Just how bad has the Randy Gregory pick been?

In recent weeks, we received the long-awaited news that Randy Gregory had formally applied for NFL reinstatement (news that was officially confirmed earlier this week by the Dallas Cowboys). That is all well and good, but Cowboy fans seem to have wisely conditioned themselves to expect nothing from this development. Despite the fairly surprising return of Josh Gordon to the field after two-and-a-half years of suspension, it is best to tread lightly in the face of the league's harsh treatment of Dallas when it comes to punishments.

Between these expectations and the very limited impact he had a chance to offer prior to his banishment, Gregory has been labeled a "bust pick" by the fanbase of the team, as well as leaguewide. Indeed, many now firmly declare the use of the 60th overall pick of the 2015 draft on the pass rusher to be an outright mistake.

That's fair enough, I suppose. Or is it? There is absolutely no doubt that the selection has produced disappointment, and the pick most certainly would be a bust barring a major turnaround going forward. But does that make the move a literal mistake? What did the team sacrifice in order to take the Gregory gamble? Have we overshot the mark in terms of our projections for alternative directions the team could go in, or has our intuition of the cost been on the spot?

Setting The Bar

Before we can consider whether the Gregory pick was a mistake, we need to first establish the requirements for such a label. To be a mistake, the decision has to have been made in the face of good reason to have acted otherwise. That means having a sense of what Dallas was projecting to get back in the event that the move worked out, and of how likely it would be for the move to fail.

It turns out that it isn't all that difficult to establish these expectations; it's so straightforward, in fact, that it's a wonder it hasn't been done before. First, we can create a simple (and rather accurate) sense of the upside of this pick. It has been firmly established that Randy Gregory, off-the-field issues aside, was seen as a "top 5 talent" in the draft, and that was across a rather broad consensus (I won't take the time to present citations here, for reasons you'll see in a moment, but feel free to investigate if you have your doubts). Now, to be fair, long is the list of players with a certain positive acclaim who ended up being drafted later than the consensus, so we won't lock in this this broad consensus. But there is no question that, as a pure football player, Gregory had the talent level of a player who would go in the top half of the first round. To that end, we'll look at the established values for picks 5, 10, and 15, along with that of the actual draft pick that Dallas used on the player.

What expected values? Well, we have two respected measures of draft pick value that we can reference: the "Jimmy Johnson chart" that was created based on historical trades and for decades has served as the baseline for draft pick trades, and the "Chase Stuart chart" that was the product of an investigation into the historical production results for draft slots. The CS values are more conservative for higher picks than the JJ values, lending to the formation of a nice value range. Here are the results:
It is more difficult to objectively determine what the odds were for the Gregory selection to not work out, given how much factors in to such a calculation. For now, rather than attempt to pinpoint this probability, we can simply determine the "bust percentage" that would result in "breaking even" for each of these three football-talent-only draft slots for Gregory. The CS and JJ values naturally include players who exceeded expectations, those who disappointed, those who got hurt, etc. All that was already baked into Gregory's "high first round" draft status. What remained unaccounted for was the very negative influence of the possibilities that drug use, drug suspension, and/or other issues stemming from his anxiety-issue/bipolar status would lead to Gregory failing outright in his NFL career. To be as conservative as possible, we can assign an outright "0" value for Gregory's career in the case of failure due to one of these factors. All that remains is to calculate the weighted average probability assigned to this "0" that would result in dragging down the CS and JJ values for each of the three draft slots to 8.5 (for CS values) or 300 (for JJ values), as follows:
Wow. I won't lie - when I first thought up this doing this exercise and writing about it, I did not expect the results to be this stark. I knew that the "break even" would be more favorable for Dallas than fans seem to allow for this draft pick, but apparently even my own intuition was way off.

In case it isn't clear, here is what these results say: even if we assign an ultra-conservative football upside to Gregory of the 15th overall pick at Chase Stuart values, Gregory's "bust rate" due only to his personal demons (as in, not even including the non-zero chances of him busting due to football weaknesses, injury, etc) would have had to be worse than 50% before the move failed to be "profitable". It is easy to overestimate the typical return offered by the 60th overall pick, and therefore easy to underestimate how bad Gregory's chances would have had to be be as of draft day for the gamble to not be worth it.

