Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Baseball Newsstand > Baseball Primer Newsblog > Discussion
Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Monday, December 15, 2008

ESPN: Rank the top pitchers in MLB history

There are certainly some missing, but it’s worth checking out.  With 8 #1 votes, I suspect that Don Sutton’s mother is still alive.

CrosbyBird Posted: December 15, 2008 at 07:27 PM | 123 comment(s) Login to Bookmark
  Tags: awards, hall of fame

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

Page 1 of 2 pages  1 2 > 
   1. depletion Posted: December 15, 2008 at 08:34 PM (#3029613)
I started looking for Walter Johnson. Alas, he was not there, not Lefty Grove, Satchel Paige.
Sans everything.

I don't think this can be done in closed form. You need to run the Poisson distribution for the pitchers allowed runs per 9 and the summation of the opposition runs per 9 from 0 to 1 run less than the pitcher of interest to find the likelihood of win. Then find the likelihood of a tie in a similar fashion. The likelihood of a loss is 1 minus the sum of tie and win likelihoods. By multiplying win probability + 0.5 tie probability by IP/9 you get how many games the pitcher should have won over the given IP. I did this in Matlab for a couple pitchers some time ago. I could probably dig it up if provoked.

Peace.
   2. T.J. Posted: December 15, 2008 at 08:36 PM (#3029620)
Other than ranking Ryan last ("SportsNation" has him ranked #3; he's the John Elway of baseball), I wasn't more than 3 off from SN on any player (Carlton: me #3, SN #6; Clemens: me #4, SN #7; Koufax: me #5, SN #2).
   3. jacksone (AKA It's OK...) Posted: December 15, 2008 at 08:38 PM (#3029625)
I could probably dig it up if provoked.


Consider yourself provoked, sounds interesting.
   4. T.J. Posted: December 15, 2008 at 08:38 PM (#3029626)
I started looking for Walter Johnson. Alas, he was not there, not Lefty Grove, Satchel Paige.


"ESPN.com's Rob Neyer has compiled a list of the best pitchers dating back to the mid-1940s"
   5. HGM Posted: December 15, 2008 at 08:38 PM (#3029627)
This really should be renamed "Rank the top pitchers of the last 50 or so years".
   6. The importance of being Ernest Riles Posted: December 15, 2008 at 08:39 PM (#3029629)
It's more like a Weibull distribution, and I played around with this on THT a little while ago. I only integrated from 0 to the league-mean ERA, but your idea sounds pretty sweet, too.
   7. T.J. Posted: December 15, 2008 at 08:39 PM (#3029631)
This really should be renamed "Rank the top pitchers of the last 50 or so years."

Agreed. The title is terribly misleading.
   8. Famous Original Joe C Posted: December 15, 2008 at 08:40 PM (#3029634)
Sandy Koufax is criminally overrated. What's the argument for Koufax over Pedro, let alone Clemens/Maddux/Spahn?

I guess I shouldn't bang my head against the wall over what ESPN.com readers think, but seriously!

Pedro ~2700 IP, 2.91 ERA, very high offense era, bulk of career in hitters park

Koufax ~2300 IP, 2.76 ERA, extreme pitchers era, and best pitchers park of said era to boot

I know Koufax pitched a bunch more innings than Pedro in their respective peak seasons, but Pedro's were of clearly better quality per inning.
   9. HGM Posted: December 15, 2008 at 08:41 PM (#3029636)
Really rough ranking. Didn't put much thought into it:

1) Clemens
2) Maddux
3) Seaver
4) Johnson
5) Spahn
6) Pedro
7) Koufax
8) Carlton
9) Wynn
10) Glavine
11) Perry
12) Niekro
13) Ryan
14) Sutton
   10. JJ1986 Posted: December 15, 2008 at 08:47 PM (#3029650)
Mine:

1. Seaver
2. Maddux
3. Clemens
4. Spahn
5. Koufax
6. Pedro
7. Big Unit
8. Carlton
9. Ryan
10. Niekro
11. Perry
12. Glavine
13. Sutton
14. Wynn

I didn't look anything up so I'd probably change them if I actually researched it.
   11. Howie Menckel Posted: December 15, 2008 at 08:48 PM (#3029652)
Winningest since mid-1940s, as required, and not listed are Tommy John, Blyleven, Robin Roberts, Kaat, Mussina, Jim Palmer, unclear if Feller is eligible, Gibson, Moyer, Dennis Martinez...

Palmer and Whitey Ford and Gibson could have made it 17, with all the 300-game winners and all the others worth mentioning, I think (assuming he skips Feller)

(Schilling Smoltz Eckersley Blyleven Roberts fight it out if a top 20, maybe Mariano at this juncture)
   12. aleskel Posted: December 15, 2008 at 08:49 PM (#3029653)
since no one else said it, I'll just go ahead and point out that it's a travesty that Sutton is on this list and Blyleven is not.
   13. depletion Posted: December 15, 2008 at 08:51 PM (#3029655)
Phil Niekro pitched boatloads of innings. Even if a lot of them were only slightly better than average, I tend to think of him as underrated.
Hoyt Wilhelm had fewer IP than most of these guys but with a very low ERA.

I wish I had more time to look at these issues.
   14. Santanaland Diaries Posted: December 15, 2008 at 08:54 PM (#3029661)
since no one else said it, I'll just go ahead and point out that it's a travesty that Sutton is on this list and Blyleven is not.


Gibson's the more glaring omission though. and you'd think the legend of his 1968 would make him a natural for this sort of exercise.
   15. Khrushin it bro Posted: December 15, 2008 at 08:56 PM (#3029663)
#1 Tim Lincecum
   16. NJ in NY (Now with Big Girl!) Posted: December 15, 2008 at 08:57 PM (#3029665)
Mine (didn't research and would love to hear which guys my subjective analysis is really off base for)

1. 1.Roger Clemens
2. 2.Warren Spahn
3. 3.Greg Maddux
4. 4.Pedro Martinez
5. 5.Sandy Koufax
6. 6.Randy Johnson
7. 7.Steve Carlton
8. 8.Tom Seaver
9. 9.Gaylord Perry
10. 10.Phil Niekro
11. 11.Early Wynn
12. 12.Nolan Ryan
13. 13.Don Sutton
14. 14.Tom Glavine
   17. The Most Interesting Man In The World Posted: December 15, 2008 at 09:06 PM (#3029682)
I don't think this can be done in closed form. You need to run the Poisson distribution for the pitchers allowed runs per 9 and the summation of the opposition runs per 9 from 0 to 1 run less than the pitcher of interest to find the likelihood of win. Then find the likelihood of a tie in a similar fashion. The likelihood of a loss is 1 minus the sum of tie and win likelihoods. By multiplying win probability + 0.5 tie probability by IP/9 you get how many games the pitcher should have won over the given IP. I did this in Matlab for a couple pitchers some time ago. I could probably dig it up if provoked.

