In my extremely amateurish spreadsheet I use "WAR per PA" as one of several weighted values to determine value. It is by far the least weighted and I'm not sure how I feel about it. It elevates the ranking of guys who retired before their decline phase and makes people like Lenny Dykstra look like borderline Hall of Fame players.

Damon is actually an example as I have him in a dead heat with Mike Cameron despite Damon playing almost 3000 more games.

Damon may be an excellent example of why I think, for HoF purposes, we should look more at WAA (average) than WAR. There's really no difference in the measures, just shifting the zero point. But it provides one method of adjusting for peak vs. career (and "penalizes" guys who hang on just for milestones) and it seems a more natural zero point for an HoF discussion -- i.e. you've got to be above-average to have a shot at the HoF. Also the concept of "replacement player" makes sense within the context of a single season -- i.e. if Damon got hurt on opening day, his team would need an emergency replacement -- but not for a career -- i.e. if Damon didn't exist, his teams would have (on average) had an average player out there.

Damon has 523 RAR. 339 of that is simple playing time. For his career, he's only about 18 wins above average or about 1 win per season. That's still way above Brock (a mere 3 WAA), ahead of Rice (13-14) and Perez (about 15), behind Dawson (22-23) and Puckett (20 WAA despite just 45 WAR).

As a contrast, you've got somebody like Berkman who has the same WAR total as Damon but has about 30 WAA. Or, more controversially, Ron Cey, Jose Cruz and Cesar Cedeno who are all about 25 WAA with the same WAR total. Or Mike Cameron with 47 WAR but about 23 WAA.

Here again the replacement-level concept falls over a bit. Damon has 3000 or so more PA than Cameron. So, sure, in our thought experiment comparing the two, Camereon's teams need to put somebody else out there for those 3000 PAs. But that's 4-5 full seasons worth of PAs -- the "sensible" counterfactual to "Cameron gets 3,000 more PA" is not "a replacement player gets 3,000 PA" but rather that an average player gets those 3,000 PA.*

WAA probably goes too far in the other direction of ignoring playing time value so I wouldn't necessarily have a problem with somebody claiming that Damon is Cameron's equal or just ahead of him. I would find it hard to make the case that Damon should be in the HoF while Cameron should not ... although the line will always be arbitrary and some group of players will get screwed.

*You can, if you want, treat it as slightly below average given Cameron's absence lowers the overall average.

I like Wins above average but zeroing out any below average season.

I am cool with the idea, for a HOF discussion, that below average play should not add to a player's case. I reject the idea that a player contributing to a ballclub (below average but better than replacement level) should ever hurt his HOF case.

3 - I agree - that's what I do.

I like Wins above average but zeroing out any below average season.

I am cool with the idea, for a HOF discussion, that below average play should not add to a player's case. I reject the idea that a player contributing to a ballclub (below average but better than replacement level) should ever hurt his HOF case.

So, what about Andruw Jones then? If you just look at his above average seasons then he's at least a borderline case, and I don't know if he fell bellow 0 WAR in any season except that first with LAD. How would ya'll treat his career?

With Jones you have to make up your mind how to treat his defense. By Total Zone he's the best defensive center fielder who ever lived. It's not just Total Zone, John Dewan's numbers are similar for the years we can compare, as are his UZR on Fangraphs. Michael Humphries in Wizardry ranks him at the top as well.

On the negative side, STATS based UZR and Zone Rating have him closer to an average defender.

If TZ/DRS/UZR/Wizardry is actually true then his peak value is enough to make him a deserving HOFer. Rate him merely a great CF and he belongs in the Hall of Very Good.

Jones looks pretty good by WAR as it is, he looks better by WAA since he's done that in a relatively short career.

Damon has produced 51.6 wins through his career over what a replacement-level player would be expected to produce in his place. By dividing career WAR by career games played for Damon, we wind up with .021. What does this mean? My interpretation is that over the course of his entire career, his team would expect to have a 2.1% greater chance of winning a game with Damon in the lineup versus a replacement-level player.

You could quibble about games vs. seasons vs. PA (or 550/600/650 PA) but roughly speaking this approach doesn't seem to tell you anything new.

He looks at it as Damon improving his team's WP by .021 -- well that's 3 wins per "full" season (is 150 games a full season or 162?). Or you take Damon's 51.6 WAR, divide by 17 seasons and you get -- 3 wins per season. Do it per 600 PA and it's 2.9; use 650 and it's about 3.2.

Which is fine -- WAA doesn't tell you anything new either and it's really a matter of which expression you find the easiest to relate to. Still you're left with the question of whether 17 years of being a consistently good but not excellent much less great player is more/less HoF-worthy than being an excellent/great player for 8-10 years.

I'm not a big fan of WAA because I think it fails to properly reward in-season durability. I could understand if you said an extra season or two of average play at the end of one's career shouldn't add anything to a HOF case, but I think an extra 30 games of average play each season during a guy's prime means something - it's not easy to find average players and especially not as one-month fill-ins.

Well, that's true as well - I also look at WAR.

Put another way, we know that the "true replacement value" isn't constant (within the year, across shorter v. longer timeframes, by position, by era) - but we accept certain assumptions for the purpsoes of simplifying comparisons and allowing us to all speak the same language (among other things). By also looking at WAA (or pWAA or the like) as well, we're adjusting the replacement level to account for the change in timeframe. (I don't think there's a 'right' answer in terms of how much to adjust - which is why I don't simply sub in an alternate replacement level and consolidate metrics.)

You gotta have WAA.