This is our old blog. It hasn't been active since 2011. Please see the link above for our current blog or click the logo above to see all of the great data and content on this site.

Do star safeties have shorter careers than players at other defensive positions?

Posted by Jason Lisk on April 5, 2010

A couple of weeks ago, I looked at Peter King's comment on Eric Berry to examine whether safeties were risks at the top of the draft. In that post, I determined that safeties were not risky in the sense that they didn't pan out. In fact, compared to other positions, a higher percentage of safeties panned out in that they made at least one pro bowl and were at least good starters (career AV > 49).

However, King did raise a point about safeties being less likely to stay healthy, citing recent injuries to star safeties Troy Polamalu, Bob Sanders, and Ed Reed. Today, I try to take a stab at looking at that issue. The best way we have of doing that is to look at games played data, though games played is also susceptible to talent and ability as well as health. In an attempt to deal with that fact (though I am sure I fail in limited specific cases), I decided to look at players who already proved to be good players at a young age (age 25 and under) and see how many games they played from ages 26 to 29, how frequently they retired before age 30, and how old they were during the last season they were able to play 10 or more games for an NFL team. The hope here is that if a player has played well through age 25, he should continue to play in games through age 29, unless he missed those games due to injury, or was benched or forced to retire due to injury-related decline.

So, I pulled the top fifty players (and ties) as measured by Approximate Value, for all defensive players who turned 25 between 1978 and 2005, at the following positions: defensive tackle/nose tackle, defensive end, inside/middle linebacker, outside linebacker, cornerback, and strong/free safety. A couple of footnotes to that. First, a player had to qualify within one of those positions exclusively. If a player switched positions between, say, outside linebacker and middle linebacker, before he turned 25, then he would have to qualify based only on the points at one of those positions. If he had 25 points of AV at outside linebacker and 5 at inside linebacker, he may have made the list. If it was an even split of 15 at OLB and 15 at ILB, he did not. So, players could have switched positions and still made this list, but it had to be on the strength of only one position. Second, switches within a position did not matter and were counted together--safeties could switch from free safety to strong safety, defensive ends could switch sides, defensive tackles could move to nose tackle. Finally, we always have to balance sample size and sample relevance. I would loved to have a larger sample to examine, but I felt by going much past the top 50 over a span of less than 30 years would bring in guys who were playing because of need/draft position only, and would miss games because of non-injury reasons more frequently.

After coming up with the list of players for each position, I recorded the percentage of games played from ages 26 to 29, and, for retired players, the age at which they last played in 10+ games. Let's get to the results. The columns represent the number of players included in that position grouping, the average percentage of available games played by the young "stars" at that position from ages 26 to 29, the "final age" is the average age that the retired players at that position last played 10+ games, the 33% column represents the percentage of players that missed at least one-third of the available games from ages 26 to 29, and the "done by 30" column represents the percentage of players that played their final season of 10+ games at age 29 or younger.

position no. game pct final age missed 33% done before 30
OLB 51 0.877 31.9 0.098 0.196
CB 52 0.827 31.9 0.173 0.250
DT 54 0.813 31.1 0.148 0.267
DE 51 0.808 31.3 0.235 0.333
ILB 51 0.766 29.9 0.294 0.489
S 50 0.763 29.8 0.260 0.533

Well, look at that. Despite my earlier criticism of King, safety does check in as more injury prone than other defensive positions, in a virtual dead heat with inside linebackers (and my data set does not include any of the three safeties he cited). Just over half of the star young safeties were effectively done or retired before they reached age 30. Compare that to the other secondary position, where 25% of the cornerbacks were done before age 30. The other thing that stands out to me is the size of the difference between playing outside linebacker versus inside linebacker on your life expectancy. I am sure there might be some agents and players interested in seeing how much a switch from the outside to inside will cost them in terms of future career length, as the young star outside linebackers played 2 seasons more on average, and played in almost a half-season's more games from age 26 to age 29. In fact, the outside linebacker who played in the fewest games after turning 26 was Thomas "Hollywood" Henderson, and his demise wasn't because of on-field injuries.

Even though King was proven correct in his assertion that safeties may break down due to being smaller bodies who throw themselves around, I don't think it then follows that drafting safeties early is a bad idea. In fact, using this information, the opposite may be true. As I pointed out in the previous post, a team should be primarily concerned with what a draft pick provides while playing under the initial rookie contract. If we have a position that has a relatively low chance of providing complete busts, and where veteran players in their late twenties tend to fall apart more frequently than other positions, that strikes me as the exact situation when you would want to use your draft picks on a position. We see that on the other side of the ball with running backs, as teams tend to use a high frequency of picks at running back, but are hesitant to give big money long term deals to veterans past a certain age.