Film Study: A look back at Melvin Gordon’s rookie season

In light of Melvin Gordon’s breakout performance in week 1 of the 2016 regular season, it’s worth taking a look back at what he accomplished as a rookie and trying to put it into context the way only exhaustive tape-watching and number-crunching can.

INTRODUCTION

Just before the season starts, I would like to show BFTB the results of my analysis of Melvin Gordon’s runs of 2015. I watched every single of his runs (no receptions, no pass-blocking!), wrote down formation/personnel details, game circumstances, how the defense aligned and a subjective PTS score ranging from 1-5 with 5 being the best. I also summarized the play qualitatively. All of this can be found in the Excel file. Note that I did all of this between May and June, so long before the Gordon pre-season hype train gained steam.

Melvin Gordon does for the most part what he should do. What is really obvious is how the Oline completely fails to block on a consistent basis. Even if there is (not often) a solid push and they transfer the LOS 2-3 yards downfield, there is no hole to run through. Then we have plays that look solid, 4-5 yards easily and Boom!, backside pursuit is not cut off, resulting in a tackle for loss or little gain. What is also striking is the need to cut back/adjust/reroute almost every other play. Very rarely can Gordon take the ball, hit the hole where the play is intended and go from there. Very often, there is no hole at all where it should be and Gordon needs to improvise from the get-go.

As a point of criticism, I sometimes think he wants to “invite” tacklers or blocked defenders to one side, push that side and then cut it back. This sometimes makes the blockers look bad as Gordon fails to cut it back early enough and runs right into the tackler. He also fails, as many know, to keep his feet running sometimes, but as you will read later on, this is not the huge issue it is made out to be.

QUICK MYTH CHECK SUMMARY

Now, for those who want a quick summary, I decided to present my main findings in “MYTHBUSTERS”-style. Let’s get right into it. For details, please refer to

MYTH 1: Melvin Gordon is not a great fit for the many Shotgun/Pistol runs Frank Reich called!

BUSTED! Well, without adjusting for other factors, Gordon actually averaged a good deal more yards when running from the Shotgun (3.32 PTS & 4.22 YPC on 79 attempts) compared to Under Center (3.20 PTS & 2.53 YPC on 59 attempts). Even from the Pistol he averaged more, 3.29 & 3.69 YPC on 49 carries. Over 42% of his carries came from the Shotgun, which is a lot, yes, but Gordon did not struggle running from the gun.

MYTH 2: Gordon’s vision is his biggest weakness and therefore he cannot execute the hundreds and hundreds of Draws/Delays Frank Reich called!

BUSTED! Again, without adjusting for other factors, Gordon averaged 3.38 PTS & 5.66 YPC (36 attempts) on Draws and Delays, significantly more than the 3.25 PTS & 3.04 YPC gained on all other types of plays (151 carries). By the way, if you relate this, only 19% of his runs were draws/delays (you thought that would be more, right?). I distinguish between “Delays” where the action is mainly QB/RB but the Oline takes basically only one false step and then fires forward whereas a “Draw” would entail complete fake of pass protection by the Oline, “inviting” the defenders, and so on. Separating Draws and Delays, Gordon gained 3.42 PTS & 6.71 YPC on 24 Draw carries against 3.29 PTS & 3.58 YPC on 12 Delay carries. So I think the data shows that Gordon does a good job on these improvisational plays where you need to read the blocks by the Oline while you get the ball and then weave your way through the defense without a clear, predetermined path.

MYTH 3: Defenses did not respect the run and Gordon still could not make something happen!

BUSTED! My measure for “Respect” is derived by looking at the number of defenders in the box and subtracting the available blockers (TEs/FBs/OL). So when this is positive, the defense “respects” the run because they put more guys into the box as the offense had blockers (I sometimes counted a defender as one-half if he was close to the box or shot into the box right after the snap). On nearly all of Gordon’s runs the defense either matched (44%) or exceeded (47%) the number of blockers. So even though the Chargers had serious issues up front, defenses respected the run threat. Gordon actually has (in this sub-category) the best Point average and YPC when facing one additional defender in the box. He gained 3.95 YPC here with a point average of 3.48 on 42 carries. It is not true that the Chargers faced defenses daring them to run. However, when defenses eased up (on 8.5% of the runs, at least one-half of a defender less was in the box than the number of blockers), Gordon struggled to a 3.06 PTS & 1.81 YPC line on 16 attempts. Beyond the pure stats, the general impression was seldomly that Gordon faced a light box, obviously a side-effect of 63% of his carries being on first down.

MYTH 4: Gordon is not effective between the tackles, stop running him there!

BUSTED! For reference, over half of Gordon’s runs were intended to go through the A-gaps (97 of 187, 52%). There is little difference in the earned PTS average specified by gaps (and all are slightly above-average around 3.25-3.3). However, his YPC in the A gap was highest at close to 3.71 yards. While not perfect, I think this is a good average up the gut, especially with the issues we had upfront. Apart from the stats, the eye test showed a very powerful runner that rarely is stopped in his tracks. Gordon usually falls forward, runs through lesser tackle attempts and often drags one or two defenders for an extra yard.

