“It’s only preseason!”
This caveat will loom over every New York Jets highlight that emerges in August.
Visit our New York Jets Analytics page to learn the advanced football numbers.
Generally speaking, NFL fans and media are aware that preseason performances should be taken with a grain of salt. With that being said, the NFL still makes the decision to play three of these things every year. So, they must have some degree of value, right?
Each year, we see a few examples of preseason excellence being a sign of things to come. Jets fans will remember that Jermaine Johnson looked like a man amongst boys in the 2023 preseason, signaling the second-year breakout he was about to have.
Jermaine Johnson's performance in the preseason so far is a substantial improvement over where he was during last year's preseason
2022 preseason: 3 pressures on 36 rush snaps (8.3% pressure rate)
2023 preseason: 3 pressures on 15 rush snaps (20.0% pressure rate)
📈
— Michael Nania (@Michael_Nania) August 18, 2023
In just 4 pass-rush snaps, Jermaine Johnson drew a hold and picked up a sack.#Jets pic.twitter.com/7krrE0tHy0
— Michael Nania (@Michael_Nania) August 27, 2023
In 2022, Sauce Gardner drew praise for a pristine three-game preseason run in which he was never targeted. This, of course, has translated to the regular season.
Sauce Gardner wrapped up his first preseason with 0 targets on 24 coverage snaps — the most coverage snaps without a target by any CB in the NFL this preseason.
That doesn't happen on accident.
Gardner put some more sticky coverage reps on tape vs. NYG: pic.twitter.com/uZBST7WpbJ
— Michael Nania (@Michael_Nania) August 30, 2022
So, my question is… exactly how much does the preseason actually matter?
Three years ago, I ran a study to answer this same question. My methodology was to analyze the correlation between players’ preseason Pro Football Focus grades and their regular season Pro Football Focus grades in 2019. (Yes, PFF grades have some flaws, but it’s the only catch-all player-grading metric that can be used to compare players across all positions.)
In that study, I analyzed 12 different position groups. All 12 groups had a positive correlation between preseason grades and regular season grades, which meant players’ regular season grades did tend to improve alongside improvement in their preseason grades. However, the correlation was very weak across all positions, which meant the preseason only mattered a tiny bit.
We can understand the correlation between two sets of data by analyzing the correlation coefficient. Here is a summary of what the number means.
Correlation coefficients range from -1.0 to 1.0. Here is a generally accepted guideline of what they tell us:
- –1.000 A perfect downhill (negative) linear relationship
- –0.700. A strong downhill (negative) linear relationship
- –0.500. A moderate downhill (negative) relationship
- –0.300. A weak downhill (negative) linear relationship
- 0.000: No linear relationship
- 0.300: A weak uphill (positive) linear relationship
- 0.500: A moderate uphill (positive) relationship
- 0.700: A strong uphill (positive) linear relationship
- 1.000: A perfect uphill (positive) linear relationship
For reference, here are the results from the 2019 study:
| Position | Correlation coefficient |
|---|---|
| All Defense | 0.110 |
| All Offense | 0.089 |
| T | 0.242 |
| QB | 0.199 |
| C | 0.185 |
| CB | 0.166 |
| LB | 0.123 |
| RB | 0.088 |
| S | 0.073 |
| G | 0.065 |
| IDL | 0.049 |
| WR | 0.044 |
| TE | 0.025 |
| EDGE | 0.019 |
As you can see, not a single position could even get as far as the 0.300 threshold, which is generally considered “a weak uphill linear relationship.” The correlations were positive across the board, meaning it was better for a player’s regular season outlook if they succeeded in the preseason than if they struggled (rather than their preseason performance having zero meaning whatsoever), but the impact was marginal.
It also varied greatly by position. For edge rushers, preseason grades were essentially meaningless, but for tackles, the correlation was nearly at the 0.300 mark.
I ran this study again for the 2023 season to see if anything changed. The study analyzed 364 offensive players who played at least 50 preseason snaps and one regular season game, along with 434 defensive players who played at least 50 preseason snaps and one regular season game. Here are the results:
| Position | Correlation coefficient |
|---|---|
| All Defense | 0.155 |
| All Offense | 0.176 |
| IDL | 0.410 |
| EDGE | 0.330 |
| T | 0.295 |
| G | 0.206 |
| RB | 0.151 |
| TE | 0.133 |
| QB | 0.109 |
| C | 0.034 |
| S | 0.015 |
| CB | -0.039 |
| WR | -0.060 |
| LB | -0.090 |
In 2023, the overall strength of the correlation improved slightly compared to 2019, although not by much. It still was not close to 0.300 in either phase.
The more important takeaway was the increased variance between position groups. Three position groups (CB, WR, LB) had a negative correlation, meaning players tended to perform worse in the regular season as their preseason grades improved. On the contrary, two position groups (IDL, EDGE) had a strong enough correlation to clear the 0.300 threshold, while tackles fell just a hair short.
One very noticeable trend in 2023 was that trench players tended to have a substantially stronger correlation than perimeter and off-ball players. The four most correlative positions were interior defensive line, edge rusher, tackle, and guard. The four least correlative positions were safety, cornerback, wide receiver, and linebacker.
However, it is crucial to point out that these results are much different than what we saw in 2019, which makes it hard to place too much stock into them. Interior defensive linemen and edge rushers had correlation coefficients of almost zero in 2019. Did the league really change that much in four years to the point where the preseason is now a solid predictive tool for those two specific positions? That’s doubtful. It goes to show that these numbers can vary wildly from year to year, proving the preseason’s unreliability as a predictive tool.
Ultimately, the main takeaway is that preseason production matters very little when it comes to predicting how a player will perform in the ensuing regular season. While it seems likely that performing well in the preseason is marginally preferable for a player’s future odds of success, it is difficult to identify any trends that definitively show there is a significant positive correlation between preseason grades and regular season grades.
The preseason exists to settle roster battles – nothing more. If you use a bubble player’s preseason resume to predict their future outlook, your odds of being correct are hardly better than flipping a coin.
