I want to write about an approach to analyzing football that I haven’t seen yet. (It might exist, but I’m not aware of it.) The analysis would look at the sequence of plays to identify the distribution of positive and negative plays. (Note: Plays aren’t so binary, but tend to be more or less positive or negative. More on this later.) I think this can be important and revealing way of examining football–an approach that can reveal things that are hidden in statistics, especially the ones that are aggregates. To give a good example of this, think about the QB’s performance in relationship to the quality of pass protection. An overall rating of the pass protection, the number of sacks or pressures don’t tell a complete story in my view. What’s missing is the way ratio of good pass protection and bad pass protection and how this is spread out throughout the game. Additionally, as I alluded to earlier, pass protection isn’t simply good or bad, but falls between a range–e.g., terrible, mediocre, good, excellent. Several criteria go into the quality of the pass protection–e.g., the amount of space the QB has in front of him, the time it takes for the defenders to reach the QB, the number and size of passing lanes, the shape and integrity of the pocket and how long it maintains that shape, etc. My hypothesis is that there is a generally a ratio of good/bad pass protection that is critical to good QB performance. If the OL falls below this ratio, then the chances of a QB performing well will be extremely low. (The WRs/TEs ability to get open is another important variable in this, but I’ll leave that aside.)
Now, the ratio doesn’t really account for the distribution of these plays, so we should actually look at each play and look at the quality of pass protection distributed across all the plays. Are there are a few negative plays clustered together, or are there many spread out evenly. This kind of thing can make a difference in QB performance in my opinion. I think the looking at the distribution can matter from game to game as well. For example, suppose you have three games with consistently good pass protection and one bad game. A good QB can shrug off the effects of that bad game. But let’s say you have three games with shaky pass protection and then two more after that with fairly good pass protection. The three earlier games may have rattled the QB in such a way that they have a hard time gaining confidence and rhythm in the two later games with fairly good pass protection. After watching a lot of games, I see patterns like this. What I’m suggesting is to chart out each play and lay them out so that you can see these patterns.
Here I’m speaking about the relationship between pass protection and QB performance. But we can look at WRs/TEs (the degree to which they get open and how consistently they do this) and the relationship with QB performance. I also think this approach can be helpful when analyzing the impact of the RB on an offense. Doing this can reveal important insights that total rushing yards and efficiency numbers can obscure and mislead. For example, in the recent Cowboys-Seahawks game, I believe Ezekiel Elliot gained more yards and had better run efficiency that Chris Carson. But I think Carson had a bigger, more important impact on his offense, then Elliot. The stats don’t show this, but if you looked at each of the plays, I think this would become more clear.