Tuesday, June 16, 2009

Comparing the MLB and NFL drafts

Inspired by a Twitter suggestion from Will Carroll of Baseball Prospectus, I figured it would be interesting to compare the washout rate of first-round prospects across the four major North American sports. Carroll suggested 2004 as a year, so I started there, but the baseball evaluation was difficult, as several of those prospects are just beginning to crack the league. Thus, I went back a couple years further to the 2002 draft. It's a nice distance, as prospects' status is usually pretty well set after seven years. This MLB draft also has the added appeal of being featured by Michael Lewis in Moneyball. Here's the first part of the comparison, looking at the MLB and NFL first-round drafts; I'll have the second half, looking at the NBA and NHL, up tomorrow.

First, a note on the evaluation system used. I was looking for a simple way to compare prospects across sports, so I went with a five-point scale. Under this system, players ranked as a five are presumed to be among the best in the league at their position, while players ranked as a four are labeled as regular starters, threes are marginal players who were still in the league last season, twos are busts and ones are the worst busts.

In order to figure out who slots in where, I tried to come up with a system that could be relatively comparable across positions and sports. Obviously, single stats like on-base percentage or rushing yards won't work for this. However, all of these leagues do have stats for how many games a player was involved in and if they were selected to play in the league's all-star game. Thus, I was able to create a system that can be applied across sports with only minimal tweaking. The specific criteria for each sport are listed before that sport's table. This clearly isn't going to be a perfect system, and there are other variables such as team quality that aren't taken into consideration here; a starter on a poor team might be worse than a prospect or backup at a better team The seven years of distance does help with this, though, as teams have usually traded players who could start somewhere else by this point. This also doesn't account for injuries, as it made the most sense to base this on last year's stats; however, I have included some notes on certain players who might have fit into a higher category if not for injury or other circumstances. This is intended just as an overall look at the differences between the drafts, though, not an absolute evaluation of any of these players. Moreover, limiting this to one year means it may not be applicable generally; it's more of an attempt to try and get a handle on the differences between the leagues. Ratings and notes are after the tables:



[Table from Wikipedia]

System:
5 = All-star (appeared in at least one all-star game)
4 = Solid (in league, played in 2/3 of games or more last season for everyday players, pitched 40 or more innings for pitchers)
3 = Marginal (played in at least one game in the big leagues last season)
2 = One-off (made it to the big leagues at least once, not in league last year)
1 = Bust (never played in the big leagues as of last year)

Ratings:
5 = Fielder, Saunders, Kazmir, Hamels: four players, 13.3 per cent of all 30 draft picks
4 = Upton, Greinke, Francis, Hermida, Greene, Swisher, Loney, Span, Guthrie, Francoeur, Blanton, Cain: 12 players, 33.3 per cent
3 = Bullington, Loewen, Moore, Ring: four players, 13.3 per cent
2 = Meyer, Adams: two players, 6.67 per cent
1 = Gruler, Everts, Brownlie, McCurdy, Santos, Grigsby, Fritz, Mayberry: eight players, 26.7 per cent

Notes: Grienke and Swisher could potentially crack the all-star ranks; Upton might as well if he returns to previous form. Loewen (who I went to high school with) was decent, but struggled with control; he was badly hurt last year and had to give up pitching, but is now trying to come back as an everyday player in Toronto's system. Bullington, the number-one pick overall (largely thanks to his low contract demands), is now a Blue Jays' reliever. Francoeur started off well, but has been horrible lately and might fall down to marginal status if his slide continues. Mayberry made his MLB debut this season with the Phillies.

NFL:
Draft:


[Table from Wikipedia]

System:
5 = All-star (at least one Pro Bowl selection)
4= Solid (started 2/3 or more of team's regular-season games last year)
3 = Marginal (started at one point in time, played in the league last year)
2 = One-off (started more than five games in their career, didn't play last year)
1 = Bust (started less than five games in their career)

Rankings:
5 = Peppers, Williams, Henderson, Freeney, Shockey, Haynesworth, Walker, Reed, Sheppard: nine players, 28.1 per cent of all 32 picks,
4 = Jammer, McKinnie, Jones, Buchanon, Graham, Thomas, Colombo: seven players, 21.9 per cent of all picks
3 = Carr, Sims, Stallworth, Duckett, Lelie, Harris, Grant, Stevens, Simmons, Thomas, Ramsey: 11 players, 34.4 per cent of all picks
2 = Harrington, Williams, Green: three players, 8.8 per cent of all picks
1 = Bryant, Rumph: two players, 6.3 per cent of all picks

Notes: Jones is not expected to start this year. Williams has been out of the league for a couple of seasons, but is trying to get back in, as is Bryant. Stallworth played in 11 games last year, but only started seven and is facing legal trouble. Grant started all of the Saints' first eight games last year and then was injured; he could easily move up to a four with a healthy season. Sheppard is a two-time Pro Bowler from 2004 and 2006, but only started three games last year. Colombo is also a metal guitarist. Simmons started the Steelers' first four games last season and then got hurt, but was released in the off-season. He has not yet found a new team.

How the leagues stack up:


As you can see from the above table, the NFL first-round draftees from 2002 were much more successful. Only 6.3 per cent of all the NFL first-round draftees examined ranked as ones, compared to 26.7 per cent of all MLB first-round draftees. That's even more impressive when you consider that even the NFL player considered the biggest bust by these metrics (Wendell Bryant) played in 29 career games and started nine; the 26.7 per cent of MLB players ranked as ones never played in the big leagues. Moreover, 28.7 per cent of NFL first-rounders from 2002 earned at least one Pro Bowl nod prior to this year, while only 13.3 per cent of MLB first-rounders from that year made the All-Star game.

That doesn't mean that NFL teams are necessarily inherently better at evaluating talent, though. NFL picks have several other factors working in their favour. For one thing, NFL players are selected out of college, not high school, so GMs already have a good idea of how they perform against tough competition; some MLB picks are taken out of the college ranks, but many are chosen straight from high school.

Moreover, the NFL draft is limited to seven rounds while the MLB draft goes 50 rounds; the sheer numbers of players selected means that first-round picks face much stiffer competition for roster spots. The extensive minor league system in baseball and the longer expected development time means that many top picks will never see the big leagues, while a first-round pick in the NFL is all but guaranteed to at least play in the league (which is why the evaluation of what's considered a class-two and class-one bust in each league is offset in my system). Still, the success/failure rates do make for an interesting read. What will perhaps be more illuminating to look at is how the NHL and NBA compare to the NFL, as the development systems and the numbers of players selected are more similar to those found in football. I'll have a post with the NHL and NBA numbers tomorrow.

2 comments:

  1. Andrew, you're a machine. Fantastic post.

    ReplyDelete
  2. Love the analysis, Andrew! I've always thought that people are too harsh on the MLB draft; it is not the total crapshoot that some would have you believe.

    Good teams with competent front offices always seem to pick some decent players - The Red Sox have been picking in the 25 to 30 range for the past couple years, but still have one of the strongest farm systems, and it's not all because they're willing to go above slot.

    ReplyDelete