Tweet
If you have a system that you use to rank players, tell us about it. Most of us have a ranking of players, and I am curious as to how those lists are generated. Matt Souders, I won't ask you to explain PCA again, but if you've got any new tweaks to the system, feel free to let us know, like you always have. Basically, though, I am interested in seeing the methods other members at BBF have for making their rankings.
-
"Any pitcher who throws at a batter and deliberately tries to hit him is a communist."
- Alvin Dark -
My ranking system is pretty in depth, but I'll try to keep the explanation simple.
For pos players, I use:
Runs Created Above Avg
Batting Runs Above Avg
Batting Win Shares Above Avg
For pitchers, I use:
FIP ERA (runs saved compared to average)
Component ERA (runs saved compared to average)
ERA (runs saved compared to average)
Pitching Win Shares Above Average
I then find the standard deviation of all of these figures for each year and then calculate how many standard deviations above and below the average each player is, and then combine those results.
It's a lot more complicated than explained above, but the results look pretty good.
I believe that the only way to compare players of different eras is to compare each player to it's own era and assume that each era's level of play/difficulty was the same. My method above does a really good job accomplishing this. -
Dang, can we at least get your top 50 in order?Originally posted by Railsplitter View PostMythical SF Chronicle scouting report: "That Jeff runs like a deer. Unfortunately, he also hits AND throws like one." I am Venus DeMilo - NO ARM! I can play like a big leaguer, I can field like Luzinski, run like Lombardi. The secret to managing is keeping the ones who hate you away from the undecided ones. I am a triumph of quantity over quality. I'm almost useful, every village needs an idiot.
Good traders: MadHatter(2), BoofBonser26, StormSurgeComment
-
Lately I've been partial to Rally's quick and dirty method for position players:
B-Ref's Batting Wins
PLUS
Fielding runs converted to wins (TotalZone, STATS ZR, or UZR depending on years; could also include Walsh's arm ratings from recent seasons)
PLUS
Tango's positional adjustment (based on defensive innings if available from B-Ref or else PAs)
PLUS
Replacement Level (2 wins per 650 PAs)
It's quick, it's easy, it's pretty damn accurate.Comment
-
I agree on the quick and dirty method. We're at the point where it's no use really to argue about a 0.5 win difference between players. I have someone at 3 wins and someone else says 3.5... fine, let's move on.
The real disagreements are on the young players over a period of years, your Troy Tulowitzkis and Adam Jones, and your pitchers, who blow out their arms (either in injury or reduced effectiveness).
I would prefer seeing evaluations of talent with a 3 to 5 year time frame.Comment
-
I'll give you the top 25 position players and the top 25 pitchers:Originally posted by RuthMayBond View Post
01 - Pedro Martinez - 2000 - 44.6
02 - Cy Young - 1901 - 41.1
03 - Walter Johnson - 1912 - 41.0
04 - Pedro Martinez - 1999 - 40.9
05 - Roger Clemens - 1997 - 40.1
06 - Lefty Grove - 1931 - 39.6
07 - Dolf Luque - 1923 - 39.2
08 - Greg Maddux - 1995 - 39.2
09 - Steve Carlton - 1972 - 38.4
10 - Ron Guidry - 1978 - 38.2
11 - Greg Maddux - 1994 - 37.8
12 - Walter Johnson - 1913 - 37.7
13 - Randy Johnson - 2004 - 36.9
14 - Lefty Grove - 1930 - 36.7
15 - Pete Alexander - 1915 - 36.6
16 - Carl Hubbell - 1933 - 36.4
17 - Roger Clemens - 1990 - 36.3
18 - Randy Johnson - 1995 - 36.2
19 - Dazzy Vance - 1924 - 35.9
20 - Roger Clemens - 1986 - 35.7
21 - Dwight Gooden - 1985 - 35.4
22 - Bret Saberhagen - 1989 - 35.1
23 - Randy Johnson - 1999 - 34.9
24 - Steve Carlton - 1980 - 34.8
25 - Sandy Koufax - 1966 - 34.6
01 - Barry Bonds - 2004 - 61.9
02 - Barry Bonds - 2002 - 59.7
03 - Barry Bonds - 2001 - 57.8
04 - Rogers Hornsby - 1922 - 55.3
05 - Babe Ruth - 1921 - 55.0
06 - Babe Ruth - 1926 - 54.4
07 - Babe Ruth - 1923 - 51.6
08 - Barry Bonds - 1993 - 51.3
09 - Babe Ruth - 1920 - 50.4
10 - Honus Wagner - 1908 - 49.8
11 - Ted Williams - 1946 - 49.5
12 - Joe Morgan - 1976 - 49.1
13 - Rickey Henderson - 1990 - 48.8
14 - Mickey Mantle - 1956 - 48.7
15 - Ted Williams - 1947 - 48.6
16 - Rogers Hornsby - 1925 - 48.4
17 - Kevin Mitchell - 1989 - 48.1
18 - Mickey Mantle - 1957 - 47.9
19 - Barry Bonds - 2003 - 47.6
20 - Lou Gehrig - 1927 - 47.6
21 - Ty Cobb - 1917 - 47.4
22 - Barry Bonds - 1992 - 47.1
23 - Babe Ruth - 1924 - 46.8
24 - Mickey Mantle - 1961 - 46.7
25 - Robin Yount - 1982 - 46.5
To understand these numbers, take Bonds 2004 season for example. He was 6.19 standard deviations better than the league average. I just multiplied it by 10 to make it easier to view.
Remember, this system compares players to their respective positions and seasons. So for example, Lou Gehrig's value in the early 1930's goes down because his comparisons include Greenberg and Foxx at 1B, and also because of Babe Ruth.
Two rankings that I was shocked to see were Kevin Mitchell in 1989 and Rickey Henderson in 1990. But standard deviation doesn't lie in showing the true dispersion of players. Babe Ruth's 1927 season finished 29th overall at 45.5.
Also, defense is not calculated into these numbers. They are purely pitching and offense.Comment
-
Ok, you're doing top seasons and not careers. But hard to argue much with your choicesOriginally posted by Railsplitter View PostMythical SF Chronicle scouting report: "That Jeff runs like a deer. Unfortunately, he also hits AND throws like one." I am Venus DeMilo - NO ARM! I can play like a big leaguer, I can field like Luzinski, run like Lombardi. The secret to managing is keeping the ones who hate you away from the undecided ones. I am a triumph of quantity over quality. I'm almost useful, every village needs an idiot.
Good traders: MadHatter(2), BoofBonser26, StormSurgeComment
-
Agreed, Railsplitter that looks like a very good list. I might have to adopt that system for the time being. Let me know if you make any changes to it.Originally posted by DomenicComment
-
Well, here's what I do (start at bottom and read up):
Works out to be basically the same as the method that Sky mentioned above. The only difference, really, is that I calculate my own linear weights instead of using BRef's. Like Tango said, it generally won't matter a whole heck of a lot, especially for current players.
-j---
My blog: On Baseball and the RedsComment
Comment