One of my favorite, non-Chiefs-related football sites is Football Perspective. Chase Stuart, who runs the place, recently had a couple awesome posts that got me thinking more than I already was about Knile Davis as the potential future starting Kansas City running back if/when/let's-not-talk-about-it Jamaal Charles begins to slow down. Unfortunately, the awesome posts confirmed what "advanced" stats have been saying about Davis for two straight seasons... but more on that shortly.
This post will go from stat stuff to some charting on Knile's carries vs. league average to some All-22 film snapshots. Together, hopefully they give us a better idea of what we have in Knile Davis.
This is possibly a controversial subject to many traditional analysts and perhaps to some of us here at AP, too. But it is a simple question: how many carries would you have to take away from a running back to bring his yards per carry (YPC) down to league average. The answer? Nine.
The best back in the league last year by YPC was Justin Forsett at 5.39 yards. If you remove a mere 9 of his 235 attempts, his YPC drops to league average.
The most productive back in the league last year was Demarco Murray. But take away just 7 of his 393 attempts, and his YPC also drops to league average.
Out of this very simple information emerges a stunning fact: on 386 of Demarco Murray's carries last season, he was indistinguishable from an average back according to YPC.
What about the corollary? How many attempts would you have to remove from a below average back to bring his YPC up to league average? This is where our own Knile Davis first appears. You would need to take away 13 of his worst carries to give him an average YPC.
All of this can be pretty mind-blowing at first, depending on how much you liked YPC before-hand. After all, this list of backs with at least 4.5 YPC over 1,000+ attempts is a pretty consistent list of the greatest to ever grace the game. Doesn't that mean YPC tells us something valuable about who is and is not a Hall of Famer? Well, maybe. The problem is that YPC is heavily affected by outliers. If we assume taking away 10 of an RBs best runs each season gets him to league average, then you would only have to remove 50 (likely less) of Jamaal Charles' runs from his five full seasons to make him look "average." That's right, the greatest yards per carry stat of all-time over a whopping 1,200 attempts is made to look average by taking away merely four percent of the total carries.
Thus, a few good runs can make YPC look elite. A few bad runs can drag it down. But a huge part of what makes a running back great or not is found in those 96 percent other attempts that, according to YPC, are average. Over time, outliers matter less, but in the short term we need a better way of measuring who is making the most of 100 percent of his carries, not four percent of them.
A RB Heat Map tells you what percentage of a back's carries went at least 0 yards, at least 1 yard, at least 2, at least 3, at least 4, and so on, for however many intervals you want to include.
What makes it really cool (aside from the fact that it's a freaking heat map) is how it gives us a nice visual presentation of the various running backs and their skill-sets.
Returning to Forsett, for example, a lower percentage of his carries (red) make it to the line of scrimmage than the average back. Same goes for carries that make it at least 1 yard, and 2, and 3. But, beyond that, Forsett starts to shine. He is blue (above average) at every other interval, including leading the league in the percentage of his rushes that went at least 8, 9, 10, 15, and 20 yards. This is your big play back who otherwise gets caught for little-to-no gain.
The opposite is true of Anthony Dixon, who had 105 attempts last year with Buffalo. This is a back whose carries are more likely than most to get you at least a few yards, but he has no big play production.
If you are looking at the map, you will, of course, have first searched for two things -- and they aren't Forsett and Dixon. You went for Jamaal and Knile, in that order.
Charles, via a quick eye-balling of the total percentage points in each column, is the 2nd or 3rd best back. Davis, by that same metric, is the second or third worst. Charles is above average in every single column. Davis is almost entirely the opposite.
As I said above, this map confirms what other stats have been saying about Davis for two straight seasons: that he is not producing up to league average. Out of the 57 backs who played at least 25 percent of league snaps last season, Pro Football Focus (PFF) had No. 34 ranked last in both total grade and run game grade. For backs with at least 100 attempts, Football Outsiders (FO) had him in the bottom 4 of every one of their major statistics, including Success Rate.
