This prospecting season, I've done more work on Isaiah Spiller than I have on any other single player, and I have probably dedicated as much thought to Spiller as I have to the rest of the running back class combined. It's both exciting and unsettling that my process has led to me a position that is near opposite of the conventional wisdom that has Spiller as a top-3 (and often top-1) running back in this class, and I've attempted to have conviction in my own process while also exercising openness and intellectual humility when considering arguments to the contrary of my evaluation. To this point, I've written two entire articles over at breakoutfinder.com just on Spiller, and I need not monopolize those airwaves with my continued beating of this dead horse. However, I've kept my eyes and ears open for pro-Spiller arguments because I want to explore the possibility that I'm wrong, and I want to publish my definitive take on the issue with all of those arguments in mind. This article will go through all of the good and the bad that I see in the Spiller profile, as well as address the major pro-Spiller arguments that I've come across.
The Good
Undeniably, there is a lot to like in Spiller's prospect profile. Number one, he has good size. He was listed at 6'1 and 215 pounds on Texas A&M's roster in 2021, and based on historical data, I anticipate he'll be about 6', 220 at the Combine, giving him 67th-percentile size at the position and making him built like a lot of the premier bellcow backs in recent NFL history.
He was also productive at a high-end SEC program. He broke out with a 22.3% Dominator Rating as a true freshman, a 76th-percentile season, and then -- while never responsible for a truly dominant portion of A&M's offense -- maintained quality numbers throughout his three years on a team that was consistently competitive in the best conference in the country.
Spiller's receiving ability also means that he's capable of playing on all three downs. His career-best Target Share of 9.7% is a 59th-percentile mark, and Spiller's seasonal Target Shares never dipped below the 8.4% mark. Those numbers are even more impressive in the context of the size of Spiller's overall role in the Aggie offense; his career Satellite Score -- a number that contextualizes Target Share in the face of Dominator Rating to describe how much a part of a team's passing game a running back was relative to his contributions to the offense as a whole -- is a 39.6, a 72nd-percentile number in the same range as guys like Darren Sproles, Chris Johnson, Najee Harris, and Saquon Barkley. Spiller was not particularly efficient as a receiver over his career -- his career marks in Yards per Target, Yards per Reception, YAC per reception, and Catch Rate are all between the 22nd and 39th percentiles -- but he filled a fairly dynamic role in the passing game, lining up out wide or in the slot on 7.1% of passing snaps (45th percentile) while seeing an average depth of target of 0.7 yards (64th percentile). He was not just a dump-off running back, and his deployment and usage (he averaged over 2 receptions per game over his three years, 85th-percentile receiving volume) speak to his potential to be a receiving threat at the next level.
Spiller is also very young. He's got an August birthday, meaning his true freshman breakout season came when he was barely 18-years old, and only Darren McFadden and Jamaal Williams have been younger as breakout true freshmen at running back since at least 2007. Spiller is also an early declare, and if he's selected in the first two rounds come April (which it seems a lock he will be), he'll be the youngest running back taken that high since McFadden in 2008. We want our running backs to enter the NFL with years of shelf-life left and after having proven themselves as early as possible as amateurs, and Spiller checks those boxes like few other backs before him.
It also seems to be true that Spiller looks great on tape. A lot of smart analysts that I respect, including Ray Garvin, Jordan Reid, Fusue Vue, Nate Liss, and dudes on Twitter whose full names I don't know like @angelo_fantasy and @devydeets, are fans of what they see from him on film. That's certainly worth something, and it's the reason why I take the pro-Spiller arguments seriously.
The Bad
My reservations about Spiller are 100% centered around his inability to run the ball efficiently relative to the per carry output of the other running backs at Texas A&M, a problem that was consistent for Spiller over his entire three years.
