As for dividing by PA, that was simply to find the players who put up good numbers but for various reasons had not played the full season...for curiosity sake, not as a way rank players.

Your solution is interesting because it sets a finite maximum (6) and minimum (0) to each total result. I experimented with dividing each player's raw numbers with "LgAvg per 600 PA" (40/7 for SBs in the case of Wills in 1963) instead of "largest number of x" (max(B:B)). In that system players like Wills, Taylor, Clendenon and Gilliam rank much higher because of their dominance in SBs that season compared the the weak league average of 7.

So instead of Aaron 5.38, Pinson 4.44, Mays 4.38, White 4.25, FRobinson 4.08, Cepeda 3.92, Mathews 3.88, BWilliams 3.87, Flood 3.82, McCovey 3.79 - you get Aaron 13.62, Pinson 11.08, Wills 10.82, FRobinson 10.55, Mays 9.35, White 9.28, Taylor 9.10, Flood 8.90, Clendenon 8.43, and Gilliam 8.34.

I am curious to hear your thoughts regarding dividing by LgAvg per 600 PA rather than (max(B:B)).

I'm sitting here pondering your decision to divide by PAs. On the one hand, I'm intrigued, but I'm thinking that statistically, it might be suspect. To this point, we've been doing dividing the stats accumulated by a given player, by the player who was #1 in that particular stat, obtaining a number that represents that ratio, between 0 and 1, repeating it for several stats, and then adding those ratios together. This represents, in a somewhat arbitrary (but still fairly grounded) way the comparison of that player with the dominant players in the league. It gives no weight to one stat over another, and summarizes the season output of that player.

You seem to want to evaluate the average amount of 'stats' a player accumulates, on average, per PA, using this method, and then compare. To obtain the seasonal stats above, and then to simply divide those values by PA...causes a problem, at least in my opinion, as you've not taken into account the PAs of the top player in that stat. If, for example, Player X get 100 runs in 500 PAs to lead the league, and Player Y gets 80 runs in 300 PAs, the raw number for player X would be 1.0000 (leading the league) and Player Y would be .8000 (80/100). In other words, Player Y did 80% as well as Player X on that stat. If you then divide .8000 by 300, you get .00267. But does that really mean anything? Player X is supposed to be the best in that stat, but if you did the same calculation for him, you'd get (1.0000/500) or .00200. So...it would seem that Player X is not the top player, on a per PA basis.

In my opinion, if you wish to evaluate players, not on their season numbers but on a value adjusted for PA, then you need to divide the stats by PA BEFORE you start. SB, BB, R, RBIs, TB can all be divided by PA to get a SB/PA, BB/PA, R/PA, RBI/PA and TB/PA, which is literally "what should I have expected, on average, from a player, in that P." If a player gives you, say 100 R on 500 PAs, then the value would be .20000 R/PA. Calculate this for all players. Then find the top value and repeat the calculations as before. This would, in my opinion, be closer to what you want.

It comes down to the question you want to answer (total season vs. expected per PA) and goes back to a JohnnyTwisto's comment about not accounting for park or position. This ratio stats account for none of that. They are simply a way of evaluating a single player against the top players in that league for those stats. A mini black ink/grey ink evaluator. Quick and dirty. Quite obviously a defensive stud, speed demon, heart of the dugout catcher that gives you 4.2340 out of 6.0000 points is more important than an all bat, no legs clubhouse cancer that DHs, but who happens to get 4.2500

What it does show is pure, unadulterated dominance. Coming in first in a category is impressive. Coming in first by 30% over second place is way more impressive.

http://www.baseball-reference.com/pi/shareit/kbuC5

(Gerald Laird?)

As for the Angels not getting their due, I was very surprised a few weeks ago during a Yankee telecast which polled the audience on who was Teixeira's main competition for MVP. 4 choices were given; I don't remember them all but Mauer was one. The poll winner was Kendry Morales. I was shocked because he's having a very good season but not an overwhelming one and I didn't think the average fan would be that aware of him.

I use my old system at various points through an active season to see which players come up that aren't in popular conversation that year. It was especially interesting last year when there was no single stand out in the American League and my system began singling out Pedroia before the media did. Pedroia wound up winning last year within the constructs of my system. (He also won the MVP).

This year the AL is again unresolved (MVP/best-season-of-the-year-wise) at this point in the season, while the NL has been wrapped up for some time (Pujols). A particular name continues to crop up lately (and leading the pack lately) in my old system that no one in the media is discussing for 2009 AL MVP rights. I was interested to see where this particular name ranked with this new system. Indeed, the same name currently leads the list using the new system too. Has anyone heard Chone Figgins' name crop up in the media as a contender for MVP this season? Have any of you considered him? Close behind in 2nd on the updated list is Abreu, who wasn't even in the top ten of my old system's list. Anyone have Abreu on their list of MVP candidates? Certainly, the Angels players are not getting their due respect in the mainstream media.

Using the new system, the top ten 2009 AL leaders are: Figgins 4.641, Abreu 4.635, Teixeira 4.492, Bay 4.480, Crawford 4.440, Jeter 4.400, BRoberts 4.325, CPena 4.289, Longoria 4.235, Ellsbury 4.205. Statistically, it's still a dead heat within these constructs. In terms of leadership, team result, popularity, media attention and intangibles, Jeter probably has a lock on it. For what it's worth, Figgins leads AL 3rd-basemen in Assists and is 3rd in DPs.

The system rewards quantity. How much raw data can a player amass. But what happens when we divide the raw number by plate appearances? Limiting it to those with 440 PAs or more, the leaders are: ARodriguez (29th on the raw list), Mauer (17th), Youkilis (14th), Bay, Zobrist (20th), Bartlett (30th), Nelson Cruz (34th), Abreu, CPena and Crawford. If you drop the PA limit to 300, Rajai Davis and Torii Hunter top the list. This really highlights what interesting years Zobrist, Cruz Davis and Hunter are having.

Figgins, Abreu and Hunter are dynamic players. The Angels are fun to watch and the team will get their deserved media attention in October.

What I think makes it more of a fun/junk measure and not something serious is the stats selected. They seem a bit arbitrary. Total base and walks tend to compete against each other. and no one thinks stolen bases should be weighted that high. Its kinda like metrics for ranking players in a roto-league they correlate with a players real offensive value quite a bit, but there are some "interesting differences".

Sabermetricians have been working on offense metrics for decades and in terms of run estimators, they've gotten quite good at it. We're all stats geeks here, we know how involved those estimators can be. A fun metric like this one doesn't really compare to those.

I don't want to spoil the fun here because these black ink measures are cool, but lets not get too carried away about how much they should be used in MVP discussions.

1. What would be the appropriate stats for a pitching version of this?
W? ERA? ERA+? WHIP? SO? SO/BB? 10 other stats?

2. Are there any hitting stats that should be considered added to the hitting version?
BB/SO? OPS+? (honestly...I like the 6 in there)

Of course, stuff like this can be overrated. We know how good Manny is offensively, and yet when he was replaced with Juan Pierre for 50 games, the Dodgers did even better.

So often, if a guy finishes first in HR and RBI in his league, he wins the MVP regardless of how his team did or how good the player's offensive performance really was. I like Slothbaby's system just as a single data point reality check on how good the season was.

1. Ryan Howard (2) 4.6681
2. Albert Pujols (0) 4.6582
3. Jose Reyes (1) 4.6427

Wow.
That being said, I still think BB/SO should be relevant here. 181 vs 50. Just sayin'.