I work as an actuary, yet I now have a headache.
   18. Fridas Boss Posted: December 15, 2008 at 09:16 PM (#3029695)
17. The Most Interesting Man In The World....

I work as an actuary....



Does not compute..
   19. eric Posted: December 15, 2008 at 09:20 PM (#3029701)
Off the top of my head I have:

1. Roger Clemens
2. Greg Maddux
3. Tom Seaver
4. Warren Spahn
5. Steve Carlton
6. Gaylord Perry
7. Phil Niekro
8. Randy Johnson
9. Pedro Martinez
10. Sandy Koufax
11. Tom Glavine
12. Nolan Ryan
13. Don Sutton
14. Early Wynn

I'm very comfortable with all those rankings except the exact placement of Randy Johnson, which I can't see lower than 8 but also of course no higher than 3. I took Bill James' lead and just ranked him as low as I think I could get away with until I can gain a little perspective.
   20. DL from MN Posted: December 15, 2008 at 09:20 PM (#3029702)
Gibson and Blyleven are both huge omissions, Robin Roberts is another. Sutton and Wynn? Did they need consensus picks for the bottom?
   21. Famous Original Joe C Posted: December 15, 2008 at 09:22 PM (#3029706)
I don't think this can be done in closed form. You need to run the Poisson distribution for the pitchers allowed runs per 9 and the summation of the opposition runs per 9 from 0 to 1 run less than the pitcher of interest to find the likelihood of win. Then find the likelihood of a tie in a similar fashion. The likelihood of a loss is 1 minus the sum of tie and win likelihoods. By multiplying win probability + 0.5 tie probability by IP/9 you get how many games the pitcher should have won over the given IP. I did this in Matlab for a couple pitchers some time ago. I could probably dig it up if provoked.

I work as an actuary, yet I now have a headache.


Yep. MS in Stats and I completed glazed over.
   22. Repoz Posted: December 15, 2008 at 09:23 PM (#3029707)
Ford?
   23. eric Posted: December 15, 2008 at 09:35 PM (#3029718)
Definitely a few omissions. If they were trying to keep to a nice round number like 10 or 15 I could understand that there might be some discussion over who really deserves to be at the bottom. But 14? No Roberts? No Gibson?

Like Howie said in #11, just go to 20, add Roberts, Gibson, Ford, Palmer, Blyleven, and Wilhelm. Or Rivera. Hell go to 21 and add Rivera, too.

It's obvious whoever compiled this list just took all the 300 game winners and added the best of the modern pitchers, and then realized, well, you can't have a best pitchers list without SANDY KOUFAX, and called it a day.
   24. snapper (history's 42nd greatest monster) Posted: December 15, 2008 at 09:43 PM (#3029728)
Palmer and Whitey Ford and Gibson could have made it 17, with all the 300-game winners and all the others worth mentioning, I think (assuming he skips Feller)

Ford, Gibson, Palmer and Marichal need to be on this list, and Wynn, Niekro, Sutton, and Glavine need to be off. Ryan doesn't really belong, but he's a freak, so he can stay.
   25. Hugh Jorgan Posted: December 15, 2008 at 09:53 PM (#3029744)
#1, I thought the same thing. Where the hell is the big train and Mathewson, surely they are in the conversation? Lefty Grove? c'mon surely there's a metric where these guys are in the top 14, well at least ahead of Sutton and Niekro.

Sure, Koufax wins that very short peak rating game...maybe Pedro beats him there, but now you are talking the best of all time. So you need to be better than 5 incredible years. Gibson? Palmer?

Let's put it this way, you are the GM of the all time team for 9-10 years, who are your 4-5 starters for those 9-10 years and you can only have the same guys for those 9-10 years? What I'm getting as is your guys need to have some level of longevity along with brilliance.
   26. Hugh Jorgan Posted: December 15, 2008 at 10:01 PM (#3029756)
and another thing, Clemens is way too low for that final list, he's down in 7th. Heck if I had to pick 2 starters on that list that I needed for at least 10 years I probably go with Clemens and Seaver(very close call with Maddux).
   27. snapper (history's 42nd greatest monster) Posted: December 15, 2008 at 10:03 PM (#3029761)
Sandy Koufax is criminally overrated. What's the argument for Koufax over Pedro, let alone Clemens/Maddux/Spahn?

I agree completely. Compare Koufax to a contemporary, Whitey Ford, who you never hear in the "greatest pitcher" debates.
Ford has him (by a lot) in IP and W's, and the quality is there too, ERA+ Ford 133, Koufax, 131.

All Koufax has is four otherwordly seasons, but in one of the greatest pitching environments ever contructed (1960's Dodger Stadium), with an illegally high mound. I have a sneaking suspicion that if the Dodger'stay in Ebbetts Field, Koufax's rep is more like Sal Maglie's.

Look at his Home/Road ERA's
1963 Home 1.38 Road 2.31
1964 Home 0.85 Road 2.93
1965 Home 1.38 Road 2.72
1966 Home 1.52 Road 1.96

Pedro has him beat in quantity and quality.
   28. Replacement-Level Primate Posted: December 15, 2008 at 10:04 PM (#3029764)
It's obvious whoever compiled this list just took all the 300 game winners and added the best of the modern pitchers, and then realized, well, you can't have a best pitchers list without SANDY KOUFAX, and called it a day.


Agreed. Rob Neyer should lose his BBWAA membership over this. And where is Pitcher of the '80s, Jack Morris?!?!
   29. The Most Interesting Man In The World Posted: December 15, 2008 at 10:05 PM (#3029766)
17. The Most Interesting Man In The World....

I work as an actuary....


Does not compute..


Don't hate! Calculate.... and depreciate!
   30. Juan V Posted: December 15, 2008 at 10:22 PM (#3029780)
Pretty strong ranking. I voted:

# Clemens
# Maddux
# Seaver
# Carlton
# Big Unit
# Spahn
# Pedro
# Perry
# Ryan
# Niekro
# Koufax
# Glavine
# Wynn
# Sutton
   31. Voros McCracken of Pinkus Posted: December 15, 2008 at 10:47 PM (#3029810)
"So you need to be better than 5 incredible years."

Why? I'm not saying you're right or wrong, it's just not obvious to me one way or the other. We all understand why one year isn't sufficient: it's too small of a sample and lots of guys can have one freak Rick Wilkins season that makes them look like a HOFer. Two seasons is probably out for the same reason.

But I'm not sure why five wouldn't be enough, particularly in a row.
   32. ERROR---Jolly Old St. Nick Posted: December 15, 2008 at 10:54 PM (#3029820)
Palmer and Whitey Ford and Gibson could have made it 17, with all the 300-game winners and all the others worth mentioning, I think (assuming he skips Feller)

Ford, Gibson, Palmer and Marichal need to be on this list, and Wynn, Niekro, Sutton, and Glavine need to be off. Ryan doesn't really belong, but he's a freak, so he can stay.


