Saturday, February 07, 2009

Best Tigers Base Runners

Last year and the year before, I did a long series on base running. This year, I'm going to use the Baseball Prospectus base running stats. Their data is based on an algorithm developed by former BP writer and current Pittsburgh Pirates statistician Dan Fox. Part of what Dan did was similar to what I did but his algorithm is more complex and the results are a little more accurate. Using the retrosheet database and probability theory, Dan Calculated six base running statistics (all of measured in runs above average):
  • EqAAR (Equivalent Air Advancement Runs) - Contribution of base runners advancing on fly outs
  • EqGAR (Equivalent Ground Advancement Runs) - Contribution of advancement on ground outs.
  • EqSBR (Equivalent Stolen Base Runs) - contribution of stolen bases including runs subtracted for caught stealings and pickoffs.
  • EqHAR (Equivalent Hit Advancement Runs) - contribution of runners taking the extra base on a hit: first to third on a single, second to home on a single, first to home on a double.
  • EqOAR (Equivalent Other Advancement Runs) - contribution of other base running advancements - passed balls, wild pitches and balks (evidence shows that those events are not entirely randomly and are influenced by base runners to an extent).
  • EqBRR (Equivalent Base Running Runs)- the sum of the five above statistics above or total base running contribution.
Each of the measures considers the following:
  • Number of opportunities
  • Each runner is compared to league average for each situation mentioned above
  • Base runners are penalized for making outs on bases attempting to advance
Adjustments are made for the following:
  • Runners on base and numbers of outs before and after the play determine the run value
  • Run value is also influenced by where the ball is hit - left field, center field, right field
You can read Dan's entire algorithm in detail here, here, and here. Table 1 below presents EQBRR for Tigers base runners in 2007 and 2008. Curtis Granderson has been the leading Tigers base runner over the past two years with a total of 12.38 runs above average. Other above average Tigers include Gary Sheffield (4.93), Placido Polanco (3.85) and newly acquired Gerald Laird (3.31). Their worst base runners are Magglio Ordonez (-9.87) and Miguel Cabrera (-6.84).

Table 1 - Base Running Runs Above Average for Tigers in 2007-2008

Player

EqBRR (2007)

EqBRR (2008)

EqBRR (two yrs)

Curtis Granderson

6.42

5.96

12.38

Gary Sheffield

4.99

-0.06

4.93

Placido Polanco

0.92

2.93

3.85

Gerald Laird

2.15

1.16

3.31

Carlos Guillen

-0.17

0.86

0.69

Adam Everett

-1.72

0.76

-0.96

Marcus Thames

-0.10

-1.49

-1.59

Brandon Inge

0.35

-2.32

-1.97

Miguel Cabrera

-4.30

-2.54

-6.84

Magglio Ordonez

-1.89

-7.98

-9.87

4 comments:

  1. Great stuff, Lee. A couple things:

    (1) I wonder why BP don't have these data going back to 1954? It seems all you'd need is PxP info in which all the advancements or lack thereof are reflected. It's not like advanced fielding analysis, for which you need the camera system. Or am I missing something here?

    (2) These figures reflect bottom line EqR generated, which is affected by playing time. How valid would it be review this info on a per-600 PA basis? I know that injury-proneness is always a factor, but a rare injury to somehow not otherwise prone may yield a projection for them that's ... ahem ... off-base. Valid?

    ReplyDelete
  2. Chuck, I'm not sure why BP only goes back two years. You are correct that retrosheet gives them all the info they need to go back to 1954.

    As for your second question, you would have to do it on an opportunity basis rather than a PA basis. This is a little tricky because you have to consider the opportunities for the five elements (air advancement, ground advancement, etc)first before getting the final number.

    ReplyDelete
  3. Opportunity basis -- interesting. OK, I get why, because you have to account for varying levels of performance across the league for each player's team, but how can you project opportunity level? I understand that different teams create different levels of opportunity because of the unique player makeup of the team. My question is, how do you factor the opportunity likelihood of the team into projections? Create a projected team figure of likely players on a roster and layer player projection onto that?

    ReplyDelete
  4. I was thinking of something simpler than that:

    Granderson had 46 opportunities to move up on a ground out and was -.1872 runs below average on that type of advancement. If he had the same opportunity rate over 160 games rather than 141 games, he would be -.1872*(160/141)=-.2124. Then you would do the same thing for all the different advancements and add up the results at the end.

    ReplyDelete

Twitter

Blog Archive

Subscribe

My Sabermetrics Book

My Sabermetrics Book
One of Baseball America's top ten books of 2010

Other Sabermetrics Books

Stat Counter