The SIP Database: the fuel that gets us where we want to go

To accomplish our mission of predicting injuries, we need to start with robust injury data. Any attempt to predict injuries will be garbage if your input information is garbage.

The Sports Injury Predictor staff has spent the past year building up an already rich database of injuries and refining our processes for tracking those ailments in order to better project their future impacts.

The result: a database of hundreds of NFL players with injury histories dating back to college.

It all started with hours spent manually poring over years of injury reports. We looked beyond the pathetic excuse for a “report” that the league officially puts out every week. You can’t learn much from a player name and a body part in parentheses, and we all know about league-wide issues with inconsistent injury reporting.

Sure, it’s heroic that Tom Brady has battled through that tragic shoulder injury he suffered while sliding out of the birth canal. But including it on every week’s injury report doesn’t tell us anything. We’d be much more interested, on the other hand, to find out how many of the concussions Dr. Gisele reported actually occurred.

Some situations proved easier to discern than others, but we tracked down as many publicly available details as possible, including news reports from team, local and national sources. (We don’t have access to private medical records which are protected by privacy laws or inside info from NFL team doctors.) And we’ll continue doing so for all reported injuries to fantasy-relevant NFL players – on both sides of the ball – going forward.

We wanted to add “fantasy relevant” players to the data because they are the population who are most important to predict and they get enough media attention to have reliable injury reporting. As of August 2017, we have vetted a combined 611 offensive and defensive players. The offensive players played between 2012 and 2016 and in some season in their careers averaged at least 16 touches per game for QBs, 4.5 for RBs, 1.6 for TEs and 2 for WRs. The 611 players are not exhaustive of these criteria, but are representative. The IDP players were selected based on a top-40 fantasy performance somewhere between 2014-2016.

College injuries prove trickier. The NCAA and member conferences don’t compel schools to report even to the inconsistent level that NFL teams do. But you’ll still find a number of them included in the database.

For every tracked injury, the SIP database includes details such as:

To date we have vetted 611 players; tracked down and verified all the public details. That group has produced 3,165 total injuries.

Here’s a look at how those records spread out across positions:

We also drew on the expertise of Jeremy Funk, a clinical epidemiologist and contributor for consultation helped us to define:

The 6 injury classes break down like this:

The 13 body parts, in order of injury frequency according to our data:

The following sites have proven to be the most commonly affected by injuries in the data we’ve collected so far:

And here are the top conditions that football observers (including fantasy players) can expect to occur:

And you’ll find still others listed as “unspecified” to cover injuries that don’t fit neatly into our spectrum of categories.

As we move forward, an ever-richer database of injury data will only give us more to work with. We’ll continue to add as more injuries occur around the league throughout the season. Our list of vetted players will continue to grow as will the details we track. And along the way, we’d welcome your submission of any previous injuries you find missing from our database.