Win Shares for MVP

September 14, 2004

I’ve posted an article over at the Hardball Times, called Win Shares for MVP, that makes the best case I can make for the usefulness of Win Shares.

Win Shares is a retrospective, value-based statistic, which makes it the right stat for defining “Valuable.”  There’s also a pretty good discussion of the article over at Baseball Primer.



Tuesday September 22/04
BARRY BONDS
Look at one division and Giants

Team           OBP   OPS   RUNS
Giants-BONDS   .357     .794     796
St. Louis         .345     .807     796
Houston         .343     .777     747
Chicago Cubs     .328     .793     736
Cinncinati Reds   .331     .751     688
Giants-NO Bonds .326   .736   ~660
Pittsburgh         .322     .725     642
Milwaukee         .321   .706     585


These numbers speak for themselves.
Barry Bonds is the MVP.

He means 1 run per game he plays to this Giant team if the Giants can insert a bat from the bench into left field that is about equal to the rest of the team (probably about right) when he is not playing. His fielding and running are probably plus a few runs also. 

On base percentage again is the dominant statistic, greater than slugging percentage in predicting runs.

Why is Bonds so valuable?
1)It is helpful to group or concentrate very productive bats (even a couple of great hitters). Bary’s OPS and particularly his OBP are so outlandish that he can be put together with ..oh, say Felix and Piezynski (take any two) and with them create an OPS grouping equal to Edmonds, Pujols and Rolens of St. Louis.

Bonds always centers the two hitters. Concentration of power.

He gives you the luxury of perhaps putting all this OPS at the front of the order.

He provides an OBP GREATER than that of the St. Louis troika. BONDS, Pierzynski and Felix BEAT Pujols, Rolens and Edmonds in OBP.

BONDS once again needs studying. Many answers to Saberm… lie in the studies of Bonds.

Posted by .(JavaScript must be enabled to view this email address)  on  09/22  at  11:07 PM

MY Favorite stat. Does someone use this?

I call it   ”  BP ”  It is so simple, and I think it is better than OPS

BP = Base Percentage

TB’s + BB + HBP/PA - SF

*Add reached by error for more advanced version and do not count SAC’s or SF’s.

Singles hitters that do not walk a lot will not like this measurement-but then, how good could they be????

In any case, baseball accords them plenty of fame (could Suzuki get a second MVP and now with a last place team?) ..Gwynn, Suzuki, Carew

It’s advantage over OPS is that it does not double count the OBP attributed to hits that OPS does.

What it does is adjust slugging percentage upward for times reaching first base other than by a hit (walk or error). If one’s slugging % is above 1.000, these would cause an adjustment downward. This is probably the way it should be.

You would not really want your opponent to walk Barry Bonds if he slugged 1.000.  It would look like this…

AB   Hits Double Triple   HR   BA   BB
400 160   40     10     60   .400   250

Walks go well with slugging because they give credit for something apart from slugging that is essential to evaluating a batter’s performance, and yet the two can be equated. An OB Percentile is more valuable that a slugging percentile, all things being equal. But since walks do not advance runners, they have a less valuable component also..it adds up to about a wash.

TUES SEPTEMBER 28

        PA- TB’s BB+  BP     OPS
        SF       HBP
BONDS 594 302 231   .897   1.453  
PUJOLS 661 376   89   .703   1.072
BELTRE 631 374   53   .677   1.042
SUZUKI 728 309   51   .495     .871


It can be used as counting stat, or as a percentage.
              Bases     Base Percentage
BONDS       533           .897
PUJOLS       465           .703
BELTRE       427           .677
SUZUKI       360           .495  


Outs made-  Bonds 226   Suzuki 426
Bases made-Bonds 533   Suzuki 360

The measure looks good, does it not?

Posted by .(JavaScript must be enabled to view this email address)  on  09/28  at  02:24 PM

Craig, I think that’s pretty much “Total Average” which was invented by Tom Boswell in the 1980’s, I think.  This article:

http://knology.net/~johnfjarvis/runs_survey.html

compares many of the most popular run creation stats, including TA.  It does do better than OPS, and about as well as OBP times SLG.

Neither one beats Runs Created, or any of the more sophisticated run estimation stats.

Posted by studes  on  09/28  at  02:35 PM

Thanks for the excellent reference to baseball run evaluators. There are better and more sophisticated run evaluators than total average or BP. TA includes stolen bases -it should not do so, insufficient runner advancement contributions.
I STILL LIKE BP (Base percentage) and simply “BASES”.

BETTER             WORSE
BASES             TOTAL BASES         Base Percentage     Slugging Pct.

It accounts for walks (unlike TB’s and Sluf. Pct) and creates a better marriage of OB Pct and Slug Pct than does OPS ... resulting in a more reliable indicator of value.

I first used it as a way to ecompare the great amount of walking Bonds did in 2003 to Pujols greater amount of hitting…

I felt walks went very well with slugging because:
A) They equally value the bases attributed to a batter
B) They have offsetting components and values that come very close to neutralizing each other when it comes to runner advancement.

Bases from base hits advance runners more and are more valuable in short sequence scoring.

Walks are equal if followed by Homeruns later in a inning.
Walks do make runner advancement contibutions.
Walks reduce base advancement deficits to bases from hits by clogging up the bases for future hits…
- a homerun worth 4 bases gives no future advancement potential, 4 walks gives a potential of 12 advancement bases for future hitters
... a key reason why SFrancisco scored so many runs in 2004. Walks never contribute outs, creating extra hitting opportuinities for future hitters…. another key reason why S Francisco scored so much in 2004.


Total Bases + walks=BASES
Bases/PA =BP (Base Percentage)

Better than OPS for many reasons
1.  simpler
2.  OPS is more apples and oranges
3.  OPS gives zero credit to the base advancement from walks

BP’s best feature is that it marries bases from walks and bases from hits ..and they are nearly equal in value.. so one can get a simple base total like this one on Sept 24 ...or a rate.

Sept24/04   BASES   BASE PERCENTAGE
BONDS     533           .897
PUJOLS     465           .703
BELTRE     427           .677
SUZUKI     360             .495

OPS is a combination stat that the public has come to understand very well… kind of like Power/Speed.

Bases is really a single counting stat.  BP is like Batting Average or Total Average or Slugging Pct.

 

Posted by .(JavaScript must be enabled to view this email address)  on  10/05  at  08:07 PM
Page 1 of 1 pages
Commenting is not available in this channel entry.
<<Previous Article:  All-time Leaders in Percent of team Win Shares Next Article:  2004 Win Shares now Available>>