I undertook the development of the Percentile Ranking System as a way of interpreting projection data. With so many projections out there, it helps you to compress a player’s expected performance down to one number, much like a dollar value. In testing out PRS in mock drafts, I’ve noticed some real weaknesses:

Even though they are the rarest offensive stat, steals are seemingly undervalued. Even though I think Jose Reyes is overvalued for his steals, he’s still much more valuable than PRS rates him.

Also, saves appear to be undervalued. Using a straight PRS projection would recommend drafting Rafael Betancourt, Jonathan Broxton, and Heath Bell ahead of 2/3 of the closers out there. While those players are not without value, unless your league scores holds, that is probably not a winning strategy. I recommend grabbing one of these three in the last round or two of your draft, BTW. Betancourt is preferred, since the current closer, Joe Borowski, is of dubious ability and is most likely to lose his job. These middle relievers will help stabilize your ERA and WHIP, and contribute significantly in the strikeout category.

As another way of approaching fantasy value, I’ve come up with a point system based on the rarity of stats in actual Yahoo! fantasy leagues. If you count the overall total of each stat, you can get a sense of how valuable that stat is in relation to the others. Here’s what I get using Cory Schwartz’s composite MLB Fantasy 411 projections, using numbers from four separate roto leagues last year:

Offensive stats

Steals are the scarcest commodity, and scaling everything in relation to pilfered bags, we see that 1 SB (1 point) is approximately the equivalent of 1.68 HR, 6.94 runs, and 6.76 RBI. Rate stats such as batting average get a little tricky, so here’s how I approached that. The league average BA was .285, but to calculate impact on your team average, you have to talk in units of hits, not batting average. So I awarded points based on how many more hits a player produced compared to an average player for his projected at-bats. Mathematically:

Points awarded for BA = (Projected BA * Projected AB) – (.285 * Projected AB)

So Hanley Ramirez is awarded 16.7 points for being projected to hit .312 in 2008. And Adam Dunn is penalized a whopping 17.2 points for being projected at .252.

Pitching stats

For pitchers, wins edged out saves as the rarest stat (unfortunately, they are also the hardest to predict because of run support). Anyway, assuming one win is one point, that is worth 1.20 saves and 13.33 Ks. ERA and WHIP are treated in a similar fashion to batting average above, except we award points based on earned runs saved over average for ERA (3.89) and walks + hits saved for average WHIP (1.286) for the projected IP.

I know that using constant scales like this is very crude mathematically, but this is meant as a kind of first approximation here. Using these point valuations gives us relative values for hitters and pitchers separately. Here’s the top 10 hitters:

  1. Hanley Ramirez
  2. Jose Reyes
  3. Carl Crawford
  4. Alex Rodriguez
  5. David Wright
  6. Albert Pujols
  7. Matt Holliday
  8. Miguel Cabrera
  9. Ryan Braun
  10. Ichiro Suzuki

As you can see, this method gives a little too much weight to steals, as no one would ever choose Carl Crawford ahead of Alex Rodriguez.

And now the top 10 pitchers:

  1. Johan Santana
  2. Jake Peavy
  3. Jonathan Papelbon
  4. J.J. Putz
  5. Joe Nathan
  6. Takashi Saito
  7. Brandon Webb
  8. C.C. Sabathia
  9. Mariano Rivera
  10. Francisco Rodriguez

Here again, wins and saves seem to dominate this list. Clearly not a good evaluative method. But where this points system errs, it balances out the weaknesses of PRS. So I converted to dollar values and averaged to get a new ranking. Hence, a hybrid evaluation system, yielding this top 20:

  1. Alex Rodriguez
  2. Hanley Ramirez
  3. David Wright
  4. Matt Holliday
  5. Albert Pujols
  6. Johan Santana
  7. Ryan Braun
  8. Miguel Cabrera
  9. Chase Utley
  10. Jake Peavy
  11. Jimmy Rollins
  12. Vladimir Guerrero
  13. Carl Crawford
  14. Jonathan Papelbon
  15. Jose Reyes
  16. David Ortiz
  17. Carlos Lee
  18. Alfonso Soriano
  19. Joe Nathan
  20. J.J. Putz

Still some small glitches, but more agreeable than either method on its own, I’d say. Crawford edged out Reyes by an insignificant amount, so you could invert the two easily. That has more to do with the projection than anything else.