Given that this is Neyer, I'm more than a bit surprised at these inane omissions and commissions. Seems like a function more of a lack of video highlights than anything else, which would be typical of ESPN.


Top 15 (since 1945):

1t. Maddux
1t. Clemens (not considering steroids for purposes of this list)
3. Seaver
4. Pedro (# 1 on peak value)
5. Spahn (almost never pitched against the Dodgers; big minus for that)
(You could argue forever about the relative position of Pedro and Spahn)
6. Johnson
7. Carlton
8. Koufax (would be higher on a pure peak value that included the postseason, but obviously lower on a purely career value basis)
9. Gibson
10. Palmer
11. Ford
12. Marichal
13. Perry
14. Glavine
15. Schilling
   33. CrosbyBird Posted: December 15, 2008 at 10:57 PM (#3029822)
Sorry about reinforcing the crappy ESPN title with my own, although I did warn you that there were some omissions.

I had Clemens, Maddux, RJ, Seaver, and Carlton as my top 5, but I'm not married to the order of 3-5. What stands out most as a difference is the placement of Ryan and Koufax, both of whom I ranked fairly low (but ahead of Sutton).

I may be criminally underrating Spahn. I don't really start to grok a comfortable adjustment for environment until around the early 70s.
   34. Hugh Jorgan Posted: December 15, 2008 at 11:12 PM (#3029842)
#31
I just reckon when you are picking the all time greatest MLB pitchers career length needs to come into at some stage. Sure if your managing a team for 3 years you can pick peak value, who wouldn't want Pedro and Koufax going back to back during their peak? I just think that at least 8 peak years is required to be part of the discussion. Isn't that what makes most of these guys the best of all time? Enormously long peak periods where you can count on them year after year after year.
I'm sure there's a length of peak we are all comfortable with, I just put mine at a minimum of 8-9 years...
   35. with Glavinesque control and Madduxian poise Posted: December 15, 2008 at 11:23 PM (#3029854)
For the record, I prefer Maddux's peak to Koufax's, and more people would agree with me if 1994-95 weren't strike years.
   36. vortex of dissipation Posted: December 15, 2008 at 11:25 PM (#3029856)
All Koufax has is four otherwordly seasons, but in one of the greatest pitching environments ever contructed (1960's Dodger Stadium), with an illegally high mound. I have a sneaking suspicion that if the Dodger'stay in Ebbetts Field, Koufax's rep is more like Sal Maglie's.

Look at his Home/Road ERA's
1963 Home 1.38 Road 2.31
1964 Home 0.85 Road 2.93
1965 Home 1.38 Road 2.72
1966 Home 1.52 Road 1.96

Pedro has him beat in quantity and quality.


In those four seasons, the Dodgers won three pennants, and Koufax won three unanimous Cy Young Awards when there was only one award for both leagues. I don't think it's possible to overestimate the respect in which Koufax was held at the time, and that really doesn't come across when just computing numbers...
   37. Howie Menckel Posted: December 15, 2008 at 11:26 PM (#3029857)
We have many peak, prime and career voters at Hall of Merit.
They're all reasonable approaches, I think.
   38. Tiboreau Posted: December 15, 2008 at 11:32 PM (#3029861)
PITCHER  . . . . . . . . IP OPSWARP3  5PK
1. Roger Clemens 
. . . 4916.2 143  209.0 65.6
2. Greg Maddux 
. . . . 5008.1 132  171.8 59.5
3. Randy Johnson 
. . . 4039.1 137  154.8 59.6
4. Tom Seaver  
. . . . 4782.2 127  146.6 55.7
5. Steve Carlton 
. . . 5217.1 115  138.7 56.4
X
Bert Blyleven . . . 4970.0 118  147.7 53.8
6. Warren Spahn  
. . . 5243.2 118  143.8 51.5
7. Gaylord Perry 
. . . 5350.1 117  130.9 52.5
X
Bob Gibson  . . . . 3884.1 127  118.6 57.0
X
Robin Roberts . . . 4688.2 113  126.0 54.6
X
Bob Feller  . . . . 3827.0 122  104.7 59.6
8. Pedro Martinez  
. . 2782.2 154  116.4 58.6
9. Phil Niekro 
. . . . 5404.1 115  132.9 48.9
10. Nolan Ryan 
. . . . 5386.0 111  135.5 47.6
11. Tom Glavine  
. . . 4413.1 118  129.1 45.7
XX
Jim Palmer . . . . 3948.0 126 99.6 47.9
12. Sandy Koufax 
. . . 2324.1 131 68.3 48.7
13. Early Wynn 
. . . . 4564.0 107  107.4 42.7
XX
Whitey Ford  . . . 3170.1 133 98.5 37.0
14. Don Sutton 
. . . . 5282.1 108  106.9 32.9 

There have been discussion before about whether ERA+ overrates top pitchers in high run scoring eras; both Randy Johnson and Pedro Martinez score rather well according to WARP, particularly Johnson.

Bert Blyleven's WARP speaks to how highly he is thought of statistically; however, he may be overrated due to how much you believe in the ability to pitch to the score, etc.

The numbers next to Bob Feller do not include any WWII credit. I know some people are leery of giving pitchers WWII credit, feeling that they may have benefited from time off and the potential for injury, while others just don't give WWII credit. I did put above Pedro due to on this list due to Feller's missed time, however.

Phil Niekro's been a bit underrated, both on ESPN and this thread, IMO.

Nolan Ryan, one of the most overrated ballplayers in MLB history, may be underrated by the more statistically inclined. . . .

Jim Palmer scores a little lower than is reputation suggest, at least in this thread. The issue is related to the brilliance of his defenses behind him--how much credit goes to which, particularly if Palmer, recognizing an advantage, pitched to his defenses.

Whitey Ford scores lower according to WARP than his ERA+ or reputation on this thread would suggest. I haven't looked to see how good the defenses behind him were, but IIRC another facet of his game was how his manager used him--Casey Stengel, being Casey Stengel, micro-managed his pitching staffs just like he did his batting lineups, matching his best starters against better teams & starting pitchers, leaving Ford with fewer IP than he might've achieved under manager.

It looks like the list was created to get a spread of the talent elected to the Hall of Fame in a short list--Blyleven's eliminated because he hasn't been elected; Gibson's eliminated because of Koufax; Roberts is eliminted due to the presence of Spahn; Feller's eliminated due to WWII and/or debuting prior to the war. Tom Glavine's probably included due to being a long-time teammate of Maddux, while Clemens & Johnson included as contemporary future HoFers. These are all guesses, and I have no idea why they didn't include Palmer & Ford (or Marichal or Fergie Jenkins).

Clemens is way too low for that final list, he's down in 7th. Heck if I had to pick 2 starters on that list that I needed for at least 10 years I probably go with Clemens and Seaver(very close call with Maddux).

