One of the Chiefs' most dire needs this off-season is a dynamic safety, but GM Scott Pioli is reluctant to take a safety with the 5th pick in the draft. Falcons GM Thomas Dimitroff is apparently on the same wavelength. There seems to be a growing conventional wisdom that safeties are high-risk picks at the top of the draft. As Peter King pointed out recently, the three best safeties of the decade--Ed Reed, Bob Sanders, and Troy Polamalu--have missed 78 games due to injuries in their combined 21 NFL seasons.
The thinking is that safety (ironically) is a fundamentally dangerous position. The nature of the position, launching head-first at high rates of closure toward oncoming ball carriers, may carry a systematically higher risk of injury than most other positions. Reed, Polamalu, and Sanders suggest this may be the case, but a sample size of three is small to say the least. Are Pioli and Dimitroff rightfully concerned?
To find out, I looked at the top draft picks at safety over the past few decades. I looked at two statistics that indicate durability: total games played and seasons as their team's primary starter. I also looked at two other statistics that indicate total value as a player. Pro Bowl selection is a fair standard for top draft picks, and it at least indicates that a player was not a bust. I also looked at the "Career Approximate Value" statistic developed by Doug Drinen at PFR.
All data are from PFR's excellent draft query tool. I looked at 1st round picks from 1980 through 2003. Unfortunately, PFR doesn't distinguish defensive backs as either a safety or cornerback in its draft data. I classified each by hand according to their initial position classification during their NFL career. Some cornerbacks eventually transition to safety later in their careers. Ronnie Lott, Rod Woodson, and Antrel Rolle are prime examples. In those cases, I classified them as cornerbacks because that was their intended position at the time of their draft.
Although at first the cornerback/safety distinction was a complication in my analysis, both positions serve as good comparisons for each other. They each call for similar but not identical skills and talents, but they perform different roles on the field. The first comparison looks at 1st-round CBs vs. Ss in terms of durability and overall value.
|Yrs as Starter||5.6||5.9|
|Avg Pro Bowls||1.2||1.5|
It appears that 1st round Ss are at least as durable and valuable, if not more so, as their CB counterparts. With only 36 Ss taken in the first round during the period studies, we can't put much stock in the slight advantage they show in terms of value and longevity. Still, at the very least, we could say that these numbers do not show any statistical cause for concern when drafting a top S.
The second comparison looks at only the top 10 overall picks in each draft. Here we see a slightly different story.
|Yrs as Starter||6.7||6.2|
|Avg Pro Bowls||2.4||1.7|
Safeties taken in the top 10 overall picks play fewer games, have shorter careers and are less likely to be selected to pro-bowls. It's important to keep in mind we're only talking about 11 safeties in the sample, so these numbers should be taken with a heavy grain of salt. It's easy to see why some GMs and coaches would get the impression that taking a safety at the top of the draft can be a gamble.
The fact that safeties taken throughout the entire first round show no extra likelihood of injury or demise indicates that the differences observed between cornerbacks and safeties in the top 10 picks is due to sample error. Unless there is something systematically particular about the #11 through #32 picks that prevents safeties from being injured as often as their top 10 counterparts, I wouldn't conclude there is any particular risk drafting a top safety.
So far we've only looked at a comparison with cornerbacks, so let's expand the view and examine how safeties compare to some other positions. The next table lists the years as his team's primary starter for selected positions, both for all 1st round picks and for just top 10 overall picks. In the following tables, I used a slightly different data set. I limited the scope to the 1980 through 2000 drafts (instead of through 2003 in the tables above). The average career lengths will be longer because more recent picks won't have a limit of the present day.
|Position||1st Round||Top 10|
It appears that the longevity of safeties is in line with several other positions. In fact, they seem to have slightly longer careers than the sample of other positions listed.
Lastly, I looked at the average Approximate Value of the top picks in various positions.
Here it appears safeties are less valuable than some of the other positions. This could be because AV isn't calibrated fairly across all positions (which I doubt), or it could be because players drafted highly as safeties just aren't that great. Actually, I suspect it's because a few of the very best safeties in the last generation or two have been converted cornerbacks. Hall of Famers such as Lott and Woodson, who were perennial Pro Bowlers and All Pros after moving to safety, gobbled up a lot of the AV available. That may explain both why AV for CBs is somewhat higher than average and why AV for Ss is lower than average.
In the end, I don't think there is much evidence that top safeties are any bigger injury risks than top players at other positions. The small sample size of recently injured safeties is misleading. In fact, of the 78 games missed by Reed, Polamalu, and Sanders cited above, 49 (62%) of the 78 were missed by Sanders alone. And Bob Sanders was actually a second round pick.
Perhaps there is an arbitrage opportunity. While all the other GMs and coaches are staying away from safeties, a smart GM can pick up a potentially great player later in the draft than he otherwise could.
Edit: I was pointed to PFR's own analysis of this exact same question by Jason Lisk. Sorry for the overlap, Jason. Hopefully, I've added a good deal to the discussion.