Success Rate (SR) can tell us some of the same things we were looking at above, but with a slightly different goal. As FO explains:
This number represents the player's consistency, measured by successful running plays (the definition of success being different based on down and distance) divided by total running plays. A player with higher DVOA and a low success rate mixes long runs with downs getting stuffed at the line of scrimmage. A player with lower DVOA and a high success rate generally gets the yards needed, but doesn't often get more.
This possible split mentioned between DVOA and success rate is the same observed in Forsett and Dixon above. But where the Heat Map tells us what yards an RB produces on every carry, regardless of the situation, DVOA and Success Rate are interested in measuring whether or not a run is successful from a team perspective. Let's focus further on the latter.
SR is measured in a more complicated fashion than what I am about to lay out, but the basic premise of it goes back to the birth of advanced stats in a book called The Hidden Game of Football, which FO has improved upon. The gist of it is that, on first down, a play needs to gain at least 45 percent of the yardage needed for the play to be successful. So a 1st and 10 run needs at least 4/5 yards to be a "success." second down runs must manage 60 percent of the needed yards. Third and fourth down runs must successfully convert for a 1st down.
I thought it would be useful to chart Davis' runs using this basic criteria and compare it to league average. As I said, FO has Davis ranked bottom four in SR. For 2014, they say that only 37 percent of his carries were a success. Let's see how Davis looks in a simpler Success Rate based solely on the rules just described.
The above chart tells us what percentage of runs were a success on various down and distances over the course of the 2013 and 2014 seasons. Some of the columns for Davis have very few attempts, so don't read too much into each one; rather take in the whole picture to see that Davis is, regardless of the situation, completing runs less successfully than league average. FO's success rate for Davis in 2014 was 37 percent. Our simple rate over the last two seasons has Davis at 35 percent.
On 1st down attempts in particular, where more than half of Davis' attempts have come, the third-year-back is gaining four yards only 33 percent of the time -- which is to say he is gaining three yards or less 67 percent of the time. The rest of the NFL's backs are, on average, successful on this down and distance 45 percent of the time.
Altogether, Davis has a 35 percent "simple" Success Rate, while the NFL average is 50 percent. Again, this means, just as the Heat Map and the FO stats tell us, Davis is producing far less than league average across all his carries.
So how is this happening? We do know the run-blocking deserves its fair share of criticism. As with all football stats, it is very difficult to isolate individual performance from the context of the team and the many moving pieces. But the fact that Jamaal Charles and Knile Davis play behind the same line and yet have such dramatic differences in so many statistics does at least tell us one thing: the explanation for Davis' numbers cannot be entirely reduced to the o-line.
Knile Davis' biggest game last season came against Miami in Week 3. Charles was out and so Davis was the starting back. And he was worked like one. Davis had 32 carries for 132 yards and a touchdown. He forced 10 missed tackles and had a long run of 21, and another of 17. Not bad.
Yet PFF gave Davis an awful -3.7 run grade. Part of that is the result of two fumbles. But I think the film below might reveal a few other things that PFF saw that day and that may need to be addressed if Davis is to emerge as a viable back-up to Jamaal Charles.
Here, former Chiefs tight end Anthony Fasano (red) will move across at the snap to block any pursuit. Davis (yellow) is to follow Eric Fisher (No. 72) and Mike McGlynn (No. 75), who will take care of the two guys in front of them...
... except McGlynn instead attempts the splits as the guy he's blocking performs instant transmission and suddenly ends up on the other side of him.
This run is proof that, yes, running behind this line, and especially to the left side, was probably no fun. In fact, PFF grades runs by direction, too; and Kansas City runs to the inside left against the Miami front graded out the worst. This is just one example of why.
But, like I said, not everything boils down to blocking.
Here, McGlynn will successfully pull to his left.
Davis has a clear lane...
... but doesn't take it. He is shown above by the arrow, on his toes but leaning back, hesitating. He will get wrapped up instead by the back-side pursuit.
This hesitation and unwillingness to take the space given to him was a common mistake by Davis against the Dolphins. Another example follows below:
Former center Rodney Hudson is going to pull with McGlynn. Davis will follow them outside.
But, again, he hesitates. The blockers are running out in front, leading the way, but Davis has already slowed his momentum and considers cutting back.
