One of the main problems with the Speed Score, in its current iteration, is that it is notorious for missing on many players. For that reason, the Speed Score has to be just one tool - or metric - among many before we have a clear picture of which running backs we can expect to succeed / fail.
I already introduced one new metric, Punch Score, and gave the Speed Score a solid mathematical foundation that actually makes some intuitive sense.
It's time to introduce the third and final metric that we can use together to make educated guesses at which running backs will be busts/steals in the NFL draft: Foot Speed.
Keep in mind, making these determinations on one day's worth of a conglomeration of ability, strength, and speed tests that these athletes make is obviously going to miss on some players; even a perfect system would miss on the character guys and the guys who hit the injury bug in the pros - so no measure would be perfect.
The curious case of Darren Sproles
One of those things that I always notice when I watch football are the observational comments people sometimes make that, on the surface, are really hard to prove one way or another. The most obvious ones are the "gritty" comments, there are many of these such statements.
Among these comments is "he looks like he's running so fast", often attributed to Darren Sproles' tenure as a Charger. Depending on how you take this quote, it could mean that not only is he fast, but he looks fast while doing it...or that he looks fast but really isn't.
This piqued my interest, though. Why is it that Darren Sproles looked fast, and so much so that it caused people to actually comment on it on plays where he wasn't doing anything in which his speed resulted in a favorable result? After all, Sproles had a pretty yawn-inspiring 40 yard dash at the NFL Combine.
Clearly, Darren Sproles is shorter than your average running back. Unless you want to believe that all short running backs have really small upper bodies, the length of these players' legs are going to be shorter than the taller backs'. In order to have a similar straight-line speed, despite probably having a shorter stride associated with shorter legs, the player will probably need to have quicker foot speed.
Hence, "he looks like he's running so fast". But how can we prove it based on the NFL Combine numbers?
Here's where the approximation comes in. The closest thing we have to measuring a player's stride at the NFL Combine is their broad (horizontal) leap. If you have two players who both run 4.40 40-yard dashes, the difference in their leap distance will come down to two things: athleticism/actual leaping ability, and stride. So, it stands to reason that a player with a great 40 time, but a low combination of athleticism and leaping ability, will have great foot speed.
So, much like I did with the Z-Speed Score, we're going to compare the rarity of the player's speed against the rarity of the player's broad jump score, by z-scoring both variables. We'll then subtract the player's broad jump z-score from their z-scored 40 time in order to produce our 'Foot Speed Score'.
Players with scores above zero have a more impressive 40 yard time, compared to the general population, than their broad jump score. These players, in theory, are more likely to have fast foot speed.
Historical Foot Speed Scores
The fun part. Let's take a look at some historical foot speeds to give us some context.
Again, since this is a recognition tool - a metric that allows us to treat the best/worst as potential bargains/busts - let's take a look at the top five foot speeds at the NFL Combine since 1999, as well as the four lowest foot speeds coming from a player with a +100 Speed Score.
|Player||Year||40 Time||Broad Jump||Foot Speed||Speed Score|
Take a look at those top five scores. The top three are perhaps the three players who most come to mind when I think of players who "look like they're running so fast". Edwin Baker is a similar player, but he didn't really get a chance to play before the end of this last season with the Browns. Andre Hall has been out of the NFL for awhile, but the very first line in the 'strengths' field for Andre Hall in the '07 Football Prospectus reads "Hall is a very quick-footed back".
Meanwhile, those other guys all had impressive broad jumps that may have indicated that their 40-times were largely a result of a longer stride. During a cursory look over the players with a meaningful negative score, the lowest thirty-plus players didn't have anything that could be considered NFL success. Joique Bell has a chance to buck that trend, but with infamous busts Maurice Clarett, Ryan Williams, and William Green on the same side of the ledger, it feels fairly safe to make a rule like 'stay away from guys with poor Foot Speed'.
The player with the most NFL success who scored the lowest was Ricky Williams - but with a score just around -1, it's difficult to say that the spread between his 40 time and broad jump really means anything anyway. Again, this is a recognition tool. One that successfully says 'look at these guys who may be useful in the screen game or on returns' on one end of the spectrum, and 'maybe don't believe these speed scores' on the other.
Why Foot Speed matters on the field
Foot Speed matters for a variety of reasons. First and foremost, it implies neurological speed. Neurological speed has positive repercussions all over the football field.
Secondly, a runner who takes more steps to achieve the same speed has more opportunities to change directions. I'd imagine our Foot Speed metric would correlate with shiftiness in some fashion.
Indeed, we can anecdotally state that the top Foot Speed score players are those guys who are most useful in the screen game, on third downs, and returning kicks on special teams.
Note: We could attempt to adjust the broad jump based upon the vertical jump - in effect isolating the x-direction distance - but I think it's important to introduce the fewest assumptions possible. After all, this is being used strictly for to identify those who may potentially be steals and those who may potentially be busts because of this statistic; we only really need to be concerned with the really good and bad scores, and not so much the granularity of it all. In a future iteration, I may attempt to adjust for more, but I want to be certain I'm not making it noisier.
Finally, I can bring Foot Speed, Punch Score, and Z-Speed Score together to present the group of running backs who will be most over/underdrafted in the 2014 draft.