Super Stat Sticky: Get Your Learn On!

[Mashed Potatoes]

Thanks to the work of Baltimoron and now, we have the beginnings of a comprehensive advanced statistics list. Anyone with more links, important and concise notes, or organizational ideas just post them here or PM me, and I'll edit the original post. It will become a sticky, and eventually incorporated into the hangout.

Availability is only the first step though, once the list is finalized I'll start a weekly series on a specific statistic. How is it calculated, where is it applied, what are its strengths, deficiencies, and limits...etc. Soon we will all be "cooking with gas."

Quick Links to Super Stats - the best and most comprehensive measures

EQA (Equivalent Average)

VORP Hitters

VORP Pitchers (Value Over Replacement Player)

Winshares

Runs Created

WPA Leaderboard

wOBA (Weighted On Base Average)

2004-2006 Park Factors

OPS (On base Plus Slugging percentage)

RC27 (Runs Created per 27 outs)

DIPS (Defense-Independent Pitching ERA)

PMR (David Pinto's Probabilistic Model of Range)

WXRL Basically VORP for relievers.


------------------------------------------------------------------------

Notes:
EqA - EQA and the EQA based counting stats EQR, RAP, RAR, RARP. Comprehensive (includes stolen bases), highly correlated to actual runs produced (EqR is the companion counting stat), park and league adjusted, consistent in that it is relative to a league-average baseline that is set to the same number (.260), calibrated to mimic the scale of batting average (i.e. a .300 EQA is comporable in scarcity to that of a .300 batting average).

VORP - Value Over Replacement Player. The number of runs contributed beyond what a replacement-level player at the same position would contribute if given the same percentage of team plate appearances. VORP scores do not consider the quality of a player's defense.

OPS - simple to calculate, adding OBP to SLG. Correlates well to Runs Created (RC/27), but is inaccurate to the extent that OBP is actually at least 1.5 times more important than SLG.

RC27 - Runs created per 27 outs (Estimates how many runs per game a team made up of nine of the same player would score)

DIPS - A pitcher's ERA, independent of the defense behind him. This formula, based on essays by Voros McCracken, assumes that all pitchers have consistent BIPA (Balls In Play Average. Batting Average Against, not including home runs or strikeouts), and adjusts accordingly. The DIPS ratios on ESPN use the DIPS 2.0 formula, are not park-adjusted, and do not adjust for knuckleball pitchers.

WPA - Win Probability Added, is a system in which each player is given credit toward helping his team win, based on play-by-play data and the impact each specific play has on the team's probability of winning.
Here's a link from January 06 at the USS Mariner that talks about the defensive fielding stats available. link It doesn't have great info on the Fielding Bible (because it hadn't yet come out), but is not bad otherwise. Here's a good link discussing the Fielding Bible

wOBA - wOBA

Tangotiger's website is a great resource. Not really easy to digest, but very valuable.

PMR - Its system for evaluating fielding range. It calculates the probability of a ball being turned into an out based on six parameters:

1. Direction of hit (a vector).
2. The type of hit (Fly, ground, line drive, bunt).
3. How hard the ball was hit (slow, medium, hard).
4. The park.
5. The handedness of the pitcher.
6. The handedness of the batter.

For each ball in play, the program sums the probability of that ball being turned into an out, and that gives us the expected outs. Dividing that by balls in play yields expected defensive efficiency rating (DER). That is compared to a player (or team's actual DER), or how many balls in their "area" were turned into outs. Its the next step beyond UZR and unlike the fielding bible, is freely available on the web. PMR is explained here and here, or in a short video 5 minute video here (the video is pretty easy to understand).

WXRL - It takes into account the game situation when a reliever enters, when he leaves and gives credit for the difference by evaluating the relative win probabilities. Inherited runners, bequeathed runners, and leverage are integral.

[Art Wing]

Thanks to the work of Baltimoron and now, we have the beginnings of a comprehensive advanced statistics list. Anyone with more links, important and concise notes, or organizational ideas just post them here or PM me, and I'll edit the original post. It will become a sticky, and eventually incorporated into the hangout.

I love this idea and all of your research is greatly appreciated. Despite being an academic stats and math are sometimes difficult for me to digest- time to "get my learn on".

[PeteCanes]

Yeah, this is great stuff. Can we get a sticky on this action?

Nevermind, just saw it in the hangout feedback section.

[rolliefingers]

Yeah, this is great stuff. Can we get a sticky on this action?

Seconded. This needs to be a sticky on the Orioles forum, just because more people use that.

[Migrant Redbird]

Great idea! Need to fix your link on Winshares. The current link points to Runs Created.

[Nuclear Dish]

Am I wrong in thinking that Drungo has had a list like this up on the Hangout for well over a year?

[Mashed Potatoes]

Am I wrong in thinking that Drungo has had a list like this up on the Hangout for well over a year?

Check the OPS Misleading? thread, he was one of the posters who advocated the idea.

[1970]

I added a link for wOBA.

wOBA is a pretty good stat. The Book tells us it should match OBP, and sure enough, the overall MLB wOBA does match the overall MLB OBP for each of the last three seasons.

