If you have any questions about what we do on the site, don’t hesitate to contact us, but first, have a read through this FAQ – we just might have been asked them before, and you might find the answer you’re looking for below!
1) Why is Player X rated so low or so high in your rankings?
This is easily the most asked question we get and so it is going to get pride of place at the top of our FAQ. The first thing to note about our “rankings” is that they are not rankings in the traditional sense. We don’t measure talent; we’re not telling you who the best players are. Our rankings are more of a performance evaluation, and a reflection of how efficiently a player made plays in the time he was on the field.
The second thing to take into consideration is that Player X really might be that good or that bad. Just because a player made the Pro Bowl last season doesn’t mean he played well — in fact, he may have played very poorly. NFL coaches don’t watch nearly the amount of tape on other players (certainly not ones they’re not going to play that season) as conventional wisdom would have you believe. When people watch a football game they tend to create a mental highlight reel, remembering a few of the best and a few of the worst plays, and then (based on those few) guessing the overall performance. That is rarely accurate. We are able to go back after the games and grade people on a play-by-play basis, totting up the numbers as we go. Just because a player was largely anonymous during the game doesn’t mean he played poorly. If he edged out a win in every block he made, he could well have had an excellent game overall, even if he didn’t leave anybody on the turf from a pancake block.
Similarly, just because Player X isn’t a household name, don’t discount the possibility that he is having an excellent season that is passing by the notice of most people.
A good case study from the 2009 season is 49ers TE Vernon Davis. Listen to anybody talk about him and they’ll point to his league-leading TD count and proclaim him an All-Pro TE.
Here are some of the things we would say about Vernon Davis and how he rates in our system:
• Blocking plays a big part in our grade:
With the size/strength ratio that man has there’s no way on earth that Davis couldn’t block effectively if he used good technique and put in the effort. When we initially graded him in some of the early games of the 2008 season it was apparent he had all the skills but just didn’t seem to care at times; he would make some difficult blocks but then miss a whole host of easy ones. When coach Mike Singletary took over and had “words,” it seemed to light a fire under Davis and for the last half of the season he was very good indeed. We expected this to continue but, after he began getting plaudits for his receiving, his blocking seemed to regress almost to a worse position than previously. Regarding what coaches and writers say on the matter, we take it with a pinch of salt. This is because coaches hardly ever tell the truth and most writers don’t have the time to analyze every player in detail to come to an accurate conclusion.
• Penalties Matter:
Most people will forgive players these but when you see the stats for how a penalty effects the chance of a drive progressing, why should they? Davis gave up more penalties than all but three players in ’09 (all of them offensive tackles) and twice as many as most TEs. This has to be taken into account.
• Pure Receiving:
We don’t grade people on how many yards they get because in a lot of cases that’s irrelevant. Many of TEs could probably get 1,000 yards if you threw to them enough. Our system is not impressed by dropped passes (13 is a ridiculous number and again league-leading), yardage that doesn’t pick up a first down or at least come close and yardage in garbage time. Frankly when you compare him to Antonio Gates, Tony Gonzalez or Dallas Clark as a receiver he’s wasn’t in the same zip code; he may well be soon but not last year. There are maybe another half dozen TEs that could pick up the same type of numbers as Davis playing in his place but would do it without screwing up as much as he did in ‘09.
• Overall Rankings:
At the moment these simply weigh everything equally. We believe blocking is a vital part of a TEs responsibility in the NFL and so players that struggle with this will be brought down in the overall rankings. We acknowledge that not everybody believes the same thing and so there is nothing to stop people sorting the columns by the attribute that they believe holds the most weight and viewing them that way.
2) Why do you have only one “L” in your Twitter account @ProFootbalFocus?
As you’ve seen, this has changed and we’re now found on Twitter with the handle: @PFF, but in case you’re feeling nostalgic, here’s the old response to a very popular question: Twitter has a character limit on its usernames. That character limit is one character fewer than ProFootballFocus, our ideal account name. We decided to drop one “L” and hope nobody noticed.