Clemens reputation is shot, just like McGwire's--his talent will be seen as an product of foreign substances and he will forever be underrated according to his statistics, just as Nolan Ryan & Sandy Koufax will be considered overrated.
   39. dze27 Posted: December 15, 2008 at 11:33 PM (#3029863)
Just looking at one pairing -- Spahn/Randy Johnson -- Spahn had exactly two seasons ('47, '53) with an ERA+ higher than Johnson's career ERA+ of 137 (as well as 4 fewer Cy Youngs). I think Spahn should be around 6th-8th or so on that list (and Johnson somewhere in the 3rd-5th range). Spahn does have 1200 IP on Johnson, which is a lot, but Johnson's peak is ridiculously better. I'm more of a peak guy, and it's not like Johnson had a short career with 295 W and 4000 IP, and counting.
   40. vortex of dissipation Posted: December 15, 2008 at 11:36 PM (#3029865)
Yeah, there's no argument for Koufax over Clemens/Maddux/Spahn.


For career performance, obviously not, but I think Koufax quite clearly has an argument over Spahn for peak.
   41. Famous Original Joe C Posted: December 15, 2008 at 11:38 PM (#3029868)
In those four seasons, the Dodgers won three pennants, and Koufax won three unanimous Cy Young Awards when there was only one award for both leagues. I don't think it's possible to overestimate the respect in which Koufax was held at the time, and that really doesn't come across when just computing numbers...

All true, and a fair point. It doesn't change the fact that, when you compare the breadth of his entire career to other all-time greats, he pales in comparison. You'd have to only consider a pitcher's best four or five consecutive seasons to consider him inner circle, or else use the Jim Rice "yeah, but you had to be there!" argument.

Deserving Hall of Famer? Yes, absolutely. Amazing peak? Yes, one of the very best ever, even adjusting for his context.

In the same class as Pedro, let alone Maddux/Clemens/Spahn, etc? No chance, no matter what people thought of him at the time, and it's not even particularly close.
   42. NetOwl Posted: December 15, 2008 at 11:58 PM (#3029883)
You need to run the Poisson distribution for the pitchers allowed runs per 9 and the summation of the opposition runs per 9 from 0 to 1 run less than the pitcher of interest to find the likelihood of win. Then find the likelihood of a tie in a similar fashion. The likelihood of a loss is 1 minus the sum of tie and win likelihoods. By multiplying win probability + 0.5 tie probability by IP/9 you get how many games the pitcher should have won over the given IP. I did this in Matlab for a couple pitchers some time ago. I could probably dig it up if provoked.


Thanks for practically forcing me to write my own script for doing this, though I needed to get some practice with the new computation program anyway.

It works so far with test cases involving a guy who gives up almost no runs (can't do 0 because of the poisson calculations) and a guy who gives up as many as his team scores on average.
   43. ERROR---Jolly Old St. Nick Posted: December 16, 2008 at 12:04 AM (#3029885)
Deserving Hall of Famer? Yes, absolutely. Amazing peak? Yes, one of the very best ever, even adjusting for his context.

In the same class as Pedro, let alone Maddux/Clemens/Spahn, etc? No chance, no matter what people thought of him at the time, and it's not even particularly close.


I put Koufax below all four of those, and below three others, but it's not that lopsided. He loses out because of the brevity of his peak, but within that peak, all those extra innings represent considerable added value over any modern pitcher. Not enough to make up for other differences (career lenghth; Pedro's insane peak), but enough to keep him within a respectable distance.

And none of those pitchers were (in hindsight) as guaranteed to come up big in the biggest games as Koufax. Small sample size may have much to do with this, but of the pitchers above him on my list, only Spahn (barely) and Carlton (by about half a run) have postseason ERAs lower than their regular season's. Koufax's World Series ERA was a third of his regular season's.
   44. OCF Posted: December 16, 2008 at 12:23 AM (#3029897)
1. Clemens
2. Seaver
3. Maddux
4. Spahn
5. Gibson
6. R. Johnson
7. P. Martinez
8. G. Perry
9. Carlton
10. P. Niekro
11. Blyleven
12. Jenkins
13. Roberts
14. Ryan
15. Sutton
16. Glavine
17. Schilling
18. Mussina
19. Smoltz
20. Bunning
21. K. Brown

Going by FWP based on career RA+ equivalent record, with a slight bias against 1990's-2000's pitchers because of the decentralization of ERA+/RA+ (which has essentially nothing to do with it being high scoring; instead, it's a byproduct of the reduction in individual IP.) It's a career perspective, and Koufax doesn't crack the top 20.
   45. GGC Posted: December 16, 2008 at 12:33 AM (#3029911)
I don't think this can be done in closed form. You need to run the Poisson distribution for the pitchers allowed runs per 9 and the summation of the opposition runs per 9 from 0 to 1 run less than the pitcher of interest to find the likelihood of win. Then find the likelihood of a tie in a similar fashion. The likelihood of a loss is 1 minus the sum of tie and win likelihoods. By multiplying win probability + 0.5 tie probability by IP/9 you get how many games the pitcher should have won over the given IP. I did this in Matlab for a couple pitchers some time ago. I could probably dig it up if provoked.


This sounds similar to the support neutral stats at BPro.
   46. Eugene Freedman Posted: December 16, 2008 at 12:52 AM (#3029927)
How many mothers does Nolan Ryan have to come in 3rd in the polling. He and Don Sutton are essentially identical pitchers. Oh, and Blyleven too. Where is he? And, Fergie Jenkins...
   47. Mirabelli Dictu (Chris McClinch) Posted: December 16, 2008 at 12:52 AM (#3029928)
With no research done and without consulting this thread first, I came pretty close to the consensus here:

1. Clemens
2. Maddux
3. Seaver
4. Pedro
5. Johnson
6. Koufax
7. Carlton
8. Spahn
9. Perry
10. Wynn
11. Ryan
12. Glavine
13. Sutton
14. Niekro
   48. Jeff K. Posted: December 16, 2008 at 01:01 AM (#3029944)
That is a ##### to rank. I started throwing guys at the top or bottom, just to weed out, kind of slowly maneuvering them around. I look and think "Okay, I'm comfortable with my bottom 5, but damn, Carlton is 7th? That just doesn't sound right. At least I only have two more to put in here." And then realized those two were Seaver and Spahn.

I don't like how my list is so heavily slanted to recent at the top end, but I think they're all justifiable choices where they are:

1 - Maddux
2 - Johnson
3 - Clemens
4 - Spahn
5 - Martinez
6 - Seaver
7 - Carlton
8 - Koufax
9 - Glavine
10 - Perry
11 - Wynn
12 - Sutton
13 - Niekro
14 - Ryan
   49. Voros McCracken of Pinkus Posted: December 16, 2008 at 01:02 AM (#3029945)
Here's a DIPS/replacement thing for giggles (1946 to 2007):

1. Clemens - 103
2. Ryan - 85
3. Unit - 85
4. Craig Calcaterra's Love Slave - 74
5. Blyleven - 69
6. Pedro - 66
7. Schilling - 63
8. Gaylord - 63
9. Seaver - 59
10. Carlton - 57
11. Kevin Brown - 54
12. Mussina - 53
13. Roberts - 53
14. Smoltz - 52

With 2008 mixed in Johnson passes Ryan into 2nd place. Pedro doesn't catch Blyleven. Mussina passes Brown, not sure about Carlton. Don't know whether Smoltz catches Roberts.

