The NFL needs better tools for evaluating whether or not a particular team’s draft class was a success or not. By now, most fans acknowledge there should be a moratorium of at least three years before we attempt to declare if a class was “good” or “bad,” but even most of those valuations are based on imperfect tools. Perhaps foolishly, I set out to create a less imperfect tool, and discovered why so many of the minds much smarter than mine accept these lackluster tools: this stuff is work.

We’ll get into what my initial plan had been and why it was relatively quickly abandoned, but I did make a new draft evaluation tool, even if it is not as complex or deep as I would have liked. This tool, called a Total Points Earned Contribution Score (I know, I can really name ’em, huh?) builds on the Total Points Earned (TPE) metric developed by Sports Info Solutions (SIS). TPE is a player-specific version of EPA, which we’ve looked at here. TPE ascribes a point total on a per-play basis for how much each player contributed to the EPA generated on a particular play. Is it perfect? No. Is it more trustworthy than PFF’s grading system, I believe so. The work below is largely inspired by wanting a version of this ESPN article that considered the draft capital investment in each pick.

Total Points Earned Contribution Score

The Total Points Earned Contribution Score (TPECS) starts by using the number of snaps each player contributed (S) as a factor of their draft round. Each draftees snap count was taken from PFF for the regular season. The snap count total was divided by the product of the most snaps played at that position (M) and a variable based on draft round (R).

TPECS = TPE*(S/(M*R))

The draft round variable was determined by how many snaps can or should be realistically expected from picks in that round. So, first round picks get an expectation of playing 50% of snaps, meaning their draft round variable is set at 0.5. The rest of the rounds broke down as such:

• RD 2 = 0.4
• RD 3-4 = 0.3
• RD 5+ = 0.2

The Bills’ Scores

The visuals for this article may be too small to view, so if you’d like to see the chart, you can access it here (You can also find TPECS career totals and yearly averages there too). Some of the highlights:

• The best rookie class under this format was in 2018. No surprise there. But the class wasn’t carried by Josh Allen. It was Tremaine Edmunds’ TPECS of 75.74 that was far and away the best of the class.
• The single best TPECS in a rookie season was Tre White’s 90.26 in 2017.
• The most productive season a Bills’ draft class had was not the 2018 class in 2020. It was the 2018 class in 2021 with a TPECS of 444.4, beating the 2018 class’ score in 2020 of 407.33.
• The best season on a rookie deal while accounting for original draft position was Allen’s 2021. The best non-QB value was Matt Milano’s 2019 (154.34). Milano’s 2021 and 2022 seasons scored a higher TPECS, but he was on his second contract by then. Dane Jackson’s 2022 is not far behind in value (141.42).
• The best-performing draft class for TPE/Snap for their careers to this point is the 2021 class, which is really bolstered by Damar Hamlin’s 2022 performance.

I would love to tell you I have plans to do the other 31 teams so that the BIlls’ TPECS could be put into perspective, but I can’t. If any of you would like to hire me full time to complete it and write about football all the time, I’ll do it tomorrow. But until I get that DM….my hope is that you found value in the exercise and thinking about the draft in a more comprehensive way. It has also made me appreciate the tools we do have.

What TPECS Doesn’t Include

The tool I initially wanted to create included several more pieces that were simply too involved to broach for even just the Bills let alone the whole league. If there was infinite time and I stopped sleeping, this tool would include scores for how well a player ranked in the three most important stats for their position. The scores would modulate based on how where that player was drafted, how well they scored in the stat, and how many years they had played in the league (it is reasonable to expect a third-year QB to perform better than a rookie QB). I recruited some of the best minds in Bills’ content creation to help ascertain which stats are best for each position, and I created the modulation chart.

I pretty quickly realized that tracking all of that down as one person doing this in his spare time was untenable. But I think there is value for you all to see what is considered the most important stats by positions, and I’m tossing in the modulation chart because I did the work and you can write your own articles if you don’t like it please, and thank you.

The stats are from a variety of sources including ESPN, PFF, Football Outsiders, and SIS. Erik Turner helped pick out the LB stats. Ant Prohaska gave an assist with the DL. And Bruce Exclusive recommended the DB stats, and the QB ones are essentially a truncated version of his STEW metric. There might be another article in reviewing why each stat was selected.

The modulation chart at least lets you see there was serious consideration given to the attempt and offers a framework if you were ever interested in trying it out for yourself. For instance, if a first-year RB, drafted in RD 2, finished 25th in EPA/Att, that would score 1 point. The points from the rankings would then become part of an equation with TPECS, but I never got far enough to justify calculating that formula.

Potential Changes

• Switching out the draft round variable for the draft slot valuations from Chase Stuart’s Draft Chart (or the new Spielberger). This would give a score that was specific to each draft slot, not just the round they were selected in.
• I’m eventually going to include TPE/Snap for a career. I want to see how that corresponds to Approximate Value.
• Highlights to indicate when a player is on a second contract.

I’d love to hear any thoughts anybody might have about this process – especially if that includes how to make it faster.

Updating the Football Outsider’s Chart

When I finished the TPECS chart, I still wanted to offer some league-wide context for the Bills’ drafting. When I did that first draft evaluation article, I included a chart that Football Outsiders had done a few years ago, but it only went through 2019. FO included their formulas in that article, so I was able to bring the chart up to date by pulling data from Pro Football Reference and Chase Stuart’s website. You can at least compare how the Bills have drafted through 2022, even if this chart has some problems. I wrote the following in HOW TO EVALUATE IF AN NFL DRAFT WAS A SUCCESS to describe the FO chart:

There are a variety of ways to capture draft capital, beginning with the now long-fabled Jimmy Johnson trade chart, followed by the Rich Hill chart, and the Chase Stuart chart. They each have pros and cons but their goal is important – find translatable values for draft picks. Football Outsiders has done extensive work, and their results can be found in Ben Ellinger’s NFL Drafting Efficiency article from 2020.

Ellinger’s article begins with determining each team’s amount of draft capital based on the Chase Stuart chart. Each pick a team has that year is assigned a value (there are small differences year over year because of the compensatory pick formula), and then each team’s total is divided by the total for the full draft, generating a percentage of draft capital for each team. For the return on how each team spent those draft resources, they used Career AV from each draftee, added the Career AV from each of that particular year’s draftees for each team, then divided each team’s total by the total Carer AV from that year. This process created a percentage depicting how much of the possible AV from that year’s draft each team garnered. Football Outsiders went the final step and divided draft return by the draft capital year by year.

Bills rank 15th in 5 Year Average, but, as we’ve seen with the Chiefs, one good year can swing that perception all the way back, because the popular opinion is that KC has drafted really well, but their 5 Year Average ranks 29th.

Conclusion

Drafting is difficult, and so is assessing the draft in a comprehensive fashion. Most of our tools for doing are barely on the surface level of analytics because of the difficulties in collating that data into manageable formats. Someday we might have access to systems that allow for easier data management for sports analytics, but for now, we’re limited by time and capacity, or at least this author is.

Starting next week, we’ll be examining how 2023 prospects fit into the Bills’ archetypes that we shared in this series: Bills’ Measurables (the OT article has the links to all the others). If you’ve followed me since last draft season, you know what’s coming.