Skip to main content

This is only a sample of what could be done, and not meant to be the only thing done.

 

Been thinking about what was said about “tweaking” the stats in the "Success and Failure" thread to make one player appear to be superior to another, or that allow two people looking at the same numbers to come up with two different interpretations. A simple example of what I often see is in the link.

 

http://www.infosports.com/scor...per/images/rank6.pdf

 

“Run Production” is a stat I’ve long thought was better than either RBIs or Runs by themselves. Its pretty simple in that its nothing more than RBIs+Runs-HRs. Like all stats, while just the raw number does show which player reached the highest number, it really doesn’t tell anyone a great deal until its put into some kind of context. See the 1st 2 pages where the order is simple total runs produced. One context it could be put into is Runs Produced per PA as in the 2nd 2 pages where that’s the order the data appears. Another context is Pitches per Run Produced as can be seen in the last 2 pages.

 

The thing is, all 3 examples use exactly the same data, but since the order the data appears is changed, 3 different people can easily make different interpretations based on what they’re looking for and the context they’re seeing it in. That happens very often when dealing with parents looking for any way to see their child as being somehow superior to his peers.

 

To me, the most telling order would be the 2nd where the order is runs produced per PA because it would tell me the higher on the list the player, the more likely he’d produce a run.

 

But to some, the last order where its number of pitches per run produced would be the most telling.

Original Post

Runs and RBIs are both flawed when used this way because they both rely heavily on forces outside of the player's control.  Additionally, doing this on a per pitch basis distorts things even further, since better hitters tend to see more pitches, but it's easy to get flukily high RBI totals from otherwise average hitters who might see a lot fewer pitches.

 

All of this is re-inventing the wheel, anyway, when things like linear weights, RC, and other stats are available.  Even if you don't necessarily have the data to do such a system complete justice, there's enough work that's been done on the college game to give you a decent ballpark for plugging in HS numbers, or you could just do the necessary regression on the data you do have and get a flawed, but at least better system than inventing new stats.

Add Reply

×
×
×
×
Link copied to your clipboard.
×