3) How do you know exactly what a player’s job is on any given play and whether negatives were his fault?
This question obviously changes depending on the position and the specific play. Much of our methodology can be found in our explanation of our grading, but we will go into a quick summary:
• Pass Protection
This is a measure of how much total pressure a player gives away during a game. It isn’t a simple formula like -1.5 for a sack, -1 for a hit, as time taken to get the pressure is also very important. This is then normalized by adding a small positive factor for every drop-back he played. It is usually simple enough to determine what a player’s assignment is in pass protection, as we have the benefit of being able to watch one player closely and specifically multiple times. It is possible that a blocker is directed by the quarterback to take a specific man, resulting in a pass rusher that appears to be his responsibility getting a pressure and us grading that blocker down. This is an inherent inaccuracy in the grading, but despite this potential inaccuracy we have had NFL sources, including Bengals OG Evan Mathis, confirm that our grades are accurate, and reflect closely what they receive as internal feedback. There may be an inherent margin of error in what we do, but it is still more accurate than anything outside of a team meeting room.
• Run Blocking
Again, it is rare not to be able to determine where a run was supposed to go, and what a blocker’s assignment was on a particular play. Players don’t make the kind of mental lapses often that would see them going to a totally different place than where they should have been. If they did they wouldn’t be in the league long. If a player attempts a block on somebody, whether they win the encounter or lose it, it was almost certainly their assignment on the play. This is obviously not easy to pick up live and in real time, but again we have the benefit of being able to go back retrospectively and watch multiple replays of a play developing to get the information.
Assignments in coverage can be a more difficult matter, specifically in the middle of the field when applied to LBs and safeties. Because when a TE, HB or WR finds a seam, the determination of whose responsibility that is becomes trickier. Some other analysts have decided to apportion the yardage out amongst the players involved, but frankly we don’t believe this is worth the effort. In addition, if it’s a touchdown, do you award half a TD? Because of the inherent issues with the statistics mentioned above we just go for the simple approach and take the closest player when the ball is thrown, understanding inaccuracy is built in.
4) Why are your numbers different to the ones I see elsewhere?
These differences can usually be explained in one of a few ways:
• We record our figures differently in a couple of instances: we don’t like (and won’t use) half sacks. We don’t see why a player should be penalized (or an OL rewarded) for coincidentally getting to the QB as the same time as someone else. We also don’t record the same play in multiple categories for players (more on that later), so a sack is not also a hit, and a pressure, and a tackle. An interception is not also a pass defensed, etc.
• When we have a chance to go back and check, things like our tackle statistics are far more accurate as we don’t have to produce them immediately. We are also far more consistent in the way we apply stats like assists or run position. We have just a couple of people applying rigid rules to the collection of this data, whereas other sources use one person per team, operating in real time, without the benefit of replays, with seemingly no guidelines on specifics. In the vast majority of occurrences, our data is more accurate because of this.
• Errors on our part … As much as we strive for perfection, human error can sometimes creep in. If you see something you’re not sure of just let us know; we’re happy to tell you why there’s a difference between the official stats and ours.
5) How do some players have more sacks than QB hits?
We treat our stats as unique. Unlike the official stats, we don’t combine statistics into multiple fields and end up counting them several times in the final reckoning. A sack on the QB does not also go down as a hit on the QB and a pressure on the QB. For example in 2009, Cardinals OLB Bertrand Berry notched 8 sacks (including playoff games), but only 2 tackles and no hits on the QB. Obviously each one of those sacks was also a tackle behind the line of scrimmage and also contact on the QB, but we don’t count the same play on multiple occasions.
6) How do you determine defensive players’ passer rating against?
It’s as simple as it seems. The passer rating against is simply the QB rating for throws made into that player’s coverage.
