clock menu more-arrow no yes

Filed under:

Net Runs Above Average

New, 8 comments

I want to do a little more work with Net Runs, a stat I've been working on as of late. I've set it up so that I have per 100 game values (since Rate2 is per 100 games played, that is where I needed to start) but you can also divide that figure by 100 to get a decimal of Net R/G. That number can then be multiplied by whatever number you want to get Net Runs per n games. Here is what I just said as a formula:

1. Net Runs Per 100 Games
(MLVr*100)+ (Rate2 - 100) = Net Runs/100 G

2. Net Runs Per Game
Net Runs/100 G / 100 = Net R/G

3. Net Runs/n Games
Net R/G*n = Net R/n G

Pretty simple stuff obviously, but a statistic I'm starting to like. Part of that is probably because I'm not the math whiz that anyone else on this site is as far as creating my own statistics goes. Net Runs should be considered in the context that it is Net Runs Above Average (NRAA), since 0.000 is the MLVr of an average major league hitter, and Rate2 works off of the average major league defender.

There are two forms of this statistic that I want to use. NRAA will focus on the current number of Net Runs a player has accumulated and their current rates. NRAA2 will be used more as a predictor of future success. NRAA will use Rate2 without any adjustments. NRAA2 will use Richard Wade's Established Level of Performance formula for Rate2 in order to figure a player's defensive abilities. This will allow us to make concessions when dealing with a player who was slowed by injury or is just having an unlucky year defensively to that point in order to better predict their future performance. The formula for NRAA2 is as follows:

1. Net Runs Per 100 Games
(MLVr*100)+ (EPL Rate2 - 100) = Net Runs/100 G

2. Net Runs Per Game
Net Runs/100 G / 100 = Net R/G

3. Net Runs/n Games
Net R/G*n = Net R/n G

EPL is Established Performance Level. Richard Wade's formula for this is as follows, for those playing our at-home version of the game:

(3 * AdjGx * Rate2x + 2 * AdjGy * Rate2y + AdjGz * Rate2z) / (3 * AdjGx + 2 * AdjGy + AdjGz)

Where x = most recent year, y = second most recent year, and z = third most recent year

Let's take a look at the current Net Runs leaders with a few different graphs, because I want something other than makeshift tables today:


Interesting that names like Brandon Inge, Lyle Overbay, Morgan Ensberg and Luis Castillo made their way into the top 20 players according to NRAA. The players in that table are ranked according to NRAA/G. Notice how far ahead Derrek Lee is from everyone else. Let's look at that in a graph to really get a grasp on it visually:


Derrek Lee is averaging over 80% of a run above average per game, while no other player in the majors tops half a run at all.  At some point I'll compare Lee's season to some historical seasons to see where he stands using NRAA.

I think NRAA is a pretty useful statistic for assessing value of a player offensively and defensively. Just by reading some unexpected names on that top 20 leaderboard makes me feel that the stat has a purpose. At some point BtB should put up some leaderboards for statistics special to this site. The only real issue I have is that MLVr numbers only go as far back as 1972 on Baseball Prospectus's web site, but I'll make do with what I have just like they have, and maybe they will add more years at some point.

I want some comments and criticism for this statistic, and ways to improve it, etc. Constructive criticism, mind you. Not that we don't always play nice around here.