MYTH 5: Gordon’s ball security is really bad!

SEMI-BUSTED! I think straight from how he carries the ball, he has no business having six fumbles on 180 carries. He rarely lets the ball far out of his frame and usually keeps it tucked away decently. One of his fumbles was a crazy play by Von Miller when Gordon had both hands on the ball. On at least two others, a guy caused the fumble who had no business being there. Gordon runs straight up the hole and the backside pursuit punches it loose or an unblocked LB comes in and strips it. Obviously, Gordon needs to protect the ball either way, but none of his fumbles I think were crazy bad from a ball security standpoint. If you look at how Gates sometimes has the ball half a foot away from his body, that is something you should criticize. Gordon’s ball security, just from a technique standpoint, I think is not overly bad (obviously not perfect as well). It is not that I zoomed in to confirm this on every carry but bad ball handling definitely did not jump out of the screen. I really hope that he gets some nice breaks this year and proves me right! Many RBs had serious fumbling issues in their first years in the league and were able to shore that up (Adrian Peterson and LaDainian Tomlinson are just two of the all-time greats to mention here).

MYTH 6: Gordon hesitates too much and has really bad vision!

SEMI-BUSTED! I obviously do not have a direct stat for this, but the eye-test shows that Gordon does OK here. He might not make every cut right where it needs to be and sometimes hesitates a second too long but without exaggerating, these things are limited to I would say one or two cases per game. It is really not much more. A majority of runs were simply not working even close to how I assume they work in practice. When there is a “Power” run to the left and there is no hole for the puller to run through, it is hard to criticize Gordon for hesitating after he takes the snap. When, on an inside zone, the DTs absolutely abuse the double teams, LBs will shoot through easily and Gordon has nowhere to go, yes, he will stop his feet there and bounce it. Statistically, on 12% of his carries, he was tackled for a loss. His average PTS there: 2.93. That means he was on average never really responsible for the loss. He played almost like an average NFL RB in these situations and the majority of these losses I attributed on the blocking up front. There were also, at least one or two but also up to three or four plays per game, where a defender was unblocked or close to unblocked and I am sure they tackled Gordon paranoid to some degree. With a better Oline, I am sure that Gordon learns to trust his blockers and this point will be moot. If he can get upright to something that resembles a hole, he is fine. It is definitely not the case that he runs away from wide open holes as many suggest. He might have missed a huge opening a handful of times over the season, not per game. He can improve here, do not get me wrong, but I do not think he was as atrocious in that department as many suggest

OVERALL STATS

OVERALL STATISTICS

His total average PTS came out as 3.26 for the whole season. By my definition, as mentioned before, “3” is basically an NFL average RB. Since I have never done this before, not with great RBs, not with bad RBs, I cannot exactly say how much better than an average RB Gordon is. But a hypothetical RB who, for 100 carries is 50 times average (3 PTS), has five atrocious plays (1 PTS), has ten bad plays (2 PTS), has 25 good plays (4 PTS) and 10 extraordinary plays (5 PTS) would grade out at 3.25. So, I think Gordon is good RB, he played reasonably well last year. Only 10.7 % of his carries were plays where I expect an NFL average RB to get more yards. However, 43.8 % of his carries were better than what an NFL average RB is expected to produce. Obviousyl, this is subjective and most of them were little things like avoiding a two-yard loss and turn it into a loss of half a yard. Sometimes it means falling forward on a tackle for an extra yard. These are things a regular back is not expected in my opinion and Gordon did many of them, a lot of times. 6.4 % of his carries were far and beyond what an average NFL RB would be expected to produce. This is in-line with the information from his “Inconsistency” Score regarding the PTS (standard deviation of PTS), which came out as 0.68. That means that on average, per play, he deviated almost two-thirds of a point from his mean PTS of 3.26. Since he had little deviation below the mean (by looking at the distribution of PTS in the figure below), I confirm that Gordon played like an NFL average RB for the vast majority of his carries (many of those he did get little chance to produce anything else really) but when he deviated from that, it was usually positive. He demonstrated that he is a lot better than an average NFL RB on many of the carries where he actually had a chance.

GAME SPLITS

The game splits here are interesting. You can see how Gordon had a really bad game against the Packers with by far his worst grade. He had 2 fumbles in 7 carries and earned a PTS of 2.64, clearly below average. His best games according to my measure were the Jaguars (3.57), Dolphins (3.47) and Raiders (3.43) games. These are very good PTS averages although sometimes the YPC does not reflect how well he actually played in those games. The last column is also interesting as it shows how inconsistent his PTS were. There is a clear negative correlation here. So when he got hot, he was able to play consistently well. The “AVGRESPECT” measure shows if the defense actually put more defenders in the box than the Chargers had blockers. It is nice to see that defenses with a good front 7 (Chiefs, Lions, Bengals) were not overly concerned with the run threat (although no team “dared” the Chargers, on average, to run by having fewer guys in the box than available blockers). There is a clear positive correlation between PTS and AVGRESPECT, as should be expected. So when defenses opened up, Gordon could punish them. One can also that he was used mainly on first and second down with on average over 8 yards to go. Also, as the Chargers lost many games, he usually got the ball when the Chargers were trailing.