When prospecting players -- regardless of whether your method is film-watching, analyzing data, whatever -- it is important for your evaluations to be grounded in some sort of baseline that informs whether a given player is good, bad, average, excellent, etc., relative to other players, or, more specifically, that informs whether a given trait is so relative to that trait in other players. Essentially what I mean by that is if your conclusions about a player's ability in a particular area are not informed by how well other players perform in that area, then your conclusions are meaningless. If some film grinder tells you that Running Back X has excellent vision -- even if he/she can point to specific examples in the film that demonstrate that vision -- the claim that the vision is "excellent" is without merit if the film grinder is not operating from some sort of base-level knowledge (probably gained through experience) of what average (as well as poor, great, etc.) vision looks like. Similarly, if I told you that Running Back X was a great producer in college because of his 38% Dominator Rating and his Breakout Age of 19.4, I've said essentially nothing if I haven't done the work or accessed the data that puts into context whether such a Dominator Rating and such a Breakout Age actually do indicate greatness from a production standpoint.
I say all of this to get around to my main point: every single year, the film grinders and the spreadsheet jockeys and the player evaluators at large are going to fall in love with a bunch of random dudes who they will proclaim as generational or underrated or sleepers or whatever superlative they choose, and in most of those cases, the players being hyped up as something interesting are instead completely unspecial. Football players are not exempt from the truths of normal distribution, so much of my process is about forcing the context of the population of players at large into every aspect of every evaluation. Put simply, I want to know how well Running Back X produced, and simply knowing that he ran for 1500 yards and concluding based on my general sensibilities that that seems good is not enough: I must know what 1500 yards means in the context of his team's offense, and I must know what his contributions in the context of his team's offense mean in the context of everybody else's contributions to their team's offenses. The same is true of every other aspect of a player's profile, which brings us back to Isaiah Spiller's rushing efficiency numbers.
On his 541 carries in three years at A&M, Spiller averaged 5.53 yards per carry. My base of knowledge from having been mostly aware of football for 20 years on this planet tells me that 5.53 yards per carry is pretty good, but I don't actually have any idea whether or not it's impressive for an NFL prospect. Luckily, playerprofiler.com does know, and they have Spiller's yards per carry as a 60th-percentile number. That's great, but what my base of knowledge from having been mostly aware of football for 20 years on this planet also tells me is that yards per carry on it's own doesn't tell me much, at least not much about a particular player outside the context of the performance of his offensive line or that of his offense in general. Maybe Spiller is playing in a wide-open offensive system in which running lanes are big enough for Matt Waldman himself to average 5.5 yards per carry, or maybe Spiller is running behind a complete sieve of a line and is Barry Sanders-ing his way to that 5.5 each down. Because raw yards per carry is void of that context, I think it's important to look at a player's per carry output relative to the per carry output of the other running backs on his team.
In accounting for that context, we find that during the same time Spiller was churning out 5.53 yards per carry, his teammates were going for 6.26 per tote. My having been mostly aware of football for 20 years on this planet does not tell me how good that is, so it's convenient that I've already done the same calculation for every running back drafted in the last fifteen+ years and therefore know that a YPC+ mark of -0.73 is in the 12th percentile. The same process also reveals to me that Spiller's 10-yard run rate, one that is 3.18% percent lower than that of his teammates, is in the 15th-percentile. Those are clearly not impressive marks relative to a decade-and-a-half of other NFL prospects at the position, and even just within the 2022 class of runners they are 2nd- and 3rd-worst in their respective categories.
And ay, there's the rub: by all accounts, Isaiah Spiller is one of the best, most-refined and skill-nuanced runners in this class, and yet the numbers show a player that is offering substantially less per carry than his teammates are. The first and perhaps most logical question my mind goes to upon this revelation is, "ok, but how good were his teammates?" If Spiller was playing at Treasure Valley Community College with guys who walked on from the end of their high school benches, sub-efficiency relative to those guys would suggest that he simply sucks. If he was playing at Super Bama University with the best, most highly-touted group of running backs to ever come through the recruiting ranks, maybe underperforming their per carry output isn't that big of a deal. In reality, he played at a quality SEC school with a group of backfield mates who averaged 3.72 stars as high school recruits, a rating in the 70th percentile among teammates of backs drafted since 2007. His teammates were pretty good.
