Analytics in football
For a number of years, analytics have been a very useful tool in the evaluation of baseball players. In recent years, it has also become an important part of football evaluation, but it will never replace good old fashioned scouting. Part of the reason is the difference in the two games. Baseball is a much more individual game. Batter vs. pitcher. Statistical numbers can be very important. Football, on the other hand, is the ultimate team game, and while numbers are useful as a part of the evaluation process, tape study will always be much more important.
Yesterday, on my final day in Mobile at the Senior Bowl, I had breakfast with a friend of many years who, a few months earlier, had left the employ of an NFL team for a situation that was more to his liking. In the course of the discussion, analytics were brought up, and he told me that his former team spends a lot of money on analytics each season. We delved deeper into the discussion and how analytics were used by that organization.
He told me that were used as a supplemental source in the evaluation process, but by no means did they outweigh the scouting reports. I couldn’t agree more. Statistics can be useful, but they will only give you what you ask them to give. You can also use stats to give you strength in an argument, especially when the person you are arguing with doesn’t possess those same stats.
I remember watching a postseason press conference on the NFL Network a few years back when a general manager was asked about the awful play of one his offensive lineman. His answer was “according to (insert analytics service they subscribed to), this player was the fourth best player in the league at his position the final eight games of the season”. He used that answer to get away from answering the question asked. The truth is, if you watched tape, the player was awful, he was a revolving door when it came to giving up quarterback pressures but he didn’t give up sacks. Most coaches will tell you that while a pressure isn’t as good as a sack, it still has a positive effect on the play for the defense. That might not show up in the stats, but it does when viewing tape.
Scouting is an art and not everyone is qualified to do it. Analytics can supplement what a scout/coach sees on tape, but it doesn’t replace what he sees. Analytics is particularly useful when talking about offensive skill positon players. For a receiver, useful stats can be something like drops, yards-after-catch, or the percentage of first downs made on third down receptions. When I used to evaluate receivers, I would keep track of these type of stats while watching tape. Now with the stats right at our fingertips, I don’t have to pay as close attention to those type of things. Still, even with stats, a good report will discuss these areas of the player’s game.
Stats can be particularly useful when evaluating quarterbacks. I have always charted completion percentage on big plays when I watch quarterbacks. Analytics can give me that and many more stats that are useful in my film study. For instance, what is the completion percentage on throws to the right between 10 and 15 yards? What is the quarterback completion percentage when he makes throws on the run? What stats can’t tell me is ball placement. Ball placement, moreso than completion percentage, is a true gauge of a quarterback’s accuracy.
When studying defensive lineman and linebackers, I always kept a chart for every game I evaluated on tackles, assists, missed tackles, pressures, sacks, and forced fumbles. These stats obviously give you an idea of a player’s production, but what they won’t tell you is anything about the key factors that go into evaluating every player.
Football is not only a physical game where strength, power, and competitive nature are a part of every play, but it is also a game of instinct. Regardless of the position, instincts play a key role in a player’s productivity. Non-instinctive players don’t last very long in the National Football League. When scouting a player you have to grade his athleticism, change of direction, balance, strength, and power. Then there are the specifics for each position. For a defensive lineman, it may be initial quickness, key and diagnose, hand use, shed blocks, lateral agility, pursuit and obviously production. While stats can help me with production, it can’t help me with the other specifics.
There is a place for stats and analytics in football evaluation, but if anyone thinks that it can replace exhaustive tape study they are getting. There is an old saying in football “the eye in the sky doesn’t lie”. What analytics does is make the tape study a little easier, but it will never replace the old fashioned way.
When I read that an analytics service rates a certain player as the “fourth best player” at his position, I take it for what it is: an opinion. Unless the person knows exactly what to look for on each play, that opinion doesn’t carry much weight. If that same person tells me that a certain player has a certain level of production in certain situations, than I will buy that 100% of the time. Because it is a stat based on context. There isn’t a stat around that can evaluate instincts, competitive nature, decision making on the move, etc. A person can use stats to help in the evaluation process, but don’t ever try to use stats as the evaluation process. Not in football!
Follow Greg on Twitter @greggabe