Yesterday’s release of the 2017 compensatory draft picks by the NFL came even earlier than usual this year. However, this is for good reason, due to the fact that this is the first year that compensatory picks may be traded.  As several people have suggested (Miguel Benzan being one of the first I saw), it’s only fair and proper that teams should have precise knowledge of all the draft picks they will have before the combine–an event in which trade discussions really start to ramp up. I would expect future official compensatory pick releases to come on or near the Friday before the week the combine starts.

Now that the release out, it’s time to keep me honest and see what I did well in my projection, and what I got wrong.

Highlighted in green, I got 24 of 32 awarded comp picks correct with the correct player and the correct round. In addition, I got 6 of the 7 comp picks that missed the 32-pick limit correct, for a total of 30 out of 39. Although the press release did not state the precise order of the comp picks over the 32-pick limit, it did state the number of comp picks that were missed: 1 for the Rams and Packers, 2 for the Steelers, and 3 for the Cardinals.  With the exception of the Rams pick (see below), I got this quantity correct–and also noted how odd and painful it was for a team to have 3 comp picks miss the 32-pick limit, something I don’t think has ever happened before.

Highlighted in yellow, there were 4 comp picks that I had the correct player for, but was off by one round.  The round cutoffs are always the most challenging to predict, and I hope with 2017’s comp picks in the books I can refine those cutoffs further.  Thankfully, I anticipated that all 4 of these round misses were within the margin of error that I had noted in my projected.

Highlighted in red, there was just one pick I missed without any anticipation that I may have missed it. I had projected the Rams to get a 6th round comp pick for Nick Fairley, on the basis of his APY being $3 million over three years. However, what I did not notice is that Fairley had for all practical purposes signed a one year deal with two void years solely for the purpose of spreading out his signing bonus (a practice that the Saints are notorious for).  When I ignored his Paragraph 5 base salaries on those void years–mere placeholder dollars that were never going to be paid out–this reduced his APY to $1.583 million. Upon that adjustment, Fairley was placed exactly where should have been–a 7th round value that just barely missed the 32-pick limit.

Highlighted in blue, there were 4 picks that I missed both the round and the player on, but were misses that I did anticipate might happen.  One of them was due to Sam Young not qualifying that opened up a 5th round comp pick to the Dolphins for Rishard Matthews. (See this article to explain why.) The very good news is that Young was the only player near the qualification bubble that I missed on, suggesting that I’m tightening up that part of the formula well.

The other three, however, all resulted from my most consequential miss this year. I had thought the comp pick formula would only count the first year of the contracts of Russell Okung and Kelvin Beachum, but instead the formula counted the entirety of them.  This caused Pittsburgh’s 5th for Beachum to be upgraded to a 3rd, gave Seattle a 3rd for Okung instead of a 5th for JR Sweezy, and gave Denver a 7th for Vernon Davis instead of a 3rd for Brock Osweiler.

The best precedent that I am seeing from this point, that also involved the Broncos, comes from Ian Whetstone:

This is consistent with how LB Ian Gold's structurally similar contract was valued for compensatory purposes way back in 2005. https://t.co/DIpVOYeEyI

— Ian Whetstone (@IanWhetstone) February 24, 2017

As he notes here, Gold had a $9.6 million option bonus in his contract.  If it’s any consolation, AdamJT13, the pioneer of projecting compensatory picks, made a similar mistake to me back in 2005 from Gold’s contract (bolded mine):

The one place where there is the biggest question is between the third and fourth rounds — in particular, the comps for Bert Berry ($5.0 million, 16/16, Pro Bowl) and Damien Woody ($5.17 million, 16/16). I’ve projected that both of them will be fourth-round comps, although they could end up being third-rounders (one Denver newspaper has even reported that the Broncos would get a third-rounder). However, in each of the past two years, there have been players who signed for more than $5 million per season and were worth only fourth-round comps, so that’s where I’m projecting those two. One factor could be Ian Gold being released by the Buccaneers and being re-signed by the Broncos. He had signed for $5.464 million and, if the Broncos’ comp pick was for him instead of Berry, that might have been a third-round pick. But because he was released, he made only $2 million for one season. Since the Broncos signed two low-value qualifying players and lost three players, their comp will be for their highest-value player lost, which would be Berry. And I don’t think Berry’s contract is large enough to merit a third-round pick, despite his great season. But I could be wrong, and I could be with Woody, too.

Gold indeed earned the Broncos a 3rd round comp pick that year, along with Berry (due to a different mistake Adam made). Unfortunately, that pick was wasted used on who I believe was the worst draft pick in Denver Broncos history: Maurice Clarett.

To end this evaluation on a high note, one aspect of the projection that I did exceptionally well on was getting the order of most of the comp picks within rounds correct.  The only picks I had incorrectly swapped were the San Francisco and Indianapolis picks in the 4th round for Alex Boone and Coby Fleener.  These two picks frequently flip flopped as the 2016 regular season progressed, and at the end, Boone and Fleener were ranked within 5 players of each other leaguewide, a very narrow window well within a reasonable margin of error.  This suggests that my programmed adjustments for snap counts and postseason honors are getting increasingly accurate.