Because the quality of -0.73 yards per carry relative to 3.72-star teammates in context with historical prospects is not part of my knowledge base, let's look at the other guys from the past who performed similarly in similar situations. From the past 10 years, here is every running back to be drafted after having underperformed the per carry output of their 3+ star teammates while handling at least 10 carries per game for their college career (Spiller handled about 15, so we want to filter out guys who saw little volume and were clearly not serving as their team's primary ballcarrier): Myles Gaskin, Cyrus Gray, Zac Stacy, Joseph Randle, Cameron Artis-Payne, Paul Perkins, Qadree Ollison, James White, Ka'Deem Carey, Samaje Perine, Benny Snell, Tre Mason, Christine Michael, Montee Ball, Bishop Sankey, TJ Yeldon, and Trent Richardson. I think I am safe in concluding that -0.73 yards per carry relative to 3.72-star teammates is not good in context with historical prospects.
Punch, Counterpunch
This finally brings me to the common objections to my process and contentions to the conclusions I've drawn here that I have seen. I want to start by saying that silly, baseless, or illogical arguments will not be considered here, and if I've taken the time to explore the viability of a given pro-Spiller assertion, it's because I recognized that the assertion was viable and therefore explored it further in order to better evaluate the player.
Almost every pro-Spiller argument I've seen made (at least in regards to my own analysis) is made vis-á-vis the ability of Devon Achane. Achane is the "other" running back at A&M, an undersized but very fast and dynamic player who joined the squad a year later than Spiller did, and who has been pairing with Spiller to give SEC defenses fits ever since. The initial Achane-adjacent arguments centered around his usage: the refrain I heard was that Achane was deployed so differently than Spiller that a 1-to-1 comparison of their per carry output would not be a legitimate or insightful representation of their respective abilities.
This argument was compelling to me because comparing raw yards per carry does miss a lot of that situational context. If Achane was being used exclusively in passing situations or on jet sweeps where the other backs (including Spiller) were running the ball by conventional means on normal distributions of down-and-distance situations, it wouldn't make sense to include Achane's efficiency in a comparison with Spiller's, and my process as it existed would simply not know that. So, I looked into it.
Using play-by-play data from Sports Info Solutions, I looked at the deployment of both Spiller and Achane in all sorts of situations: on RPOs, on jet motion, out of shotgun, under center, in obvious running situations, on passing downs, with different personnel groupings, and so on, and found that there was virtually no difference in the rate of carries that the two backs had in the presence of these variables. This dive into the numbers did two things for me: it gave me confirmation that including Achane's numbers in a comparison of Spiller's efficiency to that of the rest of the team was completely legitimate given he was not being used in incomparable ways, and it inspired me to change my process to account for the situational context lacking in raw yards per carry.
That inspiration led me to the creation of Box-Adjusted Efficiency (or BAE) Rating, which is a comparison of a given player's per carry output relative to the per carry output of his teammates in context with their respective performances against each amount of defenders in the box that they faced, conveyed using a percentage value (defenders in the box is the tail that wags the dog of rushing success, according to research here, here, here, and here). The breakdown of Spiller's team-relative performance in each box count is below:
Spiller was more efficient than his teammates on 55 carries against 5-man boxes and on 25 carries against boxes with 9+ defenders, but lagged behind the per carry output of other Aggie backs on carries against boxes of 6-8 defenders that represented over 85% of his total rushing attempts. His composite BAE Rating of 93.2% is the third-lowest in the class and in the 6th percentile among backs who I've run box count-relative efficiency for so far (which includes the 2022 class and most of the 2021 and 2020 classes). Spiller also faced an average box count that was 0.03 defenders lighter than the average box count faced by other backs on the team, meaning that, on average, his lower raw yards per carry and 10-yard run rate numbers came in situations that were more advantageous than those his teammates were operating in.
