PRAA & PRAR

[bigh0rt]

Ok, so I was having this debate on another forum, where a member was trying to convince me that this year Jonathan Papelbon was more valuable than Travis Hafner. His main argument rested on Papelbon's PRAA and PRAR (41 and 74 respectively).

Now, I did a little research, just to see where Papelbon stood, and was shocked to see that Johan Santana's PRAA on the season was a 38, 3 points lower than Papelbon. His PRAR was 103, however.

But anyway, back to the PRAA. I myself have used this statistic when evaluating worth and I see it thrown around quite a bit, but to see Santana's 2006 value appear lower than any closer's in troubling to me.

A question regarding PRAA:
---
I believe it is based on an XIP (expected innings pitched) calculation. Which, I believe, is an adjustment to the innings pitched of that player based on the importance of the innings which they pitched. Papelbon's adjustment moved him from 68.3 IP to 107.9. So, the 41 PRAR is based on the 107.9.

So, if he allowed 7 ER in 68.3, it is adusted to 11 in 107.9. Add 41 to the 11 and you have 52 ER in 107.9 innings being average...or a 4.33 E.R.A.

However, since Papelbon didn't pitch 107.9 innings and only pitched 64% of that total, wouldn't his true runs above average only be 26? If he pitched the full 107.9 then you get 41, right? But he didn't pitch that many. He pitched 68.3... so, to me, it'd seem like he should only get credit for what he actually pitched. That'd put him at 26.24 runs above average.
---

Now, what I'm thinking is that PRAA is not an accurate measure of cross-evaluating... It looks useful in weighing Starters vs Starters, Closers vs Closers, etc. but I found this shocking when I tried to cross between the two.

Any insight into this would be awesome.

[thomasps3]

Ummm.....wow.

When someone can show me a pie chart, let me know. Until then, let me know if I can go and pump you some gas, praa man

[thedude]

This is why i have problems with some of the "invented Stats" (stats derived from the basic numbers). The ones from BP, which have no formulas aviable to the public, can be questionable. I know that they are trying to keep a monopoly on producing these numbers but still. I would like to see a real mathician or economist write papers without showing set by set calculation.

[mamorris]

Firstly, I think stats compared to a replacement player are a better indicator of a player's value than their stats compared to an average player. In this case, I would be looking at PRAR instead of PRAA. Secondly, this XIP stat appears to be overstating the importance of closers. It's (obviously) true that closers pitch in more important situations, so their impact should be increased. However, this article deals with "chaining", which is important when looking at the value of closers.

[bigh0rt]

Firstly, I think stats compared to a replacement player are a better indicator of a player's value than their stats compared to an average player. In this case, I would be looking at PRAR instead of PRAA. Secondly, this XIP stat appears to be overstating the importance of closers. It's (obviously) true that closers pitch in more important situations, so their impact should be increased. However, this article deals with "chaining", which is important when looking at the value of closers.

But is it necessarily obvious that closers pitch in important situations? Is the 9th inning, facing the 7, 8, 9 hitters, with a 4 run lead a more important situation than coming in with 2 out in the 7th to face the #3 hitter with 2 on, up by 3? I'd wager there's not a manager in the MLB who would have their closer in right then, even if it's the most 'important' batter of the game.