from a deficit. The Broncos, don't often have a deficit, so Peyton Manning most likely would not get a higher Total QBR then he really should over other QB's. In the event a QB leads the whole game despite stellar performance and maintaining that lead to a win, it counts as a negative in comparison to a QB that comes back from behind and gets a win. That is significantly flawed. A QB that jumps out to a big lead each and every week is not rated as high because they are that good that a deficit is not common place. Factoring in numbers of opponents players is contingent on them, etc. All in all, its just for computer rankings sake, instead of just looking at the original statistics and above all, point differential, takeaway margin, field position, 3rd down efficiency on both sides, and redzone efficiency on both sides means more to us than a number that some algorithm crunched insides of multiple other algorithms is not as significant to Broncos CR3W. In interviews that we have listened to several +Pac-12 Conference coaches agree, the +University of Utah football coach Kyle Whittingham looks at avg./yards per play that his consistently solid defenses give up, and if it is 5 yards or less then they did their job. Many of these coaches have had NFL experience as well. PFF needs to look at numbers like this and above all, factor in wins and losses, and ding teams that have to come from behind and got a deficit to begin with. As well as teams they maintain and hold leads consistently. That said, we don't really plan on using PFF's numbers as a point of reference very often.

Aside from that, these ratings appear to be much more based on how badly the guy on one side plays against another or in layman terms, how bad the opponent is. RGIII is Honorable Mention? He lost at Minnesota, Peyton Manning through for over 300 yards and 4 TDs with a better completion percentage. What is beyond understanding as well, is how is Julius Thomas not above either listed and Demaryius Thomas above any WR from week 10? Brandon Marshall had a great game, but he lost to +Detroit Lions at home and did not have a Demaryius caliber game and many would concur. Beybey Thomas is only an honorable mention?! Our point is, maybe in future don't take PFF's ratings into a discussion as the only piece of evidence or pudding. Their numbers, in the human brain, do not add up sensibly. The Denver Broncos linebackers were really good against San Diego Chargers. We really felt Chris Harris Jr. and Rahim Moore (especially Moore) were exceptional as well as key situations. 3rd & 16 with just around 4 minutes remaining in the game with a pass deflected and forcing the +Chargers to punt the ball away. Look at the clip below. Example after example. A +Oakland Raiders center against the +New York Giants. Raiders lost the game by the way. +Buffalo Bills got trounced at +Pittsburgh Steelers. We find it odd that many so-called pundits and experts use PFF for references in regards to quality of play. The +Carolina Panthers defensive line against the +San Francisco 49ers are not acknowledged or 'honorable mention(ed).' Very skewed numbers overall. Great to see Von Miller in there. Still, we find another example, Terrance Knighton is playing light out, where is he? Offensive lineman can't get a good score on pass protection? So a run oriented team is likely to get acclaim by PFF over a pass oriented team for offensive line play.
Screenshot of PFF Week 10 NFL
Team based on their (PFF) Analysis Numbers
Perhaps in the world of wizards and Mr. Harry Potter, but this is the world of Muggles. Apologies PFF, not quidditch here. Happy to wager Terrance Knighton's level of play as of late along with Chris Harris Jr. and Dominique Rodgers-Cromartie (DRC) against the players' performances, listed. +Baltimore Ravens DB's obviously faced Andy Dalton, the bad version, which influenced their rating more than anything else. Somehow, take into account what the opposing player did in comparison to what he should have done and rate his expected caliber of play into that factor. How about Hammer (Wesley) Woodyard and Danny Trevathan combining for 22 tackles against the ball control, deliberate +Chargers run game? Teams like +Tennessee Titans who dropped a loss to winless +Jacksonville Jaguars. Our interpretation is the team that plays a QB with a low pass completion rate, QB that plays well in a deficit, an O-Line is rewarded for a good running game over a good passing game. Not too mention they listed a +Dallas Cowboys player, they got manhandled by the +New Orleans Saints