Projecting Trevor Lawrence, Justin Fields and the 2021 Quarterback Class

With two years of efficiency and grading data to judge, we can now make comparisons to historical quarterback classes, and model out their likely future values

Now that the last remaining quarterback from the 2021 class has been eliminated from the playoffs, we can step back and make an initial assessment on how one of the deepest prospect groups in history has performed over their first two seasons.

The first thing I like to do before getting into the more opaque modeling is to visualize the data, and get a feel of how the current crop of players compares to historical analogs. My two favorite metrics for quarterback play are expected points added (EPA) per play and PFF offensive grades. The former gives the most accurate and relevant translation of play, by looking at exactly the expected points before and after plays the quarterback was directly involved in (dropbacks or quarterback designed runs). This points-based metric can tell us more accurately than our intuition whether a sack of five yards on 1st & 10 was more of less damaging to the team’s success than an incompletion on 3rd & 3.

EPA lacks in its ability to separate quarterbacks’ contributions from that of his teammates and opposing defenses. This isn’t a unique flaw of EPA, literally every player-level stat is a “team stat”, or one that the defense influences. That’s why combining EPA with PFF grading is so powerful. The latter attempts to judge the quarterback only on his individual contribution, by watching every play and assigning a play-level grade of -2 to +2, which translates up to a 0-100 game or season grade. While grading gives you specificity of contribution, it lacks somewhat in weighing and calibrating contributions, which is where EPA adds its superior valuation accuracy.

EPA PER PLAY EFFICIENCY

We’ll start by looking at the first- and second-year results for EPA per play and PFF grading separately. The plot below focuses only on EPA per play and includes quarterbacks who had at least 200 dropbacks in both their rookie and second seasons. It can be a bit crowded and difficult to see the results, so I highlighted the 2021 class with red, bolded names. Each player data point is represented by their team logo.

The EPA numbers here are era adjusted, as efficiency has gone up substantially over the past 20 years. I also added an “adjustment” for two quarterbacks whose rookie seasons were so inefficient that they would have skewed the entire plot: Jared Goff and Alex Smith. I moved them to -0.2 EPA per play from -0.32 and -0.40, respectively. Important point is they were historically bad as rookies, and that’s still reflected above.