Importance of Safety Play
In today’s NFL, offensive minded coaches are manufacturing yards after the catch by getting their speedy receivers in space. This has pushed defensive coordinators back onto their heels and has transformed how we view defensive positional assignments. As a result, the Safety position is asked to do more than ever. Safeties are asked to:
- Be left alone on an island in Cover 1 scenarios
- Play with high pre-snap depth in Cover 2 shell scenarios (in order to keep passes underneath)
- Play in the slot in press coverage scenarios when extra receivers are brought on,
- Play in the box as an additional linebacker, and
- Sometimes be brought in as an additional pass rusher
Chip Kelly notably detailed that they chose their personnel looks based on the calculus of where the safeties were positioned, “If there are two high safeties, mathematically there can only be five defenders in the box… With two high safeties, we should run the ball most of the time. We have five blockers, and they have five defenders.”
One way to stifle this is to have a safety in many pre-snap alignments to avoid certain tendencies (hence the use for a linebacker to drop back into coverage in a Tampa 2 scheme). A player’s ability to fit in many pre-snap looks not only helps the defense but can also boost the draft stock of a prospect.
Safeties are asked to do a lot from day one on campus. The table below highlights the average unique pre-snap alignment each position sees as they progress through college. We can see that safeties and slot defensive backs are asked to fit many different alignments as they progress through their collegiate career.
Viewing Safety Play as a Portfolio
One way to start to evaluate safety play is to evaluate their consistency in performance across many different positional alignments. We first create a matrix per player of their positive play percentage (using PFF grades) across every pre-snap positional alignment (i.e. Slot left side, Free safety right, Outside Linebacker left, etc.). As shown above, the sample size grows for players in different pre-snaps looks as they progress through college showcasing their ability to be malleable and provide impact on defense. For this analysis, we create two separate matrices for Coverage and Run Defense play (we left out pass rush given that is too noisy for safety evaluation as shown by the alignment average in college).
There are a few approaches we can draw from capital markets research such as Coefficient of Variation (CV), Sharpe Ratio, and Treynor Ratio to help us here. These measurements are key metrics used in evaluating a portfolio’s relative risk and performance.
- The Coefficient of Variation (CV) is simply adjusting the standard deviation across groups relative to the mean (i.e., standard deviation across many groups).
- The Sharpe Ratio is used in capital markets to understand the excess return of volatility. In our case, we will use it to understand a player’s play volatility in order to identify more “stable” safety play. The formula is defined as Return of Portfolio (safety play in different positions) – “Risk Free Rate” / standard deviation of portfolio’s excess return. We derived “Risk Free Rate” as mean positive play performance at a given positional alignment.
- Treynor Ratio is a ratio that calculates reward to volatility or “what is a player’s ability to play beyond their typical alignment?”. The formula is defined as Return of Portfolio – Risk free rate / beta of portfolio. The beta is calculated by taking the portfolio standard deviation and adjusting that relative to the overall market (aka overall coverage or run defense play at a given position such as safety). This can be helpful to give context to the positions a player is in relative to the overall performance.
Finally, we take a weighted average of Coverage and Run Defense metrics to which phase the position is more likely to be utilized in. Theoretically, you want a player who has a low CV, and a higher Sharpe or Treynor Ratio. A low CV would indicate that a player has low deviation in performance across many different alignments. Players with higher Sharpe and Treynor Ratio would indicate that they have a high positive player percentage relative to the benchmark.
Stability of College Play
In order to understand how these metrics can be used in predicting NFL play, we need to first understand how this metric can change as players progress through college by calculating the relative year to year correlation of the metrics. What we learn is that safeties and slot defensive backs performance is noisy on a season-to-season basis. In some cases, this may be expected in game planning (e.g., don’t throw to that player’s direction).
The metric’s stability is likely largely caused by survivorship bias of what is unobserved in the data such as the timing to which people get on the field or those who declare early for the draft. For example, Jessie Bates sat out his freshman year (redshirted), played his sophomore year (receiving all-conference honors) and junior year (missed a few games due to injury) and then declared for the NFL. When we break down the metrics stability by those who were ultimately drafted versus those who were not, we can start to gain some valuable insights regarding how this metric can be used to understand safety development and “draft readiness”.
We typically see the safeties or slot defensive backs making an increase in Sharpe and Treynor ratio as they make a big jump into their junior year (i.e., second to third year). Whereas those who were undrafted, we see the opposite or minimal effect. We see “draft ready” prospects decrease their coefficient of variation (good) as they jump into their third year as well. This is especially interesting given the relative stability in snaps from year to year for both those drafted and undrafted. 
Predicting NFL Talent based on College Report Card
Using a player’s collegiate body of work in assessing how well they will do in the NFL has shown moderate signal. The weighted average collegiate career Treynor Ratio has shown to have a 37% correlation with weighted average NFL Treynor Ratio in the first four seasons. The Sharpe Ratio has a 38% correlation and Coefficient of Variance only has shown to have minimal signal at 13%. The Coefficient of Variation may be limited if a player is a lot more volatile in a certain pre-snap alignment and given less alignment options in the pros whereas the Sharpe and Treynor Ratio works to contextualize that performance relative to a benchmark. Where the Coefficient of Variation has the highest correlation is amongst slot defensive backs versus safeties, which shows to collegiate slot corners ability to plug into many different schemes.
Application to Past Drafts
The primary application of this metrics (other than assessing development) is to identify safeties and slot defensive backs that will be able to play in many different alignments in the NFL and succeed. With regards to primary safeties (i.e., free safety / strong safety), the weighted average Treynor Ratio highlights some of the current top 15 highest paid safeties in Derwin James, Darnell Savage, Jessie Bates, and Grant Delpit. Players such as Kyle Hamilton, Talanoa Hufanga, Bryan Cook, and Jaquan Brisker are young starters who could see their way into this list in the future if their play continues to trend positively. This metric also applies well to the highest paid safeties who were primarily aligned as slot defensive backs in college such as Minkah Fitzpatrick, Chauncey Gardner-Johnson, Justin Reid, and Amani Hooker.
While safety and slot defensive back play is difficult to assess from a scouting perspective given the versatility of pre-snap alignments and assignments, we can use traditional capital markets metrics to further drill down into their “portfolio of play”. From those drafted in 2024, some of the top players in the weighted average Treynor Ratio were Cole Bishop, Malik Mustapha, Jaden Hicks, Mike Sainristil, and Kamren Kitchens.
Those with relative positive play across many different alignments can be seen as having good diversification, and defensive coordinators are looking to cash in on their play for seasons to come.
