We’re about to get into the heat of the offseason, and with rookies on NFL rosters, prospect evaluation is about to become less important (until next offseason). However, before we send off the rookie evaluation process, I did want to throw out an idea I’ve been working on. Last offseason (before I joined FantasyGuru), I started a project with my friend Nate Liss (@anoutragedjew). Something we talked about for years was the lack of context surrounding three main areas of college production research:
- The effect that teammates have on RB and WR college careers
- The collective strength of an individual prospect’s opponents throughout their career
- Rushing and Kick/Punt Return production and its importance in WR production profiles
Without addressing those areas, and creating a process for contextualizing those in prospect production research, a statistical or analytical process will always be lacking, and naysayers (folks that hate numbers!) will always have avenues to discredit all the amazing work that’s already been created. While I did have to eventually step away from that work (Though Nate and a few others are continuing!), we’d created some pretty cool ideas, that are at least conversation starters for how to evolve WR prospect evaluation. If you’ve been following along with my prospect evaluation process closely this offseason, then you’ve noticed a lot of focus on these two areas. Because of that, it’s important for subscribers to understand the logic behind valuing these factors.
The evolution in prospect research, from an analytical and statistical perspective, has been fantastic over the last five seasons. The emergence of research evaluation tools like marketshare (MS), dominator rating, Breakout Age and the Phenom Index have given an incredibly strong base to build off. Ideas and innovation in prospect research all lead back to these ideas at one time or another, and I’ll speak for myself in this statement — I’ve learned a lot from these concepts and value them highly in my process.
Intro/Theory
In order to explain Teammate Scores, it’s important to explain why there’s a need for this type of context. I’m certainly apart of the “less is more” idea when working with data, so… what does this address?
A great raw method (mentioned above) for evaluating a players value in a college offense is MS. MS is the percentage of production (for WRs specifically, it can be used for receptions, receiving yards or receiving touchdowns) accumulated by an individual out of the entire offense.
Example:
In 2018, Mississippi WR A.J. Brown had 1,320 receiving yards. Mississippi had 4,157 total passing yards in 2018.
Player receiving yards / team receiving yards = Market share of receiving yards
1,320 / 4,157 = 31.7%
So, in 2018, Brown had 31.7% of Mississippi’s receiving yards.
For a general range, anything above 25% is solid and means that player was a prominent option in their offense. In a given year of a prospect’s career (Freshman, Sophomore, Junior and Senior), this can give an initial look into what a college team thought of that player, and how integral they felt he was in matriculating the ball down the field (shoutout to Hank Stram).
This doesn’t take into account a few important factors, of course, like teammates, which will be the main focus of today. Usage in a college offense is dictated by a few factors, and one of the main factors is the quality of depth on the roster. All college depth charts aren’t created equal, and all production is not created equal, either. This becomes very blatant when comparing prospects from power 5 (P-5) conferences (SEC, BIG 10, BIG 12, PAC 12 and ACC) to prospects from smaller conferences (WAC, AAC, MWC, Sun Belt, C-USA). Some P-5 prospects go their entire career without dominating the production in their college offense, while small school prospects (the ones that get drafted, at least) are so significantly better than everyone around them, coaches have no choice but to force-feed them the ball. There’s a lot of reasons why a small school prospect has a much easier path to production than it’s P-5 counterparts, but arguably the most blatant is the depth chart.