Hudson drives a low block and McGlynn continues to pull outside.
The two men finish their blocks well, and Davis now decides he should continue his run towards the sidelines. He manages to pick up a few yards.
This next play was better from Davis, displaying his speed and potential. It produced his lone touchdown on the day.
Former Chiefs receiver AJ Jenkins (#15) has been motioned to the back-field. He will jet out at the snap to fake the pitch while Fisher (red) comes across to seal for Davis (yellow) running counter. The rest of the line sells the Jenkins toss. A well-designed play.
And well executed:
Fisher hits his block perfectly and Davis has a wide open lane. What I noticed watching Davis is that small holes cause him to hesitate, even if the blocking is doing its job. But more open lanes like this provide him enough space to be dangerous.
Davis, in space and running fast, can be a meaningful runner. He makes the right cut here back inside, runs through an arm tackle, and then another, crossing the goal line for a touchdown.
The blocking got him going, but Davis deserves credit for finishing.
Here, however, Davis hesitates. By my amateur opinion, either lane will do, and Davis has the speed to make either work for a short gain. But he never chooses one. The frame above already shows him slowing up. This play ended up as a 1 yard gain when Davis was driven out of bounds.
One of the things I noticed often is that Davis would receive the hand-off and immediately stutter or pull up. This is an example where a quick decision to hit the hole provided to him, and hit it hard, would have got him to the second level at full steam -- but instead he hesitates. He does manage to break an arm tackle from the free linebacker and pick up 6 yards, but it was more difficult than it should have been.
This was an 11 yard gain -- Davis' third-longest run of the game. But he didn't have to do much. The blocking out front is on point and there is tons of space. Davis looks a lot more comfortable on runs like this.
This was taken right after a read option. The free defender, forced to make sure quarterback Alex Smith did not keep the ball, is out of the play, and Davis can bulldoze his way through the lane. Both outside blockers have their man taken care of, and a double team on the inside defender means Davis is in a great spot to gain some yards.
As you can already tell from the still, however, Davis is leaning back and thinking of coming inside. This negates the advantage of the read option because the unblocked defender who was taken out of the play by the numbers game can now make ground on the hesitating Davis. In fact, he ends up making the tackle.
A nearly identical thing happens above. The linebacker (No. 50) is completely out of the play at this point. He has to watch Smith release the ball into Davis' gut before he can pursue. If he crashes inside too soon, Smith can pull the ball, keep it, and run it himself outside.
The blocking where Davis is heading has by no means produced any obvious lanes, but at least every guy is accounted for. Following Hudson (the center) is the best bet, and then simply pick left or right. Should be an easy couple of yards. Instead, Davis goes nowhere with the ball and the frozen defender from the read option gets back into the play, making an assist for the tackle after a small gain.
PFF has something they call "Breakaway Percentage", which is quite simply the percentage of a runner's total yards that came on 15+ yard plays. Unsurprisingly, Davis ranked second in 2014. In open space, Davis is dangerous. Of his 463 yards last season, 176 came on runs that went 15+ yards.
Davis' Heat Map confirms that two percent more of his runs went at least 15 yards than the average back -- but that's not a dominating percentage. And for 10 and 20+ yard runs, Davis was below league average. Altogether, this makes Davis a boom-or-bust back without much boom. Ideally, of course, the boom plays would be so plentiful that they would make up for a low success rate in other situations (see: Forsett). But Knile's not giving Kansas City the big play enough to make that trade-off worth it. At least, not yet.
He hesitates too often and sometimes fails to follow his blocks. There is no doubt the 24-year-old has speed and maybe even a high ceiling, but he has not been able to take advantage of either thus far in his career. The NFL is a fast-paced game, featuring linebackers who can close quickly, go sideline to sideline, and are never out of the play even if they have to chase it from behind. Good running backs create their own space in which to utilize their speed. Davis, however, rarely gets to use his speed due to hesitation at the line of scrimmage.
We shall see what 2015 brings, but the Chiefs could end up drafting a running back for competition. I would not be surprised. So here's the question for AP: is Knile Davis the undisputed heir to Jamaal Charles' throne?