DETAIL GRAPHS

In the following, you can see the detailed splits for Gordons runs as additional info for the “MYTHBUSTERS” section. You can see YPC, PTS and the frequency (COUNT) of the specific split. COUNT is shown as a gray dashed line on the right Y-axis. YPC and PTS are shown on the left Y-axis. Note that some splits are not very insightful, when COUNT was relatively low. These are statistics to support the summary up top and to provide a little more detail.

“QB” [Under Center, Shotgun, Pistol] = Alignment of Quarterback.

“RB” [Left, Middle, Right] = Alignment of Running Back.

“Personnel” [10, 11, 12, 20, 21, 22] = Personnel of Offense as usually indicated. First digit refers to number of RBs on the field, the second to the number of TEs on the field. 5 minus the sum of both is the number of WRs on the field.

“RUN DIRECTION” [Left, Right] = Self-explanatory, direction of play (as intended at the beginning).

“GAP” [A, B, C] = As seen from defense, the gap through which the play was intended to go most likely.

“PLAY TYPE” [Draw, Delay, Stretch, Zone, Power, Lead, Split, Counter] = Play type as identified by myself.

“D-LINE” [3 to 6] = Number of defenders lining up at the LOS, including down linemen, OLBs, ILBs showing blitz.

“BOX” [5 to 9] = Number of defenders in the box. Half values come from a defender shooting into the box right after the snap, or a Safety/LB aligning close to/on the edge of the box.

“RESPECT” [-1 to 1] = Number of defenders in the box minus available blockers, including Olinemen, TEs & FBs.

“QUARTER” [1, 2, 3, 4] = Quarter in which the run occurred.

“DOWN” [1, 2, 3] = Down of the run.

“FIELD POS” [DEEP OWN, OWN, MIDFIELD, OPP, DEEP OPP] = Referring to the position of the ball when snapped in 20-yard splits. E.g. DEEP OWN is between the own end zone and 20-yard line, OWN between the own 21 to the own 40 and so on and so forth.

“TOGO” [SHORT, MEDIUM, LONG] = Refers to the distance needed for first down. Anything below 3 yards is short and anything above 7.99 yards is long.

“SCORE” [-3 to +3] = Scoreboard relative at the point of the run. Down/Up more than 16 points is equal to -3/+3 Scores for example.

EXPLANATION FOR MY GRADING SYSTEM

PTS?

I tried to evaluate every single run play by Gordon on the following scale.

1. Complete mistake, fumble, big loss, etc.

2. Bad play, failure to take what was there

3. Average play, what you expect a normal NFL RB to do

4. Good play, a little more than what was „out“ there

5. Great play, above and beyond what was asked for in the situation

Yeah, it is subjective as hell, but it is what it is. I used half grades to have a little more precision. A general disclaimer: without knowing the Chargers’ playbook in detail, this grading task is theoretically impossible but I did my best. My blame/praise for Gordon or the Oline on any specific play could be doubtful or completely wrong since it is not always clear what kind of run was called. Sometimes I think I could identify Olinemen doing the wrong thing or Gordon going the wrong direction, but overall, I cannot say that this is the undeniable truth.

As a reminder of the difficulty of this task, I think back to my playing days (not that long ago actually) and we had this play “22 bounce”. Basically, it looked like a simple inside zone, hitting it right at the B-gap. But the bounce called for the RB to take it outside after pressing the B-gap initially. This play was introduced because of our previous tendency to run between the tackles and we expected to catch LBs cheating inside. Now for an outsider, it would look like the Oline was completely failing to open a hole at around the B-gap (when the RT was, in reality, inviting the LB inside), the RB was great at reading the play, when, in reality, everybody did exactly their job and nothing more. I do not think that the majority of runs I analyzed are very complicated in themselves but there are surely many wrinkles (some game-plan specific) that could make me blame/praise the wrong guy(s).

My formation descriptions might be a little random but I know what they are and I can clear them up if need be. When I call a personnel “20” I do not care whether the guy next to Gordon in the backfield is listed as a WR/TE/FB or HB. So in that regard its more about where guys line up than about roster spots. Same for RBs or TEs split out wide. When Green is outside, like a “normal” WR, he does not count as a TE in the personnel (Woodhead would be another example, he is simply playing WR when he is split out at the snap). Same with D. Johnson. Is he in the backfield, he is an RB, on the line I count him as a TE.

And all of the data I compiled can be viewed here.