I think this is the tension my “model”, more like an idea, has with the idea you need to find a player’s greatness.

Anyway, this conversation has made me think about this issue. Thanks :)

The method I outlined follows from a “players have value by helping their teams win the most games in a given season.” There are many other views to take, such as Tango’s idea that you want to figure out how great a player was based on his numbers — the longer he sustains those numbers, the more sure it’s not a result of luck. (Your Platonic model.) That’s another good method, but it suffers from the fact that players’ true talent level are constantly changing. The more playing time you require for a player to prove his greatness, the more chance you’ll be including numbers that weren’t put up when he was truly great.

If you’re going to use any sort of algorithm to spit out HoF rankings, you have to resign yourself to the fact that you can’t include all the little details of a player’s career. You can’t assign points to “best defensive player ever” or “brought fans back to the game” or “re-defined the way people play second-base”. Those things might matter in many definition of the Hall of Fame, but any objective measure needs to be strictly objective. I agree that balancing peak and longevity is the trickiest thing. Personally, I favor high peaks. I’d rather see a player have five truly amazing seasons worth 10 wins each than ten All-Star seasons worth 5 wins each. Hanging on for five years as a league-average (2-win) player should count next to nothing and is nowhere near one 10-win season.

Players have value to teams on a per-season basis. Peak performances are based on seasonal numbers and need not be consecutive.

The issue of whether a player’s peak seasons needs to be consecutive is one that I waver on, but I think of it in a slightly different way. If you think of baseball in a Platonic way, then a baseball player’s stats are a “sketch” of his form. The greater the sample, the more accurate the sketch; for Stephen Drew we have an Etch a Sketch of the type of player he is, for Jeff Kent we have a Mantegna.

When we are evaluating a player’s career, we are doing more than simply totaling his WAR, that’s the point of these posts, we are trying to figure out how good he was and who was he as a player. So, when we are evaluating Ozzie Smith, we know a great deal, which can be summarized as “maybe the best defensive player ever, a slap hitter who walked enough to be a productive hitter at his best, and a good baserunner”, and we can picture in our minds Ozzie both statistically and as he was in Busch in 1985. As a side note, this is what, I believe, people are looking to honor in Hall of Fame voting. I think this also fits in with your comment The only reason I’m doing these career profiles/graphs is because I’m not as familiar with many historical players as I’d like to be..

Consider the case of Barry Bonds (or to be less controversial, Toby Harrah) as the “posterboy” for separate peaks in a career; when considering his career, who is he? Is he the dynamic MVP from the early 90s, the slowly declining slugger from the late 90s, or the Ted Williams ‘57 clone? It seems somehow “wrong” to combine those disparate players into a FrankenBonds, although I don’t really have a solution. My next problem dovetails with another point you made

Additional seasons can only bolster a Hall of Fame candidacy.

I disagree with this (somewhat). If a player “falls of a cliff” proverbially speaking, I think we should recognize that, and not give him quite as much credit as someone who leaves the game while they still can play a little bit at the MLB level. I think less of Craig Biggio the player because I saw him suck at the MLB level for quite some time before he retired.

I agree with this next comment and think it could potentially be very interesting if researched, and be made applicable to cases such as the John Santana trade: where is the point of team quality where it becomes worth it to give away future equity for a small to medium level improvement in the present?

As far as what team you put the player on, I agree that it should be neutral, but I don’t think it has to be an average team, or one single theoretical team. You could take every single team in history, add your player, see how the win total or playoff probability changes, and then average all those scenarios. Or come up with a theoretical distribution of teams.

I’ll be interested to see that, although philosophically it’s difficult to resolve whether greatness or total value should be rewarded “more” in these types of discussions.