The Nolan Ryan thing is mostly to do with his runs allowed not matching his component stats. The DIPS predictions for hits matches up pretty well with reality there. Ryan is one of the best pitchers ever by component ERA so it makes sense he should be up there with DIPS too. So that's a discussion for another day.

The top 5 in Win% given 100 decisions:

1. Pedro - .673
2. Mariano - .671
3. Hoffman - .651
4. Unit - .644
5. Clemens - .636
6. Johan - .628
7. Doug Jones?!? - .624
8. Schilling - .621
9. Lee Smith - .617
10. Oswalt - .614
11. Webb - .609
12. Koufax - .599
13. Kevin Brown - .596
14. Ryan - .595

Obviously the 2008 numbers would affect a lot of these. Surprised to see how good Doug Jones was and for how long. Wonder what his career would have looked like had he got it started before 30.

Honorable Mention to Phil Ortega and Tommy Byrne, at the bottom of both lists.
   50. Howie Menckel Posted: December 16, 2008 at 01:08 AM (#3029953)
There's zero question that the list selection was "all post-WW II 300-game winners, plus the guys we can't not put on the list."

If Sutton wasn't the tipoff, Early Wynn was the clincher.

I'm pretty sure they considered Feller ineligible as well, a little ironic in that he's still alive and some others are dead.
He's too old to be part of a long "modern era," and yet.....
   51. Juan V Posted: December 16, 2008 at 01:11 AM (#3029955)
There's zero question that the list selection was "all post-WW II 300-game winners, plus the guys we can't not put on the list."


Pretty much. But I think Bob Gibson qualifies as "we can't not put him". Maybe someone on ESPN had a brainfart.
   52. Voros McCracken of Pinkus Posted: December 16, 2008 at 01:11 AM (#3029957)
One of the problems with Feller is that while he pitched for a good long time after WWII, most of the really good stuff (other than 1946) came before.
   53. Voros McCracken of Pinkus Posted: December 16, 2008 at 01:20 AM (#3029968)
BTW, other than Blyleven, the highest rated eligible pitcher not in the HOF is Bob Friend. Friend gave the peak of his career to the lousy Pirates teams of the early to mid 50s. However directly behind him is a pretty big list of guys (Rick Reuschel, Tommy John, Lolich, Saberhagen, Koosman, Chuck Finley and Camilo Pascual) who aren't in. They all rank ahead of Koufax who comes after Pascual.

Also I forgot to throw the knuckleball effect in there so move Niekro up a bunch.
   54. Howie Menckel Posted: December 16, 2008 at 01:26 AM (#3029975)
The list is very simple:
- 300-game winners after WW II
- Koufax, the all-time mystique
- Pedro and RJ, active and legendary

That's it. That's why Palmer and Gibson and Ford aren't on it.
   55. CrosbyBird Posted: December 16, 2008 at 01:32 AM (#3029980)
How many mothers does Nolan Ryan have to come in 3rd in the polling. He and Don Sutton are essentially identical pitchers.

I'm inclined to say that Ryan is monstrously overrated, but not that he's the same as Don Sutton. The guy is 11th all-time in black ink.
   56. Karl from NY Posted: December 16, 2008 at 01:35 AM (#3029986)
I don't think this can be done in closed form. You need to run the Poisson distribution for the pitchers allowed runs per 9 and the summation of the opposition runs per 9 from 0 to 1 run less than the pitcher of interest to find the likelihood of win. Then find the likelihood of a tie in a similar fashion. The likelihood of a loss is 1 minus the sum of tie and win likelihoods. By multiplying win probability + 0.5 tie probability by IP/9 you get how many games the pitcher should have won over the given IP. I did this in Matlab for a couple pitchers some time ago. I could probably dig it up if provoked.

Computer science major here; I skipped that paragraph at first, but after seeing further comments, went back and understood it fine.

Is a Poisson distribution correct for modeling baseball runs, though? For a Poisson distribution, the events should be independent, but baseball runs aren't. Runs are correlated to each other (2 straight doubles scores more than 2 doubles in separate innings), and successive at-bats are interdependent by clustering good hitters in the batting order and by pitcher fatigue. Or do you know this and just intentionally use Poisson as a good-enough approximation?
   57. John DiFool2 Posted: December 16, 2008 at 01:44 AM (#3029993)
No way Seaver should be below Ryan. Tom Terrific's legacy/visibility seems to be slipping a bit-you don't see him much in the news or mentioned prominently in such discussions or whatever. Didn't help that his former teammate, almost the same age, pitched 6 years longer than he did.
   58. Voros McCracken of Pinkus Posted: December 16, 2008 at 02:06 AM (#3030020)
Is a Poisson distribution correct for modeling baseball runs, though?

No. Runs are bunched far more than poisson would predict.

Almost works for soccer though (it's a little off, but close enough for most purposes).
   59. AJMcCringleberry Posted: December 16, 2008 at 02:12 AM (#3030023)
1.Roger Clemens
2.Greg Maddux
3.Warren Spahn
4.Tom Seaver
5.Randy Johnson
6.Steve Carlton
7.Phil Niekro
8.Gaylord Perry
9.Pedro Martinez
10.Tom Glavine
11.Nolan Ryan
12.Early Wynn
13.Don Sutton
14.Sandy Koufax

6-10 are interchangable.
   60. PASTE, Now with Extra Pitch and Extra Stamina Posted: December 16, 2008 at 02:19 AM (#3030028)
1. Roger Clemens
2. Greg Maddux
3. Tom Seaver
4. Randy Johnson
5. Warren Spahn
6. Steve Carlton
7. Pedro Martinez
8. Phil Niekro
9. Tom Glavine
10. Gaylord Perry
11. Sandy Koufax
12. Nolan Ryan
13. Don Sutton
14. Early Wynn

Johnson could be higher. Randy Johnson was the pitcher people think Sandy Koufax was.
   61. caspian88 Posted: December 16, 2008 at 02:32 AM (#3030040)
Before reading this thread, and off the top of my head:

1.Roger Clemens
2.Randy Johnson
3.Greg Maddux
4.Tom Seaver
5.Gaylord Perry
6.Pedro Martinez
7.Steve Carlton
8.Warren Spahn
9.Phil Niekro
10.Nolan Ryan
11.Sandy Koufax
12.Tom Glavine
13.Early Wynn
14.Don Sutton
   62. pkb33 Posted: December 16, 2008 at 02:47 AM (#3030058)
Pedro has him beat in quantity and quality.

