Join Date: Apr 2010
The Mocktopus' "Moneyball-Style" Mock Draft
The Mocktopus has returned for the 2013 NFL Draft!
Here is an explanation of what the Mocktopus is. The Mocktopus is a series of models that I created to predict the probability that a team on the clock will select a given player in the NFL Draft. Included in the Mocktopus's data set is every first round pick made since 2005. Although much derided, my research suggests that the most accurate predictors of who a player will draft are other mock drafts. The Mocktopus includes several mock drafts, each which is given a weighted score based on its accuracy at predicting the results of past drafts. As an overlay, the Mocktopus also includes takes into account Top 100 value lists from draft websites. Although mock drafts are the best predictors of which prospect will go to which team, mock drafts typically overrate the role of "team need" plays in draft decision-making. Thus, providing this value "check" on mock drafts differentiates the Mocktopus from a "consensus mock" which tallies up mock drafts like votes and slots players accordingly. To take an example of how this works, last year many mock drafts had Mark Barron going to the Dallas Cowboys at number fourteen overall. The Mocktopus, however, recognized that such a valuable player was unlikely to last until number fourteen, and it's instincts proved correct--Barron was drafted seventh overall by the Tampa Bay Buccaneers.
The Mocktopus treats some position groups differently depending on historical trends. For one, the Mocktopus does not include the "value check" for quarterbacks. In a world where Christian Ponder and Jake Locker are first-round picks, NFL teams tend to abandon value.
This is the mock from the Mocktopus that I call the "Fan Mock." Let me explain what makes the "Fan Mock" unique among mock drafts.
I'm a fan of the Detroit Lions, so when I read a mock draft, I'm naturally most interested in who the author projects for the Detroit Lions. Sure, I might read the dialogues for the first four picks, and maybe skim the rest in a casual fashion, but I honestly care little about who Mel Kiper thinks that the Atlanta Falcons are going to select at number thirty overall.
Now, one of the fun surprises I had when I created the Mocktopus is that it often projects the same player to multiple teams. This is not a rare phenomenon: it occurs multiple times for every single draft that is included within the Mocktopus. For example, for the 2011 NFL Draft, the Mocktopus would have projected Corey Liuget to both the Patriots and the Chargers, Cameron Heyward to both the Jets and the Patriots, and Gabe Carimi to both the Chiefs and the Bears.
This phenomenon occurs because, except in a few instances, there is often only a small probability that the most likely team to draft any given player actually drafts that player. Imagine, for example, that there is a 100% chance that Geno Smith slides to the Cardinals, that the Cardinals have a 50% chance of selecting him, and have a 50% chance of drafting one of five other players. Now imagine that the Bills will have to 50% chance to draft Geno Smith if the Cardinals pass, and have no more than a 15% chance of selecting any other player whether he falls or not. If our goal is to create a mock draft that "gets the most picks right," we would project Geno Smith to both the Cardinals (50% chance) and Bills (25% chance with no other player more than a 15% chance). Indeed, if we think the Jets have a 100% chance to draft Geno Smith if he falls to them, it is likely that we would project Geno Smith to go to three teams.
Because a mock draft that was interested in maximizing its accuracy at projecting each of the 32 selections in the first round would project the same player to multiple teams, it should follow that nearly every mock draft in America should do so. I have read a number of mock drafts, however, and I can safely say that I have never seen a mock draft do this, even though, in theory, every mock draft should do this. Indeed, all mock drafts will invariably have a pick or two that are forced and are not likely selections for the subject teams. It appears that an unwritten rule of mock drafting is that a mock draft should be plausible, i.e. it is something that could happen on draft day. I'm not convinced what, if anything, internal logic adds to the enterprise of mock drafting. Given that the first round consists of thirty two selections, and that there are at least twenty players who could plausibly be selected at any one of those thirty two selections, the chances that any mock draft will later turn out to be 100% accurate is probably several times less likely than winning the lottery (a much better use, if you ask me, of such amazing, odds-defying luck). It seems that mock drafters do a (admittedly mild) disservice to the fanbases of the teams that have been shoe horned into making improbable selections in the interest of chasing a goal that is a virtual statistical impossibility.
Hence, the Mocktopus's "Fan Mock." The Fan Mock is perhaps the first mock draft ever to say to hell with internal logic and to match each and every one of the NFL's 32 teams with what it views as their most likely draftees. The Fan Mock includes many glorious duplicates. It is a mock draft for the people!
This is a draft where using "doubles" could potentially pay off. For example, a trade into the first round, the mocks agree that Geno Smith will likely go to the Browns, Bills, or Jets. The Mocktopus' Fan Mock can hedge its bets by mocking Geno multiple places in the hopes that it can get one of those picks right. Dion Jordan and Dee Milliner continue to be hard players to mock. The Mocktopus thinks that they are both strong second or third choices for a number of teams in the early to mid first round.
One interesting recent trend is the near consensus that Lane Johnson will go to the Eagles, which is remarkable, because almost no one was of that opinion twenty four hours ago. Presently, the Mocktopus's picks made with the most confidence are Sharrif Floyd to the Raiders, Barkevious Mingo to the Jets, and Lane Johnson to the Eagles.
1. Kansas City Chiefs / Eric Fisher, OT, Central Michigan
2. Jacksonville Jaguars / Luke Joeckel, OT, Texas A&M
3. Oakland Raiders / Sharif Floyd, DT, Florida
4. Philadelphia Eagles / Lane Johnson, OT, Oklahoma
5. Detroit Lions / Ezekiel Ansah, DE, BYU
6. Cleveland Browns / Geno Smith, QB, West Virginai
7. Arizona Cardinals / Jonathan Cooper, OG, North Carolina
8. Buffalo Bills / Matt Barkley, QB, USC
9. New York Jets / Barkevious Mingo, OLB, LSU
10. Tennessee Titans / Dee Milliner, CB, Alabama
11. San Diego Chargers / Chance Warmack, OG, Alabama
12. Miami Dolphins / Kenny Vaccaro, SAF, Texas
13. New York Jets / Tyler Eifert, TE, Notre Dame
14. Carolina Panthers / Sheldon Richardson, DT, Missouri
15. New Orleans Saints / Jarvis Jones, OLB, Georgia
16. St. Louis Rams / Cordarrelle Patterson, WR, Tennessee
17. Pittsburgh Steelers / Jarvis Jones, OLB, Georgia
18. Dallas Cowboys / Sylvester Williams, DT, Missouri
19. New York Giants / Tank Carradine, DE, Florida State
20. Chicago Bears / Manti Te'o, ILB, Notre Dame
21. Cincinnati Bengals / Eric Reid, SAF, LSU
22. St. Louis Rams / Kenny Vacarro, SAF, Florida
23. Minnesota Vikings / Cordarrelle Patterson, WR, Tennessee
24. Indianapolis Colts / Xavier Rhodes, CB, Florida State
25. Minnesota Vikings / Manti Te'o, LB, Notre Dame
26. Green Bay Packers / Eddie Lacy, RB, Alabama
27. Houston Texans / Justin Hunter, WR, Tennessee
28. Denver Broncos / Bjoern Werner, DE, Florida State
29. New England Patriots / D.J. Hayden, CB, Houston
30. Atlanta Falcons / Desmond Trufant, CB, Washington
31. San Francisco 49ers / Kenny Vacarro, SAF, Stanford
32. Baltimore Ravens / Arthur Brown, OLB, Kansas State