I have put together an Excel file for tracking wOBA. I have every player from the last three years in the database - and am planning to go back to 2002. At the excellent suggestion of Baltimoron, I'm trying to figure out a good way to post this data on my blog.

The advantage to wOBA over OPS is that it provides a more realistic weighting to OBP than does OPS. Just as EQA matches BA, wOBA matches OBP. A .300 OBP would not be good, well neither is a .300 wOBA. A .400 OBP would be very good, as is a .400 wOBA.

Another way to look at wOBA is to compare it to OBP. If they are close to being equal then the player probably has a good balance between OBP and SLG. If the wOBA is way above OBP, then the player either has a very high SLG, or his SLG is probably a lot higher than his OBP. Conversely, if the wOBA is way below OBP, then the player either has a very low SLG, or his SLG is probably a lot lower than his OBP.

For instance, the qualifying batter with the biggest difference between wOBA and OBP in 2006 was Jermaine Dye. Yes, he had a very good OBP, but he also had an outstanding SLG of .622. Dye was one of only 7 qualifying batters with a SLG over .600 - and he was the only one of those 7 with an OBP of less than .400.

The qualifying batter with the biggest negative difference between wOBA and OBP was Joe Mauer - his wOBA was .035 lower than his OBP. He had a decent .507 SLG, but he was the only one of 7 batters with an OBP .425 or higher who had a SLG of less than .520. In fact, 4 of those 7 had a SLG of at least .619. Another good example was Jason Kendall - he had a wOBA of .332, but an SLG of only .342 compared to an OBP of .367.

The overall MLB wOBA by season:

2004 - .336
2005 - .330
2006 - .337

[1970]

wOBA by Team can now be found on my blog.

The file provides the overall wOBA of every team and both leagues from 2002 through 2006, including the overall totals for the five year period. It also lists the OBP, SLG, and OPS for each team for comparison.

I'm hoping to have the totals for every player from the same period listed (including a listing of the overall totals of every player with at least 1000 PA since 2002) later today.

LINK

If, for whatever reason, the link doesn't work, please let me know.

[now]

More thanks here for the ongoing work.

I think the opening list of links should all go to updated player rankings, wherever possible, and that links to definitions and explanations should go below in the "Notes" section. So for example, while the new link to the wOBA material is useful, it belongs in the lower section rather than in the upper section. And there maybe 1970's updated wOBA stats can be linked when it's ready.

[1970]

wOBA for every player since 2002 is now posted.

HERE

[Mashed Potatoes]

More thanks here for the ongoing work.

I think the opening list of links should all go to updated player rankings, wherever possible, and that links to definitions and explanations should go below in the "Notes" section. So for example, while the new link to the wOBA material is useful, it belongs in the lower section rather than in the upper section. And there maybe 1970's updated wOBA stats can be linked when it's ready.

Good looking out, I meant to keep them separated as you suggested, but I never clicked the first link 1970 provided so I wasn't sure if it was the explanation or the rankings. Anywho I fixed it.

Also 1970 or Tony or Scottie can you guys change the thread title from Stick to Sticky? Thanks!

[DrungoHazewood]

Am I wrong in thinking that Drungo has had a list like this up on the Hangout for well over a year?

I had a series of articles that were really stat glossaries, and I'd link to them from time to time when questions came up. When the board migrated to its current form they kind of got lost in the archives. I think this will serve as a good replacement, there's already more here than was in the old articles.

[Baltimoron]

Chone Smith's Zone rating adjustments

Basically, a large outfield wall (to a lesser extent a large outfield) can effect a players zone rating because balls that hit these walls or that fall into the extra outfield are "in-park" hits but are generally also balls that are not field able- think fly balls off the green monster or shots to deep power alleys in big parks like Coors.

RF in Camdem is pretty extreme, -0.036, while CF is somewhat less extreme, -.015. This is largely due to the the deep right center wall, which affects both CF and RF stats, and the scoreboard wall in Rf.

What this means is Camden makes the Zone Rating for a CF and a RF worse than it should be, by factors of -.015 for CF and -.036 for RF.

Worst ballpark outfield positions for ZR

Ballpark      POS  ZR adjustment
Enron         LF       -.045
Fenway        LF       -.042
Joe Robbie    LF       -.039
PNC           LF       -.039
PNC           RF       -.039
Camden        RF       -.036
Metrodome     RF       -.026
Metrodome     CF       -.024
PacBell       RF       -.022
Ameriquest    LF       -.021

Best ball park outfield positions for ZR

Ballpark      POS  ZR adjustment
Fenway        RF       .026
Skydome       LF       .022
Dodger        LF       .018
PacBell       LF       .018
BOB           LF       .017
Comerica      LF       .017
Turner        RF       .017
Angel         LF       .016

BTW here is a thread that has a link to all of the 3 year park data (2004-2006) from the 2007 Bill James handbook. link

[Baltimoron]

Another park factor reference developed by USSPatriot by that uses, when available, 5 years of data and some regression to the mean. link