Looking At The Actual 2015 Draft

While it is still early in the careers of 2015 draftees, the next exercise we can run is to look at the results thus far for other players taken in Gregory's draft. Dallas selected Gregory at 60 not just on the basis of his own profile, but also on the basis of the other prospects available. By seeing how things have turned out for Gregory alternatives, we can get a sense of what Dallas missed out on by rolling the dice.

Earlier Picks
Let's start with the players taken ahead of Gregory. For his selection to have been a mistake, surely things have worked out significantly better for players whom other teams saw as a better choice, yes? This sort of analysis is inherently subjective, so I apologize that there is certainly room for debate in terms of how players should be labeled/qualified, but to start there are already 4 players (Devin Smith, Dorial Green-Beckham, Jalen Collins, and Senquez Golson) who appear to be out of the league, and another (Kevin White) who has played fewer games than Gregory has managed in spite of his suspensions. Those players are unequivocal "busts".

Next up, we have the low performers. To keep it simple, I took each draftee's career AV total and divided it by games played. Anyone who totaled a 2-or-lower AV/game (except for Dante Fowler, as he seemed undervalued by AV) was counted, yielding 7 more names. Anyone who was already unceremoniously dumped by his team (so, not counting someone used in a trade of real value) was then counted, producing 4 more names. DJ Humphries has played in just 18 games since being drafted, and so he makes the list too.

That's already 13 out of 59 players (22%) who have been roughly Gregory-level disappointments so far in their careers, out of players taken ahead of him. Considering the higher expectations/demands of players taken earlier, another 14 players (at least, erring on the side of caution) have been major disappointments with futures in question. Just like that, we have almost half of all the players taken ahead of Gregory as leaving a bad taste in the mouths of their own fan bases. Is a Nelson Agholor (not included in the 27), for example, really a more successful draft pick after you factor in the price of being drafted 40 spots earlier? It's really quite rough.

Picks "In The Range" Of 60
It becomes even more informative to looks at the players selected around pick 60, i.e. the ones of a similar rating to Gregory. Following a similar procedure, we see that Golson, D'Joun Smith, and Jeremiah Poutasi all were essentially out of the league in 2017 in the group of players from pick 50 to pick 70. Another 6 players fell into the 2-and-under AV/game group, with one more player falling just outside of this threshold.

Do the math, and that's a whopping 9 players out of 20 (not including Gregory himself) who have offered a similar level of minimal return. Unlike Gregory, whose issues to this point have not surprised, many of these picks were made with "safe contributor" as a goal, and their teams haven't even gotten that, to say nothing of the limited (or already finished) upside that came with these high-floors-that-weren't.

Alternatives To Gregory At 60
For the Gregory pick to be an outright "mistake", it's most useful to look over the players Dallas could have selected in Gregory's place. Unfortunately, 2015 was not a leaked draft board year, so we wouldn't be able to do more than guess at which players the team actually had in the crosshairs. Instead, we can once again consult the board, checking out picks 61 to 80.

Again starting with players out of the league or not on active rosters in 2017, Smith and Poutasi are joined by Owamagbe Odighizuwa, Garrett Grayson, and Alex Carter. That's right, a solid quarter of the 20 players taken immediately after Gregory have seen their careers virtually end with no greater impact than what Gregory has offered to this point. Another 3 fall into the poor AV/game group (though Clive Walford has done enough over his 44 games that I wouldn't quite count him), though awesome-named Hroniss Grasu takes his place with only 14 games played to this point.

Again, that's approaching half (8 out of 20, or 40%) of the entire group that have been objective failures, most of whom have no hope to offer anything more going forward. It behooves all of us to realize that outcomes get quite hairy even at that relatively early juncture of the draft.

