So Long 2017 Season
It was a year of ups and downs but the 2017 NFL regular season is in the books. But for DraftTwitter, the Draft Season goes all year long. That includes review of previous year's player performance results typically for talent evaluator/Draftnik bragging rights.
The First DraftTwitter Top 100
Last year, a ragtag bunch of scouts of all ranks separately cast ballots for their 100 best prospects and those results were aggregated into the DraftTwitter Top 100 list
(seen here). During the regular season, this space tracked the progress of those players through the first 10 games of the season simply to rank each DraftTwitter scout, all in the name of fun. Now that the regular season is complete, the focus of measuring efficacy of the aggregate list shifts to the Shield, itself. Could the unthinkable have occurred? Could the DraftTwitter scouts do a better job of picking players than the NFL?!
Football Thunderdome
The top 100 players ranked by the 2017 DraftTwitter Project were lined up against the top 100 players selected in the actual 2017 NFL Player Draft to see which results got more bang for their buck.
Step I - Position Groups
The process of comparing picks can be quite subjective. How do we judge 19th picks, Marlon Humphrey and OJ Howard against each other? Both clearly made positive impressions but one generates yards and the other prevents them. Their evaluation metrics are nearly antithetical, so to go through 100 pairs of dissimilar players with such an unclear process seems a waste of time.
This analysis simply breaks up the draft into Position Groups (that is, centers, free safeties, guards, etc) and evaluates the baskets; but how are they evaluated?
Step II - Snaps, not Stats
As seen above, it would be surprising if anyone could get concurrence from a large group that using stats to compare a CB vs a TE would be a clear cut way to measure one against the other. A previous article comparing the results of the 2017 DraftTwitter draft on a per Scout basis used games started as a barometer for success in the league. The rationale there was the coach will start the players who give him the best chance to help the team win the game.
In comparing the players in the way described in Step 1, above, this quickly proved to be ineffective. It was found that a player's % of Total Snaps (and snap count trends) give a better picture of how much that player's coach relies on him to contribute to his team's success.
So, in order to remove the subjectivity from the debate, the judging was predicated not on yards or tackles or touchdowns (which stats don't apply to many players or can be misleading in many cases) but on the percentage of Total Snaps as well as Average Draft Position (ADP) of the players in that peer group.
Step III - The Cobra Kai Rule: No Mercy
One could argue that exclusion of players who were injured and placed on injured reserve or simply did not make a team was fair so as to not tank averages with a bunch of zeros but...where's the fun in that?!? No mercy, you made the picks now deal with it. So players like Forest Lamp, who started the season on injured reserve after being a top 40 pick, was a double edge sword in tanking both metrics.
The Showdown: DraftTwitter vs. the NFL Draft
In the table above, the average snap counts per the number of players is listed, for reference. This number is secondary to the % of Total Snaps on Off or Def for the average player in the group with ADP representing the Average Draft Position for each group.
By way of example, there were only two centers selected in the DraftTwitter Top 100, Ethan Pocic and Pat Elflein. They were also the only two selected in the actual draft; so, the average snaps and Snap% are identical (in other groups where the DT100 and NFL draft differed in terms of players selected, we will see differences in these). Although the snap data is the same in this case, the ADP is different because the DT100 had the players drafted higher at about 60.5 compared to their actual ADP which was 64.0. In this case, you can see the NFL got a better deal at the Center position.
Step IV- The Final Countdown
The differences between the two sets of columns are listed below, in terms of DT100 (so positive numbers are in favor of DT100, negative numbers show when the League had the advantage).
By way of example, looking at Defensive Tackles, DT100's picks' average percentage of snaps was 24% less than those DTs drafted into the NFL in 2017. That damage is somewhat offset by the much later ADP so that value cuts into the final advantage to the league which shows a Snap%-ADP advantage of 7.82% (or, read it as a disadvantage to DT100).
Looking at the WR group, on average, they reflected a Snap % of -2.94% (meaning the NFL selections had higher Snap%), however, because the DT100 ADP returned enough value by being 6.72% higher AD this more than justified the lower Snap%, actually giving DT100 the advantage of 3.78% with the WR group.
Here are a couple of observations:
- The Good - DT100's best group was the Defensive Ends and that stemmed primarily from keeping away from NFL top 100 picks Dawuane Smoot (Snap% of 24.4), Daeshon Hall (Snap% of 0.9) and Tanoh Kpassagnon (With a Blutarksy Snap% of zero point zero).
- The Bad - The aforementioned Defensive Tackle selections by DraftTwitter were not pretty. Non-NFL top 100 players Caleb Brantley and Jaleel Johnson didn't pull their weight to say the least.
- The Ugly - Of the 14 player positions drafted in the top 100, DT100 won the advantage on 6 positions compared to 8 positions owned by the NFL draft. So NFL wins, right?
WRONG! On an aggregate basis, DT100 showed a 1.52% ADVANTAGE over the NFL draft in terms of start percentage per player and Average Draft Position. Despite the victory by DraftTwitter, the road is long and winding. The intention of this project is to track DraftTwitter scouts picks against the NFL and each other over the years so look for a 2 year update on the 2017 draft same time next year. In the meantime...
Shameless Self Promotion
IT'S TIME FOR THE 2ND ANNUAL DRAFTTWITTER TOP 100 SURVEY!!!
Last year, 25 amateur, semi pro and professional scouts clicked through a simple Excel based ballot and created a Top 100 list of players that has successfully (thus far) proven to a better value than the players picked by our favorite GMs last year.
Join the fun! To receive a ballot, send your name, email and social media handled for football discussion to boombearfootballmail@gmail.com No pressure, just have fun.
