I have never invented a baseball statistic before. I have come up with several fantasy baseball strategies, several of which I've talked about here, but never a baseball statistic. After not much time and several Cheez-Its later, I've come up with one...

I'm not going to pretend there is suspense and write the name of it 20 lines down so you have to scroll down in your browser to see it because I already ruined the suspense in the article title. The name of my "new" stat is run percentage, or R%

The key component of it is basic runs and although runs is not a great stat for valuating player performance, it does have its place in the world of fantasy baseball. There have been plenty of stats that try to separate the luck and skill for other stats, such as BABIP for batting average, but never anything for runs before.

Perhaps there is a significant amount of luck involved that is identifiable to an extent... that would certainly help you fantasy-wise. A guy who gets 100 runs is more valuable than a 85 run-guy, right? Fluctuations in runs totals from year-to-year commonly are 10-20 runs. What contributes to this?

The two factors that come to mind first are player performance (bating average, home runs, etc.) and the lineup surrounding the player. There are probably others besides those two, but we will address all of that stuff later.

Now I want to discuss the methodology behind my stat, R%. It is a percentage, meaning the formula will involve a fraction. Runs will be on top and will divided by something. I don't want at-bats or plate appearances or anything like that because it should be R divided by chances to score. That means the batter must have gotten on base somehow. So hits then. R/Hits.

Well hits works, but there are obviously other ways to get on base to put yourself in a scoring opportunity. Walks is the other big one. The next thing I'm thinking is that runs resulting from home runs should be left out, because the batter never really had the opportunity to score. Remember we're trying to determine how many times a batter scored out of the opportunities he had to score.

So as of now, the formula is R - HR/H + BB

That looks pretty good, but I think more tweaking is still necessary. The formula above is strictly Runs/Opportunity. What that assumes, is that all scoring opportunities are equal, which they are not. A runner on third has a much better chance of scoring than one on first. I'm not sure exactly how much larger that chance is--for the runner on third to score--so I'm gonna keep things simple and weight it 1, 2, 3.

That makes our formula R - HR/(singles) + (doubs * 2) + (trips * 3). Plug in the numbers, and you should get the percentage (after multiplying by 100) of Runs scored over opportunity to score those runs. The Run Percentage.

Let's test it out on two players for now, and I'll pick Jack Cust and Jose Reyes. Kind of opposite
type players, let's see what their R%'s are for 2007 and 2008:

I did the 2007 computations first and when I saw Reyes with the edge over Cust I thought, "Wow, that it showed the speedy player scoring a higher percentage of the time is pretty cool." Then about 30 seconds later I did the 2008 numbers and Reyes and Cust basically traded numbers. Neither player really sustained their percentage from 2007 to 2008, but even if they did it wouldn't show anything conclusive. We would obviously need to look at more players over more years before conclusions can be drawn.

Now I want to jump back to a statement I said earlier. The two big factors on a player's run total--performance and surrounding lineup. I feel there are two ways I can introduce these factors into my statistic to make it more useful.

• The first would be to find out how much of a impact team Runs per Game affect the R%. On average, exactly how much higher of a R% would a Rangers player have this year than a Mariners player?
• Secondly, I could account for Team R/G and position in batting order and then create a "context-neutral" version of R%. This would obviously take some time and math, we'll see if I ever get to that.

Okay, that about does it for this week. I'm not going to release any numbers just yet, this was more of a methodology article than anything else. If you feel I'm missing a component, let me know! Leave it in the comments or e-mail me.

Just a heads up, Brad Stewart is planning a really great 2008 season wrap-up. Perhaps you might find these R% numbers in there. Wink. Wink.