7) When teams are in non-traditional sets (nickel, two down linemen, etc), how do you determine what position someone plays?
In order to record our Player Participation data we have a set list of positions. We’ve been tweaking since we began, and now every player falls into one of the positions on our list. In essence, if there are four down linemen (simply players lined up on the line of scrimmage with their hand in the ground) then those four players will be labelled DRE, DRT, DLT and DLE, regardless of the fact that they might be linebackers by trade. If there are three of fewer down linemen, then those players will be labelled RE, NT and LE, or perhaps just RE and LE (many teams drop the NT from their D in nickel situations).
A frequent question surrounding this is why some players appear in positions that you might not expect they usually play. Cameron Wake, for example, appears as a 4-3 DE even though we all know the Dolphins played 3-4 in 2009. The reason for this is that Wake typically only appeared in the Dolphins’ pass-rushing subsets and almost always with his hand in the ground as a defensive end.
The Jets are also a good example:
The reason we’ve put their players as being in a 4-3 is because that’s what the Jets now play in the main. It’s almost identical to what the Ravens ran in ’08 and we had a fair few questions on that too. Eric Mangini took his “pure” 3-4 with him to Cleveland when he departed.
• Of total defensive formations, the Jets played a 4-3 defense 25 percent of the time and 3-4 just under 14 percent
• In base defense (four DBs), the Jets showed a 4-3 look 65 percent of the time and 3-4 formation 35 percent of the time
• In no single game did the Jets run a 4-3 more than a 3-4, however against the Patriots in Week 2, neither was used at all
8 ) Frequently, you have two players listed as free safety on the same play. Why?
Our distinction between free and strong safety is based on a distance from the line of scrimmage. If a player is lined up more than 8 yards away from the line of scrimmage he is listed as a FS. This means that if both safeties are further than 8 yards off the line (as often happens in the cover 2, for example), they will both be listed as FS. It is equally possible that both players could be listed as a SS if they are both closer to the line.
9) What determines if a receiver is “thrown at?” How about uncatchable balls, passes thrown away, etc?
Unlike the official NFL scorers who seem determined to register every passing attempt as a target for a specific receiver, we don’t list any pass that is obviously thrown away (to avoid pressure for example). If a ball is aimed at a specific receiver, it is recorded as a ball “thrown at” that receiver. In the case of uncatchable balls, they will still be marked as a ball thrown at a specific receiver
10) Why does it take so long to get all of the week’s available statistics and data up?
Simply put, it’s a case of resources and the time it takes to complete our analysis. To get a single game completed takes around 16 hours of work, and we do not have a limitless staff of analysts and people to do the player participation. Getting a 16-game week completed requires around 250 hours of work, so it is unfortunately not possible to get them all up on Monday morning like we would all like. Rest assured, though, that the games go up as soon as we can get them completed.
11) What constitutes a “stop” or “offensive failure?”
It’s when an offense fails to get:
40 percent of the required yards for another first down on first down
60 percent of the required yards for a first down on second down
A first down on third or fourth down
12) How do you normalize your results?
Once we’ve got the raw grades we could leave it there, but this would lead to a number of problems. For example, as an offensive lineman can only be negatively graded in pass protection, the perfect score in the raw data is zero. However, what if a lineman plays half the number of passing plays of another guy and they both score zero? What allows you to understand the second has done the better job? This is where Player Participation comes in: To fully understand how a player has performed, we need to know how many plays he’s participated in and what role he performed.
So when we look at, say, a TE, we need to know how many plays he spent out in pass routes, how many times he blocked for the run and how many times he stayed in to block for the pass. To this number we then apply a normalization factor to set the AVERAGE player in that facet of the game to zero. To simplify, if you go to the by position tab and select TE, the average grade for run blocking is zero, the average grade for pass blocking is zero and the average grade for pass receiving is zero.
13) Do you use coaches’ film, or TV broadcasts for your analysis? How can you see what players are lined up where?