But while we discuss alternatives, why not also look at the successes? The most obvious name to nominate is that of Frank Clark, another pass rusher of superior talent relative to his draft spot who was taken just two picks after Gregory and has put up an impressive 22 sacks to open his career. Well, if you think that Gregory was a risk, Clark was risk squared, what with one felony (for theft of a computer) along with a lovely domestic violence charge against a girlfriend (in which Clark skated by thanks to a plea deal) already on his record as of draft day. Given the massive scrutiny the Cowboys were already under following the signing of Greg Hardy less than two months earlier, the team would have been hard-pressed to double down by taking Clark. While Gregory can not be excused for his own very real mistakes, it's pretty clear that he was more likeable and guilty of far lesser crimes than Clark, to say nothing of his superior football talent.

Perhaps the next-best candidate, AJ Cann did not play a position of desire for Dallas at the time. Tyler Lockett is the only Pro Bowler of the bunch, but that was only as a return specialist, and he is more of a supplemental player than anything else.

Really, the best and potentially only truly appealing alternative to Gregory - and the name I think might have been next on Dallas's board - is that of Tevin Coleman. Coleman has been a fairly dynamic playmaker at RB while stuck behind Devonta Freeman in Atlanta, and he could have been the non-first-round RB the team had its eyes on as a suitable replacement for DeMarco Murray. If he would have been the name on Dallas's card with Gregory out of the picture, chalk that up as a potential regret.

Conclusion: Just What Is A "Mistake" Anyways?

Let's get the more surefire point out of the way right here, right now: the selection of Randy Gregory by the Dallas Cowboys at pick 60 was not in any way a "mistake". If you actually care to use language properly, it's key to take the word for what it means: a "blunder", according to Merriam Webster, and a blunder itself is written up to be a product of "stupidity, ignorance, or carelessness". Even the best of decisions, ones with 90%+ odds of working out, can still fail to produce a positive outcome, and that doesn't retroactively turn it into a mistake. Mistakes can only be made in the moment of decision, not on the basis of results (though results certainly can point us in the direction mistakes that have been made).

If the rest of this post isn't enough to convince that no mistake was made here, that is proven simply by reviewing the response to Dallas's pick. The team desperately needed pass rush help; it had an elite pass rush prospect fall into its lap, whatever the circumstances. The team was picking in a draft slot that has FAR worse returns than the words "second round pick" evokes in the average fan, lending far more upside at the pick than a team coming off of 12 wins usually are able to get. And most importantly, the downside to picking Gregory was already quite well understood: he would be entering the league already in the drug program and thus would be at risk for early drug suspensions, and bipolar/anxiety questions imposed further risk of transgressions. Gregory was viewed as well-spoken and of good general character, and did not have other problems on his resume, so he was not a complete character mess that some other talented prospects prove to be. No team wanted to spend a truly valuable pick on a player with that much established risk, but it was just as sure that before long the reward would easily outweigh the risk. The post-pick response makes it clear that pick 60, to a team seen to be as good as Dallas, and to a team with as much need for a player of Gregory's ilk, was an appropriate spot, one that left a lot of other NFL fans feeling uphappy.

And remember, that reaction was in spite of the fact that these same fans perceive pick 60 to be "safer" than it actually is.

That reality is what allows us to question whether this pick is even "regrettable". Is it? If you are the kind of person (and there's nothing wrong with this) who regrets a decision made for the right reason if it ends up not working out, then it probably is (unless Gregory can get his act together going forward). "You always double down on eleven"...unless it's the "particular case" when it still doesn't work out.

Personally, I trust a process built on good decisions, even when results in the short term break the wrong way. To me, Randy Gregory was a 50/50 contest in which it cost me $100 to win $1000. Anyone who isn't absolutely empty on cash should leap on such an offer, but that doesn't mean you still won't lose out a good half the time. The idea is to stack the deck with as much such situations (or better) as possible, so that you can push the sampling to a size that leads you to the average $500 profit. Randy Gregory in 2015 was such an opportunity. The final piece to factor in to this analogy? Remember that you essentially have to invest your $100 one way or another, and we now know that the other "games" at pick 60 offered up $200-$300 returns at maybe 60/40 win/lose odds. You can have your $120-$180 average winning options; give me the big payoff at only slightly higher risk.

Another user-created commentary provided by a BTB reader.