This extreme divergence in depth chart talent is displayed every season in the NFL draft. Since 2010, there’s been 77 WR selected in the top three rounds of the NFL draft; 67, or 87% were from a P-5 conference. Why is that? In terms of college production, there’s almost no difference between raw production (receptions, receiving yards and receiving touchdowns) between the conferences. In fact, some of the most prolific seasons of the last decade have been from small conference prospects. Four of the top six in receiving yards in 2018 were from small schools. UMass WR Andy Isabella, Colorado State WR Preston Williams, Hawaii WR John Ursua and Fresno State WR KeeSean Johnson all produced above 1,300 receiving yards and dominated their college offenses. Isabella ended up being drafted in round two (!), but Williams (UDFA), Ursua (7th) and Johnson (6th) weren’t so lucky. In comparison, here’s what P-5 WRs look like from a draft position perspective. Oklahoma State WR Tylan Wallace and Alabama WR Jerry Jeudy aren’t draft eligible, so after excluding them (though it’s very likely both end up as top two round selections in 2020), there’s Texas Tech WR Antoine Wesley, Mississippi WR A.J. Brown, Oklahoma WR Marquise Brown and Iowa State WR Hakeem Butler. Wesley (UDFA) flamed out in the pre-draft process, but (A.J.) Brown (2nd) and (Marquise) Brown (1st) were drafted in the first two days of the NFL draft, and Butler (4th), despite falling much further than expected, still got his named called relatively early. This has been a constant theme for two decades. The NFL simply doesn’t value small school production the same as P-5 production.
The question is then — WHY does the NFL not value it the same? One piece, I’d argue, is the quality in depth chart competition. Earning a featured role on a P-5 conference team is simply more difficult. Especially at the top, with teams like Ohio State, it’s very likely that multiple NFL prospects are competing for the same targets and production. A great example from 2018 would be Jazz Ferguson. After recording just two receptions for 17 yards with LSU (losing the competition for targets with NFL-talent level prospects) in two seasons, Ferguson transferred to Northwestern State. Ferguson went onto produce 66 receptions, 1,117 receiving yards and 13 touchdowns in his 2018 season. He was dominant, but NFL teams didn’t value the production. Preston Williams is another 2019 example. After two unproductive seasons at Tennessee, Williams transferred to Colorado State and produced 95 receptions, 1,345 receiving yards and 14 touchdowns in 2018. Williams went from a player struggling to gets snaps to being one of the most productive WRs in college football. Williams ended up going undrafted. Is there extra context to Williams and his college career (Off-field issues)? Sure, but it’s fair to argue he would’ve never been a big producer had he stayed at Tennessee.
With that considered, there’s a need to filter MS and other production-based evaluation methods for the factors that distinguish P-5 and small conference production. Teammate Scores would be one of those.
This is ESPECIALLY true for breakout age and evaluating early career production. It is incredibly difficult for P-5 prospects to get on the field early, especially at the top-end schools, because there’s almost always incumbents with NFL potential on the roster. Breakout Age is defined as the first season in which a receiver achieves 20% or higher college dominator rating. Dominator Rating, for a WR, is the average of team receiving production (receptions, receiving yards and receiving touchdowns) in a given season. Considering that, it’s clear why that would be a daunting task for some prospects.
Let’s use Ohio State WR Parris Campbell as an example. In his first two seasons, Campbell accumulated 121 receiving yards and saw most of his snaps at RB and on special teams. Well, why was that? Was Campbell a deficient player that needed to learn, or was there legitimate obstacles in his way? In 2015, the top three WRs on Ohio State’s depth chart were Michael Thomas (2nd round pick), Braxton Miller (3rd round pick) and Curtis Samuel (2nd round pick). Add in TE Nick Vannett (3rd round pick) and there simply wasn’t a lot of room for young players to break through. Not only were those players incumbents, but they ended up being highly sought after NFL prospects. Miller was fresh off three seasons playing QB, and in order to truly switch positions, there were very likely guarantees at snaps and playing time. Guaranteed playing time for older players who stay and highly touted incoming recruits are an enormous factor at P-5 programs, and sometimes even if a player warrants more action, they have to wait their turn due to those circumstances. This was very real for Campbell. In 2016, Campbell had to deal with Samuel again, a player with a very similar archetype, along with Noah Brown (7th round pick), K.J. Hill (Still at Ohio State), Terry McLaurin (3rd round pick) and Binjimen Victor (Still at Ohio State). Because Samuel played a similar role and was the incumbent, Campbell had to find snaps where he could. With Samuel leaving after the 2017 season, Campbell fully broke out in his Junior and Senior seasons, but without factoring in his situation, his production profile was already very unappealing and too far gone.