I agree with your overall conclusion (if you're talking peak, Pedro is the greatest pitcher ever, imo) but the thing you do have to give Koufax is that he threw a lot more innings. Granted, the league context is part of that, but probably not all of it.

I'd still put Pedro above him, but there exists a case the other way I'd say. The case for Ryan over many of the others does not exist, unless it's all about bulk numbers
   63. Bad Fish Posted: December 16, 2008 at 03:52 AM (#3030151)
A quote I once read concerning Koufax from a prominent hitter of the time, he said other good pitchers you respected and maybe even were a little afraid of, but Koufax took your manhood from you.
   64. Eraser-X is emphatically dominating teh site!!! Posted: December 16, 2008 at 04:54 AM (#3030190)
Koufax took a little bit of manhood from prominent hitters and made Jim Rice.
   65. Howie Menckel Posted: December 16, 2008 at 04:59 AM (#3030194)
Koufax and Rice both took a lot of manhood out of opponents at home, and not only the fans but even the players expected it to happen on the road as well.

It kept not happening, especially in Rice's case, but each team just thought they got off lucky that day.
   66. ERROR---Jolly Old St. Nick Posted: December 16, 2008 at 12:06 PM (#3030272)
Koufax and Rice both took a lot of manhood out of opponents at home, and not only the fans but even the players expected it to happen on the road as well.

It kept not happening,


Jesus F. Christ, Howie. You'd think Koufax was stinking up the joint outside of Dodger Stadium. For his five peak years---which constitute everyone's reason for Koufax's status as an all time great---here are his ROAD ERAs:

1962 - 3.53
1963 - 2.31
1964 - 2.93
1965 - 2.72
1966 - 1.96

In 1962, he tried to come back from his midseason injury at the end of the year, and in three road starts he gave up 8 runs in 3.2 innings. Take that out of the equation and that 3.53 number goes down well below 3.00.

And again, small sample size, but in four ROAD appearances in the World Series, Koufax allowed 15 hits and 3 earned runs in 26 innings. Included in those road games was a three hit shutout in game 7 of the 1965 Series---on two days rest.

Koufax didn't have career durability, and it took him seven years to get started. Nobody's trying to put his overall value over the Madduxes and the Seavers. And for prolonged peak, Pedro was even better, especially considering the context of the era. But those five years alone, especially when you count the World Series, are more than enough to put Koufax way up on the list of all-time great pitchers. Those hitters back then knew damn well what they were seeing---or in most cases, not seeing. And find me any other pitcher back then who put up road numbers like that.
   67. PASTE, Now with Extra Pitch and Extra Stamina Posted: December 16, 2008 at 12:19 PM (#3030275)
1962 - 3.53 (3.59)
1963 - 2.31 (2.99)
1964 - 2.93 (3.25)
1965 - 2.72 (3.26)
1966 - 1.96 (3.28)

League average in parentheses.

That's some high quality yummy performance. But it ain't Pedro Martinez.

And find me any other pitcher back then who put up road numbers like that.


Marichal's road ERAs were generally better than Koufax's (with the caveat that Marichal's road games included Dodger Stadium and Koufax's did not). Jim Bunning's were in the same range. Bob Gibson didn't really get started until 1965, but his road ERAs were comparable for those last three years of Deadball Era II. That's about it.
   68. Jeff K. Posted: December 16, 2008 at 01:11 PM (#3030294)
Koufax took a little bit of manhood from prominent hitters and made Jim Rice.

Like how Cobra Commander made that evil guy out of Genghis Khan, Vlad the Impaler, Alexander the Great, etc?
   69. TomH Posted: December 16, 2008 at 02:06 PM (#3030311)
two, three and four are all very close. How the ESPN voters can screw up the top three spots so badly does not say much good for baseball fans and their judgement of what makes teams win; particularly analysis of context. Maybe they are punishing Clemens for steroid use?

1. Roger Clemens
2. Greg Maddux
3. Tom Seaver
4. Warren Spahn
5. Randy Johnson
X. Bob Gibson
6. Steve Carlton
7. Pedro Martinez
X. Bob Feller
X. Bert Blyleven
8. Gaylord Perry
X. Robin Roberts
9. Phil Niekro
   70. Jeff K. Posted: December 16, 2008 at 02:14 PM (#3030319)
FWIW, and I'd really rather not get into a huge debate over it, I gave Clemens a slight discount for steroids in #48. Without it, I'd flip him and Johnson. I'll just say this about it: This is a personal list, so I don't have any outside factors obliging me into higher standards (unless I'm really into the integrity of ESPN polls.) Unlike the HOF, or MVP voting, where I think the privilege of voting comes with those responsibilities, or writing an article/speaking on TV where the appeal to authority (by the fact that you're asked to speak/write on the subject) does the same, or obviously the court system. In any of those, I don't feel the standard of proof for any discount has been met for Clemens.

I say all that not to start the debate but hopefully to answer the questions about why/how for the discount that inevitably seem to lead to the debate.
   71. Barnaby Jones Posted: December 16, 2008 at 02:33 PM (#3030332)
17 people voted Glavine #1 on this list.
   72. ERROR---Jolly Old St. Nick Posted: December 16, 2008 at 03:11 PM (#3030374)
1962 - 3.53 (3.59)
1963 - 2.31 (2.99)
1964 - 2.93 (3.25)
1965 - 2.72 (3.26)
1966 - 1.96 (3.28)

League average in parentheses.

That's some high quality yummy performance. But it ain't Pedro Martinez.


Well, maybe that's why I also wrote this:

And for prolonged peak, Pedro was even better, especially considering the context of the era.


And take out those three games I mentioned from the end of 1962, when Koufax should've been on the DL, and that 1962 road ERA drops to 3.11.

As for Marichal and Gibson, both of them had better overall careers than Koufax, but neither of them had a five year stretch comparable to Koufax. And while Gibson is one of the all-time great World Series pictchers, his World Series ERA is twice as high as Koufax's. (Marichal had one inconsequential WS start.)
   73. Jeff K. Posted: December 16, 2008 at 03:36 PM (#3030415)
17 people voted Glavine #1 on this list.

Well, there's his agent, his family...frankly, I'm kind of surprised he doesn't have more than that. You'd figure there were a couple of hundred die-hard Glavine fanboys around Atlanta, much less the country.

Though it would be funny if his mom voted him 8th or something.