We use TV broadcasts for both the analysis and the player participation data. We’ve been asked a lot how therefore is it possible to determine who was on the field in every position through TV pictures. Surely that just isn’t possible? Well, just because something is not easy doesn’t mean it is impossible. Firstly, the high definition picture from which we work is excellent, and that gives us the best starting point outside of coaches’ film, which isn’t attainable outside of NFL offices.
14) How accurate are your player participation statistics?
Player participation is very difficult to begin with. Sometimes you can’t make out player numbers, sometimes you don’t get a perfect picture of a player, and you only have a strange angle from which to make a determination. But we have the benefit of being able to watch plays multiple times, pause the picture, and generally gain from doing this after the fact, something that somebody watching at home wouldn’t be able to do simply watching the game live. What you also learn quickly when doing player participation is that there are a massive numbers of players that can be identified by a physical or uniform characteristic that isn’t their number or name on the back. For example Tory Holt has a unique stance that can be immediately identified at a glance when he lines up. Does anybody really need to see Domata Peko’s number to identify him on the line of scrimmage? Do we really need to be able to see the #43 to be able to correctly identify Troy Polamalu in the Pittsburgh defensive backfield? These are obviously some of the more obvious and glaring examples, but the majority of players on the field can be identified by some physical characteristic such as these, even if they are more subtle and esoteric. We aren’t perfect, and there is a small element of human error in the process, but we have had our data verified by two NFL teams at over 99 percent accurate. Over the past few seasons we can count on one hand the plays where it has been physically impossible to identify a player from the TV footage – and in those cases, it is highly unlikely to have been a player substituted in for a single play and then removed.
Also, with the help of the two NFL teams we were able to determine 0.9-1% of our inaccuracy came from simple error in notation so this year we are going to do many more games where two people do the work, all but eliminating any problems in these games.
15) How accurate are your grades? Would they change depending on whose watching film?
Our grading is done by only a three analysts. This is to ensure and maintain the consistency of the grading. We have constantly updated and codified our methodology so that there is very little room for subjectivity and our analysts are in constant contact with one another to ensure their grading is consistent.
16) Can you compare OT, OG and C grades, or are they on different scales?
All players on the offensive line are graded using the same methodology, but the final numbers cannot be compared directly because the normalization factors for each position differ radically. This means that a player’s raw, un-normalized grade will have a much different figure added or subtracted to it to normalize it depending on the position that player plays.
17) Why do you subtract for penalties?
As much as some people wish it wasn’t so, penalties are part of the game, and deserve to be recognized. They can also say a lot about the performance of a player who would otherwise look far better if they were not accounted for. For example, a player could perform reasonably well in pass protection, but on occasion he is beaten, tackles his man to the ground, earns holding penalties, or is caught jumping offside in order to try and get a jump on the pass rusher. If we didn’t record penalties negatively, that player would score significantly higher than he merited.
18) How much credit does a QB get for a well-thrown pass on a screen that goes 80 yards?
Quarterback play is one of the areas our grading gets the most criticism, but it is also one of the areas where our brand of “intelligent statistics” can really shine.
Consider the following passes:
• Pass 1 is a short screen to a running back. Between good blocking on the perimeter and a couple of eye-popping moves, the HB takes it 98 yards to the house.
• Pass 2 is a ball right to a trailing cornerback. The corner, surprised by how horrid the ball is, bobbles it before the wide receiver steals it out of his hands, tosses him to the ground, and makes a 60 yard gain out of it.
• Pass 3 is a 3-yard bullet into tight coverage on fourth and 2. The pass extends the drive and enables a game-winning FG to be made.
• Pass 4 is a deep ball in the face of a blitz. The QB looked off the safety and put the ball in the receiver’s hands, a receiver who unfortunately looked the ball right into his face cage and ended up sprawled on the floor. Had he caught it, the receiver would have been able to walk into the end zone.