These types of situations happen every season, and while most prospects tend to figure it out AT SOME POINT in their careers, an early career breakout OR late career fizzle out could be plausibly explained. Small school prospects will never, at any point, have to deal with this type of competition for production in their own offense. The same goes for some even lower level P-5 conference teams. While it’s still a step up in competition, there are a lot of P-5 programs that won’t have a RB or WR drafted in the NFL draft each year. This is why a full evaluation on each situation is warranted.
TEAMMATE SCORES
Teammate Score – an evaluation method that determines how much competition for production there was for a prospect in their college career.
So, how would something like this be calculated you ask? Well, there’s an answer.
The NFL draft is a great way to determine how the league evaluates talent, and every year it’s a great proxy for personal process. The NFL puts a significant amount of resources into the Draft, and while the general mantra is teams don’t know what they’re doing, in a macro sense, it’s a highly efficient beast. If looking for an in-depth breakdown of this idea, check out this article discussing RB and WR draft position.
A really important thing to note. Nate and I discussed this for hours before settling on “who” would qualify as players that would shift a WR prospects production. Ultimately we landed on WRs and TEs, and not RBs. While some RBs are significantly involved in the receiving game, many of them aren’t stealing WR routes and down-field targets.
So how can draft position be used to evaluate a prospect in hindsight? For a long time, I’ve used the general mantra “the earlier a player is drafted, the more productive they were in college.” While that’s not true in every case, I assumed it was true over 100s of prospects.
So, I did some research to find out. I went to profootballreference.com, queried “Every WR and TE, Drafted, since 2000”, found the 528 WRs and 260 TEs who went to Division I schools, and looked up the best season MS of Every. Single. One.
This was a difficult task, and the number of hours of my life wasted looking up best season MS of random late round TEs are something I’ll never get back. However, I got answers. Is it true that, generally, the earlier a player is drafted, the more productive they were as a college player? Yes, Yes it is.
Here is the average best season market share, separated by draft round, for all WRs and TEs since 2000 (Samples are 528 WRs and 260 TEs).
A beautiful, orderly, set of percentages. I would like to say I wasn’t actively rooting for this result, but I was. If I wasted all that time to find out best season MS was a big jumbled mess and had no correlation to draft round, I’d be very sad. BUT HERE WE ARE.
Think of it this way. On average, a WR who would eventually go on to be a first-round selection, in their best season, took up 37.5% of their team’s receiving yards.
A scale was built from this set of numbers, subtracting a seventh-round TE’s best season MS (11.9%) from each of the other best season MS’s. The scale starts at two so that a seventh-round TE wouldn’t be a zero.
Then, all that needed to be done was to collect and give a score to each Drafted WR and TE the (529!) WRs played with in their college careers. Yipe. A painful process, but here we are.
There was also the added element of prospects who haven’t been drafted yet. How can there be an NFL draft position method of prospects that are still in college? Projecting draft position of players still in college is something I’ve been working on for years, and while it will never be a perfect system, there’s enough information (historical production thresholds and film) out there for a general projection on underclassmen. These projections are conservative, and there are only a few first-round grades given out, but it’s important to get SOMETHING on the board for prospects we’re confident are future NFL players.
Here’s Ohio State WR Parris Campbell’s teammate score, as an example.
Teammate Scores for the 2019 WR Class
*(P) represents “projected.” A prospect will only have a “projected” Teammate Score until everyone on their final season roster has graduated (In some cases, that can take up to four years).
As you can see, a Teammate Score on its own isn’t worth much. A lot of lower quality prospects who weren’t productive at P-5 Schools have big scores. What’s important is, as mentioned above, this establishes a scale for how tough their competition for production was. A Teammate Score should be used to adjust a prospect’s profile, not replace it. It’s a plug-in to help optimize an already working system.
Until next time.