I also wanted to explore Spiller's efficiency relative to his frequency of deployment in various rushing schemes. I'm not the one to wax poetic about the differences between iso and duo, but I do have access to the data that would tell me how Spiller performs relative to his teammates on various kinds of running plays.
Just looking at the 2021 season, there were five types of runs on which both Spiller and the collective other running backs at A&M each had at least 10 carries: inside zone, outside zone, stretch, counter, and lead. Each of those run types represented a similar proportion of overall carries for both Spiller and the other backs, and combined they represented 86% and 87% of their total carries, respectively. Their respective yards per carry on each scheme (and Spiller's output as a percentage of his teammates' output) are as follows:
Spiller was more efficient than other Aggie backs on 34 combined runs on stretch and iso and less efficient on the other three run types that made up the bread-and-butter of A&M's rushing attack.
The next pro-Spiller argument I've seen is made in regards to Achane's world-class speed. The argument is that, because Achane is absolutely blazing fast (which he is), Spiller's inefficiency relative to him is both to be expected and not a big deal.
This argument was also compelling to me, so I looked into it. My sensibilities tell me that the impact of speed would most readily appear in a player's efficiency numbers in two ways: on outside runs (where beating defenders to the edge is important) and in the open field (where breakaway speed would allow a player to maximize yardage beyond their navigation of the first level of the defense).
In 2021, Spiller had 38 carries on plays that Sports Info Solutions classified as "outside" runs, while the rest of the running backs on the team had 26 such carries. On those plays, Spiller averaged 5.6 yards per carry while the other backs averaged 5.3. Those outside runs also only represent 21.2% of Spiller's total carries, so the impact that speed is having on his ability to reach the edge probably isn't worth considering as a major reason why his efficiency is lagging behind that of his teammates on the aggregate.
I measure a player's performance in the open field using a metric called Breakaway Conversion Rate, which looks at how often a player converts chunk gains of 10+ yards into longer runs of 20+ yards. Spiller's career rate in this area is 31.2%, a mark just above average (it's in the 54th percentile). Achane's career rate is 38.5%, an 83rd-percentile open-field performance. Clearly, Achane is doing more upon reaching the secondary than Spiller is, and that has an impact on their respective per carry efficiency numbers.
One metric that neutralizes the impact of those longs runs is True Yards per Carry, which takes any yardage gained beyond 10 yards on a particular attempt and essentially wipes it out, with the resulting number the per carry output of a player with his contributions in the open field ignored.
Spiller's career True YPC is 4.02. I have no idea how good that is relative to historical running back prospects because I don't use True YPC in my evaluations, but for our purposes here, we're not as concerned with how that looks historically as we are with how it looks relative to Achane and the other backs at A&M. If Spiller is a better running back whose efficiency is lagging behind Achane's because of the advantage that Achane's speed gives him in breaking off long runs, then we would expect Spiller to be more efficient with those long runs removed from the equation.
And yet, Achane's career True YPC is 4.77. He is ripping off breakaway runs more often than Spiller is while also producing more efficiently when he's not doing that.
The combined True YPC for non-Spiller Aggie backs during Spiller's career is 4.31, and the combined True YPC for non-Spiller and non-Achane Aggie backs during the same time frame is 3.78. As I said before, I don't keep track of True YPC for running back prospects, so I'm not sure how impressive it is that Spiller outperformed the other running backs on his team by 0.24 True YPC (at least he did when you remove the most efficient runner on the team from the sample), but if that were his YPC+ mark it would be a 39th-percentile number. It's also worth mentioning that the only season in which Spiller outperformed the other Aggie backs in True YPC is 2021, as in 2020 and 2019 he posted True YPC+ marks of -1.46 and -0.16, respectively. That 2019 number came during a season in which Achane wasn't even on the team yet.
