Introducing JustVORP
One of the most popular sabermetric statistics ever invented is VORP, Baseball Prospectus' measure of offensive value over replacement player. Not only does it measure offense, it compares each player to his positional peers, because a catcher with an .800 OPS is much more valuable than a first baseman with an .800 OPS. We've all gotten a lot of mileage from VORP over the years, so kudos to Keith Woolner and BPro for that.
But VORP also has its flaws:
- It's based on Equivalent Average MLV, which underrates walks.
- It doesn't account for league quality.
- It defines replacement-level in a funky way. 80% of league-average RC/out isn't a bad choice (although it's slightly high), but that number changes to 85% for catchers and 75% for 1B and DH. Why? Uh... good question. VORP ends up overrating 1B/DH's and underrating catchers.
- It uses offensive averages to define the defensive spectrum instead of actual differences in defensive value. Why is that a mistake? Because it makes DH look like a tougher position to play than 1B and LF/RF, since DHs hit worse as a group. And also because you'll get seasons like 1961 when CFs outhit LFs. Do we really think CF is an easier position to play when that happens? No.
Fortunately, mere mortals can come along and stand on the shoulders of the giants before them. Justin is one of those mortals, and for the past year he has calculated the statistics necessary to produce a better version of VORP. I'm going to call it JustVORP. Conveniently, it solves all the problems listed above:
- It's based on park-adjusted linear weights, derived from BaseRuns.
- It accounts for the talent difference between the AL and NL.
- It defines replacement level as 73% and 78% of league-average offense (depending on league). That number ignores position, because...
- It uses the proper defensive spectrum, based on relative difficulties of playing each position. And the position adjustment is added in after the offensive component is calculated. Offense is offense, but it has more value depending on the position it comes from.
As a reminder, here are the position adjustments over a full season:
CA: +12.5 runs
SS: +7.5 runs
2B/3B/CF: +2.5 runs
LF/RF: -7.5 runs
1B: -12.5 runs
DH: -17.5 runs
Now, like VORP, JustVORP is obviously missing a significant piece of the puzzle, namely fielding relative to position. But that's half the beauty -- with the wide variety of fielding metrics available, you can pick the one you're most comfortable with, like UZR. Or average some together. Or use the Fans Scouting Report. Or use 2009 projections.
Click here to download full 2008 JustVORP data, complete with a pivot table ready to help you slice and dice. Here's the top twenty-five:
| Player | JustVORP |
| Albert Pujols | 78 |
| Hanley Ramirez | 76 |
| Chipper Jones | 68 |
| Manny Ramirez | 68 |
| Grady Sizemore | 66 |
| David A Wright | 65 |
| Alex Rodriguez | 62 |
| Jose Reyes | 59 |
| Lance Berkman | 57 |
| Chase Utley | 56 |
| Carlos Beltran | 55 |
| Joe Mauer | 54 |
| Mark Teixeira | 53 |
| Dustin L Pedroia | 53 |
| Brian Roberts | 52 |
| Nick Markakis | 52 |
| Ryan Ludwick | 52 |
| Josh H Hamilton | 52 |
| Brian M McCann | 51 |
| Curtis Granderson | 49 |
| Nate McLouth | 48 |
| Ian M Kinsler | 47 |
| Carlos J Quentin | 46 |
| Matt T Holliday | 46 |
| Kevin E Youkilis | 46 |
1 recs |
9 comments
|
Comments
Simply amazing.
The work by Beyond the Box Score and related people in the offseason has been outstanding.
Webmaster of Driveline Mechanics
http://www.drivelinemechanics.com - An Unconventional Look at Scouting
by Kyle Boddy on Dec 8, 2008 8:43 AM EST reply actions 0 recs
Very Interesting
I’m going to look into to it further. I see that the analysis is limited to hitters. Any plans to extend it to pitchers?
by prhood on Dec 8, 2008 9:25 AM EST reply actions 0 recs
You need a different methodology for pitchers, because of the whole DIPS thing.
I prefer RAR for pitchers based on tRA and xIP, like I presented in a recent article. Graham’s given the go-ahead to distribute his statcorner data freely, so I’ll have a spreadsheet available in the next few days.
Beyond the Boxscore // Calling BJ Upton lazy is lazy.
by Sky Kalkman on Dec 8, 2008 10:14 AM EST up reply actions 0 recs
Interesting stuff
Just one thing: I thought VORP was based on Marginal Lineup Value, not EqA, as explained in Tango’s deconstruction of VORP anyway.
Adopted Giant: Aaron King
Wearing the crown by 2011. Or at least the LOOGY hat
by baetown415 on Dec 8, 2008 10:20 AM EST reply actions 0 recs
You are correct.
RARP is based on Equivelent Average, and doesn’t have a (big) problem with walks. RARP really is not a bad statistic in and of itself.
by cwyers on Dec 8, 2008 10:39 AM EST up reply actions 0 recs
Well my bad. Should have double-checked things.
Give me a few to fix it.
Beyond the Boxscore // Calling BJ Upton lazy is lazy.
by Sky Kalkman on Dec 8, 2008 10:43 AM EST up reply actions 0 recs
production?
PORP?
Beyond the Boxscore // Calling BJ Upton lazy is lazy.
by Sky Kalkman on Dec 8, 2008 2:47 PM EST up reply actions 0 recs

by 












BtB on Facebook