"Honey, I love you, but really. You're not better than Seaver or Carlton. Look, I put you above Nolan Ryan!"
"Mom, you're always talking about how he had one season above a 120 ERA+ and how overrated he is."
"I know honey, but it's the thought that counts."
   74. PASTE, Now with Extra Pitch and Extra Stamina Posted: December 16, 2008 at 03:41 PM (#3030422)
Glavine's enough of a student of the game that I doubt he rated himself higher than about #5.
   75. Chris Dial Posted: December 16, 2008 at 03:49 PM (#3030435)
I just happened to screwing around with this the other day.
Player    **IP**    PRAR
Roger Clemens    4916.2    973.8
Greg Maddux    5008.1    797.2
Randy Johnson    4039.1    743.8
Pedro Martinez    2782.2    622.9
Tom Seaver    4782.2    602.0
Bert Blyleven    4970    542.6
Warren Spahn    5243.2    538.6
Gaylord Perry    5350.1    533.3
Phil Niekro    5404.1    532.6
Tom Glavine    4413.1    519.1
Curt Schilling    3261    498.4
Steve Carlton    5217.1    497.2
Mike Mussina    3562.2    496.7
Bob Gibson    3884.1    496.4
John Smoltz    3395    489.0
Jim Palmer    3948    485.9
Kevin Brown    3256.1    472.8
Bob Feller    3827    470.6
Whitey Ford    3170.1    450.1
Fergie Jenkins    4500.2    443.6
Hal Newhouser    2993    439.2
Robin Roberts    4688.2    433.9
Nolan Ryan    5386    419.9
Juan Marichal    3507.1    397.7 
Now, I tend to make exceptions to this list, or tweak it based on other information, but that's a pretty good list.

PRAR is calculated by: =((ERA+/100*1.1*ERA)/9*IP)-ER

This is generated to give a quick and strong look at pitchers relative ranks from a readily available set of data and an easy calculation (from B-R, of course).
   76. Chris Dial Posted: December 16, 2008 at 03:51 PM (#3030439)
There are pitchers on that list that I never thought would meet an objective criterion for the HOF (and some of you heard me say those things). It appears I was wrong, and some pitchers I previously would shut out would likely meet my standards.
   77. OCF Posted: December 16, 2008 at 04:00 PM (#3030454)
Chris - it looks like that formula spits out a value in raw runs, not the win value of runs. In so doing, it systematically overrates pitchers from high scoring times and underrates those from low scoring times. (At least among good pitchers; it would undervalue bad pitchers in high scoring times.) Evidence of this issue: Mussina tied with Gibson doesn't pass a sanity test for me. Neither does R.J. ahead of Seaver.
   78. Jeff K. Posted: December 16, 2008 at 04:02 PM (#3030458)
You're dividing by 100 to take out the "presentation multiplier", why the multiply by 1.1?
   79. Chris Dial Posted: December 16, 2008 at 04:04 PM (#3030460)
Chris - it looks like that formula spits out a value in raw runs, not the win value of runs. In so doing, it systematically overrates pitchers from high scoring times and underrates those from low scoring times. (At least among good pitchers; it would undervalue bad pitchers in high scoring times.) Evidence of this issue: Mussina tied with Gibson doesn't pass a sanity test for me. Neither does R.J. ahead of Seaver.
Thanks. I broke it down by decade, and noted the bias, but just started this (because I *really* want a Q&G;"copyable" method from B-R) but haven't refined it. I really want to avoid too much complication. So, if there is a ready adjustment, I'm all ears.
   80. Chris Dial Posted: December 16, 2008 at 04:04 PM (#3030461)
You're dividing by 100 to take out the "presentation multiplier", why the multiply by 1.1?
Replacement level versus average.
   81. Jeff K. Posted: December 16, 2008 at 04:07 PM (#3030465)
Then, ?

Average for ERA+ is 100 (or thereabouts.) Replacement level isn't 90.9.
   82. Chris Dial Posted: December 16, 2008 at 04:07 PM (#3030466)
Thansk, OCF, maybe finding that "high run era" adjustment would get those pitchers off the list, and I wouldn't be wrong anymore. Ha!
   83. Chris Dial Posted: December 16, 2008 at 04:08 PM (#3030467)
Average for ERA+ is 100 (or thereabouts.) Replacement level isn't 90.9.
It's not a perfect system yet. 1.2?
   84. Jeff K. Posted: December 16, 2008 at 04:09 PM (#3030469)
And I'm sure there's a reason, but if you're already involving actual ERA and IP in the equation, why start with ERA+ and then try to multiply and divide half of it out? Why not just use PF on its own somehow?

(I realize you didn't develop PRAR)

(EDIT)(Or maybe you did.)
   85. Chris Dial Posted: December 16, 2008 at 04:11 PM (#3030471)
And I'm sure there's a reason, but if you're already involving actual ERA and IP in the equation, why start with ERA+ and then try to multiply and divide half of it out? Why not just use PF on its own somehow?
When you copy a chunk from B-R PF isn't there. I want to create something that anyone can do just by a Play Index search, and then calculate. Utility is nearly as important as accuracy here.
   86. Jeff K. Posted: December 16, 2008 at 04:12 PM (#3030472)
More like 1.4. That would equal a 71 ERA+.
   87. Jeff K. Posted: December 16, 2008 at 04:14 PM (#3030474)
Ah, is this related to our discussion in the Lounge Sunday?

FWIW, if they're dumping it into Excel and then calculating, PF is easy. You have the lgERA* right there, it's a one shot step to figure the PF.
   88. GGC Posted: December 16, 2008 at 04:17 PM (#3030477)
Q&G;= Q&D;?
   89. Chris Dial Posted: December 16, 2008 at 05:39 PM (#3030612)
FWIW, if they're dumping it into Excel and then calculating, PF is easy. You have the lgERA* right there, it's a one shot step to figure the PF.
*lgERA isn't on the Play Index. I checked that already. That's why I had to back calculate it.

Q & G means Quick and Good vs. Dirty.
   90. Chris Dial Posted: December 16, 2008 at 05:59 PM (#3030638)
Let's talk briiefly about runs per win. If I double league ERA, does that approximate runs per win? Or is it 2xlgERA plus 1?
   91. GGC Posted: December 16, 2008 at 06:21 PM (#3030680)
Here's a DIPS/replacement thing for giggles (1946 to 2007):

1. Clemens - 103
2. Ryan - 85
3. Unit - 85
4. Craig Calcaterra's Love Slave - 74
5. Blyleven - 69
6. Pedro - 66
7. Schilling - 63
8. Gaylord - 63
9. Seaver - 59
10. Carlton - 57
11. Kevin Brown - 54
12. Mussina - 53
13. Roberts - 53
14. Smoltz - 52


Some of you may know that I'm not DIPs biggest fan, but at least it does try to separate pitching from fielding. If I'm not mistaken, however (and you can correct me on this if you're still around, Voros), it doesn't account for wild pitches. But I think that this explains some of the love for Nolan Ryan in the general population.
   92. snapper (history's 42nd greatest monster) Posted: December 16, 2008 at 06:27 PM (#3030694)

Some of you may know that I'm not DIPs biggest fan, but at least it does try to separate pitching from fielding. If I'm not mistaken, however (and you can correct me on this if you're still around, Voros), it doesn't account for wild pitches. But I think that this explains some of the love for Nolan Ryan in the general population.