(speaking of Achane not being on the team, Spiller's team-relative efficiency numbers in a fantasy world where Achane never touched the ball at Texas A&M would be good for a YPC+ in the 41st percentile and a Chunk Rate+ in the 22nd percentile. We can Thanos-snap Achane completely and Spiller was still not impressive relative to his teammates.)
All of this analysis sort of dances around the major question I continue to have: is Achane so good that we should ignore his contributions when evaluating Spiller? My evaluation of Spiller already accounts for the fact that his teammates were a talented group, but the pro-Spiller arguments all seem to be some variation of "but Achane is really good." My position here is that if we are going to asterisk the Spiller evaluation due to the talent-level of Achane, then I should also retroactively asterisk the evaluation of other guys who played with high-quality teammates. TJ Yeldon played with Derrick Henry and Eddie Lacy at Alabama, Samaje Perine played with Joe Mixon at Oklahoma, Karlos Williams played with Dalvin Cook at Florida State, LenDale White played with Reggie Bush at USC, Elijah Holyfield played with D'Andre Swift and Sony Michel at Georgia, Montee Ball played with Melvin Gordon at Wisconsin, and so on. Is Achane so much better than Henry, Mixon, and Bush that Spiller deserves the benefit of the doubt where Yeldon, Perine, and White did not? I'm skeptical.
As I finish up, I want to be very clear about what my evaluation of Spiller vis-á-vis his efficiency relative to his teammates (and specifically, relative to Achane) is not.
It is not holding the fact that Spiller shared time with Achane against him. Spiller's production profile is good, and there are many other players who also either shared time with talented teammates or just didn't command incredibly large portions of their team's offenses but had quality production profiles nonetheless-- among them are Mixon, Bush, Swift, and Michel, as well as guys like Travis Etienne, Felix Jones, Jonathan Stewart, Ronnie Brown, and Javonte Williams and Michael Carter.
It is also not dismissing Achane's talent or pretending that he's not good. The appropriate composite scores in my running back model account for how talented a guy's teammates were, and the 70th-percentile star rating that Spiller's backfield mates came into college with is the reason that his Rushing Efficiency Score in my model is a 28.4 out of 100 and not a 12 or 15 like his raw teammate-relative numbers would have him at. I'm already adjusting for how good Achane and those guys were, and Spiller is still not impressive in light of that adjustment. Even outside the scope of my normal process, I think I've proven in this article that I've taken Achane's talent seriously and tried to find ways in which his usage or skillset would throw off my evaluation of Spiller (though I've been unsuccessful in substantiating the ways in which that is supposed to have happened).
Finally, my position on Spiller is not the result of manufactured contrarianism or prospect fatigue. My process is simply the positioning of data points in context with those of historical prospects, and the results are the results. I didn't have a take on Spiller two years ago that I have since changed out of boredom, and I didn't perceive a general sensibility regarding Spiller that I intentionally aligned myself in opposition of.
I will conclude with this: there are baseline boxes that we want our running back prospects to check, and Spiller checks many of them. He's big, he was productive, and he's a good pass-catcher. But while being 6'2 and making good money might get you a lot of matches on Tinder, they don't mean you fuck good. Likewise, size, pass-catching ability, and solid production on a good SEC team will get Spiller a foot in the door on an NFL roster in the form of early draft capital (and he very well might be a fine NFL player, especially considering that rushing efficiency is driven more by situational factors than it is by the name of the guy carrying the ball), but given what he put on wax as a college runner relative to what the other guys on his team did, and especially given what that means relative to the efficiency of historical running back prospects relative to their teammates, you should be scared to death that Spiller goes the way of Trent Richardson and Montee Ball. Sometimes the guy who seemingly has it all just can't fuck good.
Comments
Post a Comment