This is bordering on the ridiculous, King. The NBA finalized its teams for the championship and no mention? No mention of Pat Riley going to the championships again? No mention of Avery Johnson, coach of the year, getting his team there? All of this on the heals of no mention about the stunning performance 50 point performance by Dirk for the Mavs?

I think you tapped upon one idea of why Greg Maddux is such a great pitcher apart from pitching: a decent hitter and a constant gold glover winner.

Maybe you could combine the PRV with errors / wild pitchers per 9 innings? I think the idea you have is to get the total value of a pitcher apart from a very one-dimensional pitching-only stat like WHIP or ERA.

-Alec

You know damn well that by the time they are over King will have addressed the NBA finals in somewhere something like ten columns if they go six or seven games. Give the man a day to talk about Pervs before he's locked into the NBA finals for two weeks.

King,

I don't understand why you want to use the 3.5 at bats per nine innings factor. Why not just use the pitchers actual number of at bats along with his actual number of innings pitched to scale things?

Of course there are problems with this because of fractions of innings, and pitchers being more likely to get pulled if they are coming up to bat soon. Even with these problems, it seems like it would be more accurate to use actual numbers than a fudge factor.

It does make the calculation a little more complicated, but really it only involves pulling a couple more columns of data into your calculation.

First, I like this idea quite a bit. The other day, Ian Snell came up to bat for the Pirates - young guy, started the season poorly, but has come around some the last month. Anyway, the radio guy mentions that he's never had a big league hit: 0-for-22 (the exact same start, BTW, as rookie Aramis Ramirez). He ended up getting a 2-out RBI single, but I couldn't help but think, Boy he sure hasn't helped himself out there.

Anyway, I just wanted to suggest that the fielding shouldn't be folded in - it's implicit in the ERA, isn't it? If he's a lousy fielder, then that presumably translates into some additional runs. What might be useful is an adjusted ERA that treats runs scored due to pitcher errors as earned. I actually used to think that's how ERA worked, but apparently it doesn't. Same deal with wild pitches - runs that score because of them belong on the ERA.

On a separate note, It seems that it would be useful (to GMs, at least) to have a stat, similar to Inherited Runners Scored for relievers, for unearned runs allowed that the pitcher could have stopped. In other words, bases-empty, 2 out, the shortstop boots one. The pitcher proceeds to give up 3 runs, all unearned. That is, to a great extent, on him. Sure, there's Total Runs Allowed, but the above situation is different from a pitcher who induces a two-out, bases-loaded popup that's dropped and lets all 3 runners score.

I might add that this would be, in some sense, a measure of "clutch" for pitchers.

Instead of runs created, you could just use VORP (also available at Baseball Prospectus, even for pitchers), and subtract it from runs allowed, dividing by innings to get a rate: (RA-VORP)/IP*9.0

I forget whose system uses it, but you could even add in Fielding runs for the total packageL (RA-VORP-FR)/IP*9.0

To be honest, you could just swap out VORP for runs created, which was already in the proper format. Dividing by 3.5 and then multiplying by essentially the same factor was kind of unnecessary.

It can be summed up in three words: small sample size. For example, Greg Maddux, who has spent his entire (lengthy) career in the National League, has 1461 career plate appearances. Your everyday position player probably has something like 650 in one season. Would you judge a position player's ability based on just over two years' batting stats? If you would, I've got a ton of can't-miss prospects to trade you for a superstar.

In the same vein, I've always thought that the most important offensive stat would be total run production: Runs scored + RBIs - Home Runs (so they're not counted twice). This represents a composite of hitting for average, hitting for power, good situational hitting, and good baserunning. It seems too obvious to me to be ignored, yet I've never seen it published this way. Am I missing something?

Why not use career batting statistics when calculating the pitcher's RC/9 portion of the PRV stat?

On the ESPN front page, hover over the MLB pull down menu and click on Teams. Then click on Stats for whichever team you want, and the pitcher's batting numbers are included.

http://sports.espn.go.com/mlb/teams

http://sports.espn.go.com/mlb/Statistics/Team/playerstats?team=chc

A stat that would get used for pitcher's batting would be percentage of successful sacrifice bunt attempts. One may dispute whether the sac bunt is good strategy, but it is the mission the pitcher is given. When he pops up, strikes out, or does not advance the runner, he fails and disrupts the strategy of the manager, and may even cost himself run suppport. Perhaps teams do track this since it would be useful to match-up obsessed managers.

I like the idea, and it should possibly reflect the greatest pitcher/hitter ever--the Babe. Though Babe-Perv would be more unseemly.

As to why fielding should be included in the statistic: the ERA of a lousy-fielding pitcher may not reflect runs given up by errors he committed. I always thought this was a stupid statkeeping loophole.

As to Runs Scored and RBIs, they are in good part contingent on other players' actions, whereas at least HR is all hitter. Two leadoff men with great averages could score a much diferent amount based on who hits behind them. Same for RBI guys. Andrew Jones hit like .250 with runners in scoring position, while Pujols hit .330 last year. Jones had more RBIs, simply because he had more opportunities and a few more HR.