Under any conventional statistics the QB grades out extremely well in the first two instances, but one of them is a terrible pass and one of them is merely a standard completion you would expect any warm body to make. By contrast, the QB gets nothing for Pass 4, and just a completion, 3 yards, and some post-game discussion from Pass 3. Under our grading system we are able to reward and punish the QB accordingly in addition to the statistics we are taking down to reflect the quality of those passes, both positively and negatively.
Fundamentally we believe this happens, to some degree, on every play. We feel players cannot simply be judged on the result of the play unless it is contextualised. This is at the heart of what we do and why we completely unique.
19) Do you give corners extra credit for not being thrown at?
We do give defenders a small normalization boost for every pass in coverage they are not thrown at. However, we do accept this is an area of weakness because there is no reliable way to quantify this. A corner might be shutting down his man every play, so the quarterback is throwing elsewhere, or he might just be playing in a defensive backfield with much more inviting targets for an offense to exploit. However we do only grade on passes thrown into their coverage or on their play against the run game.
20) How do you account for cornerbacks in man or zone coverage?
The situation with CBs is fairly straightforward. The CB almost always has initial responsibility for the man across from them; he starts off in their zone and for all intents and purposes is in man coverage until such time as he’s handed off to a LB (on a crossing route) or a safety (on a deeper route). As soon as he has handed the guy off (and that point is normally very easy to judge), he no longer has responsibility. So the CB stats are pretty accurate, at least from a standpoint of responsibility, if not performance.
Cornerback is another area where simple statistics often come up short, but our “intelligent grades” can pick up the slack:
• A team runs a WR screen at a CB who is in man coverage. The CB is double-teamed and can do nothing, but because of missed tackles by other players the screen goes for 80 yards and a TD. The CB was clearly in coverage and hence the yardage goes against him.
• A CB is beaten badly for a reception but the WR then drops the ball.
• Another CB is beaten badly and gives up a 70-yard TD by peeking in the backfield but the play is called back on a holding call.
In the first of these the CB has done nothing wrong and is penalised by having 80 yards and a TD logged against him. In the latter two the CB has been poor but is rewarded with incomplete passes to his credit. Obviously in our grading the CB would not be penalized in the first but marked down significantly in the others.
21) Why can’t I find player X in your rankings?
There could be a couple of reasons you can’t find the player you are looking for:
• He is listed under a different position. Players like Cameron Wake may be listed at OLB in the Dolphins’ 3-4 scheme, but he shows up as a DE for us because he rushed almost exclusively from that position in the Dolphins’ 3-3-5 or 3-2-6 packages.
• There is a filter on each position ranking that excludes people with less than a certain number of snaps. You can bring him in by using the drop-down and selecting “all”
22) Do you take into account *INSERT SPECIAL SITUATION HERE* (weather, double-teams, injuries, scheme, etc.)?
The ratings don’t take into account the opposition, but we do give differing grades based on the type of pressure and when the pressure is created.
For example, a 4-second coverage sack in garbage time of a 40-0 blowout may net the defender 0.5 or even 0, where a immediate hit that forces an incompletion to win a game may net 1.5-2.0. We have an extremely detailed set of guidelines that we use for this but the analyst always has the final say because situations vary and, hard as you try, you will never be able to predict every possible situation.
Double-teams are graded differently because a double-team is an expected victory for the two-man side. If the defender manages to split the double team and make a play, he will be rewarded more than if he simply beat one man. Similarly, the blockers (or at least one of them) will be downgraded more severely.
Other things such as weather, scheme, injuries etc. we usually don’t deal with. We just look at individual players’ performances on each play in isolation. We can’t attempt to quantify how much a scheme or injury might be affecting his play — that is for our readers and users to decide for themselves. The bottom line in all our data is that it isn’t the be-all and end-all of football analysis. What we want more than anything is for people to take our information and combine it with their own knowledge and information to come to their own conclusions.