I don't thinks DIPS is very good for historical comparison.

Most of the increase in strikeouts over time is probably due to the hitters approach, not the pitchers skill.

In the old days, most hitters would severely cut down their swing with 2 strikes. There was a huge stigma to striking out.
Today, average hitters swing from the heels on every count.

If pitchers from 20+ years ago were inducing weak contact on 2 strike counts rather than putting the guy away, due to the hitters approach, I don't think we should penalize them, or reward the modern pitchers.
   93. GGC Posted: December 16, 2008 at 06:45 PM (#3030736)
That would actually be a plus for Ryan, snapper. I think that some Primates underrate him, but a top 2 ranking doesn't pass the smell test for me. THen again, the man pitched a lot of innigs and Voros posted what looks like a career list.
   94. Chris Dial Posted: December 16, 2008 at 07:46 PM (#3030852)
Let's talk briiefly about runs per win. If I double league ERA, does that approximate runs per win? Or is it 2xlgERA plus 1?
Okay, it is *lgERA/100*ERA*2+2. That gives Runs per Win. The flat PRAR (now being calculated at a 1.25 because Jeff made some suggestions) is divided by *lgR/W, yielding WAR.

This is a simple, one box calculation, from a B-R Play Index dump. And the answers are (top 50):
Player    **IP**    ERA+    PRAR    *lgR/W
Walter Johnson    5914.2    147    142.8    8.4
Roger Clemens    4916.2    143    122.6    10.9
Pete Alexander    5190    135    113.9    8.9
Greg Maddux    5008.1    132    110.7    10.3
Lefty Grove    3940.2    148    103.0    11.1
Christy Mathewson    4747    136    100.5    7.7
Tom Seaver    4782.2    127    96.2    9.3
Randy Johnson    4039.1    137    95.5    10.9
Gaylord Perry    5350.1    117    92.5    9.3
Warren Spahn    5243.2    118    92.3    9.3
Phil Niekro    5404.1    115    90.6    9.7
Bert Blyleven    4970    118    88.3    9.8
Steve Carlton    5217.1    115    87.1    9.4
Nolan Ryan    5386    111    81.2    9.1
Tom Glavine    4413.1    118    79.8    10.4
Eddie Plank    4495.2    122    79.6    7.7
Jim Palmer    3948    126    78.5    9.2
Bob Gibson    3884.1    127    78.3    9.4
Carl Hubbell    3590.1    130    76.3    9.7
Pedro Martinez    2782.2    154    75.8    11.0
Robin Roberts    4688.2    113    75.7    9.7
Ted Lyons    4161    118    75.7    10.7
Fergie Jenkins    4500.2    115    75.6    9.7
Eppa Rixey    4494.2    115    74.5    9.2
Don Sutton    5282.1    108    74.0    9.0
Red Faber    4086.2    119    73.4    9.5
Bob Feller    3827    122    72.9    9.9
Cy Young    3313    137    71.4    7.8
John Smoltz    3395    127    70.3    10.3
Tommy John    4710.1    110    70.0    9.3
Mike Mussina    3562.2    122    69.5    11.0
Whitey Ford    3170.1    133    69.1    9.3
Mordecai Brown    3172.1    138    68.6    7.7
Curt Schilling    3261    127    68.3    10.8
Ed Walsh    2964.1    146    67.8    7.3
Kevin Brown    3256.1    127    67.6    10.3
Juan Marichal    3507.1    123    66.5    9.1
Red Ruffing    4344    109    64.8    10.3
Jack Quinn    3920.1    114    64.1    9.5
Hal Newhouser    2993    130    64.0    10.0
Don Drysdale    3432    121    63.2    9.1
Early Wynn    4564    107    63.2    9.6
Jim Kaat    4530.1    107    62.3    9.4
Stan Coveleski    3082    127    62.2    9.3
Jim Bunning    3760.1    114    61.4    9.5
Billy Pierce    3306.2    119    59.9    9.8
Dazzy Vance    2966.2    125    59.5    10.1
Tommy Bridges    2826.1    126    58.5    11.0
Waite Hoyt    3762.1    111    58.4    10.0 
   95. Jeff K. Posted: December 16, 2008 at 07:47 PM (#3030854)
Let's talk briiefly about runs per win. If I double league ERA, does that approximate runs per win? Or is it 2xlgERA plus 1?


Have you tried to run a regression on it to see? Or hell, I'll look at something here.
   96. Chris Dial Posted: December 16, 2008 at 07:53 PM (#3030862)
My personal HOF probably falls between Bob Feller adn John Smoltz. Cy Young obviously had nmore data that this doesn't include, so he'd be well above the 71 mark he is at, and when a gap the size of 3 wins pops up, that's real seperation. Every gap after that is about half as much (there is one 1.7 WAR, and one 1.5 WAR). If you push the "HOF" out to Jim Bunning, who is 1.5, the next largest gap after that is 0.9 WAR. No one really seperating himself from his peers. this list is 485 players strong for consideration, drawn on a 105 ERA+ and 800 IP (in Play Index).
   97. Jeff K. Posted: December 16, 2008 at 08:25 PM (#3030901)
I'm being a ####### idiot, but I can't think. I now have a spreadsheet with each year's total # of wins per league (# of teams * 81) and earned runs (ERA/9 * Actual IP for that league) amongst all the other data. R/G, RA, G, ERA, IP, H (just because)

What function is it to show the best fit equation between runs and wins? Wins isn't variable, of course.
   98. Jeff K. Posted: December 16, 2008 at 08:47 PM (#3030934)
Nuts to me of 15 minutes ago. That guy was an idiot.

Regression where Earned Runs/Win is the dependent variable and Earned Runs is the independent variable:

NL, 1901-2008
<table>
Multiple R 0.741358474
R Square 0.549612386
Adjusted R Square 0.545363447
Standard Error 0.747356284
Observations 108</table>


df SS MS F Significance F
Regression 1 72.24891332 72.24891332 129.3528312 4.51939E-20
Residual 106 59.20538997 0.558541415
Total 107 131.4543033

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%
Intercept 4.948992801 0.198756643 24.8997604 4.30013E-46 4.554938435 5.343047167 4.554938435 5.343047167
X Variable 1 0.000356503 3.13455E-05 11.37333861 4.51939E-20 0.000294358 0.000418649 0.000294358 0.000418649

-------

AL, 1901-2008

Multiple R 0.825663689
R Square 0.681720528
Adjusted R Square 0.678717892
Standard Error 0.823465069
Observations 108

df SS MS F Significance F
Regression 1 153.9550613 153.9550613 227.040643 4.15393E-28
Residual 106 71.87804034 0.67809472
Total 107 225.8331017

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%
Intercept 4.422617153 0.224088103 19.73606406 2.77186E-37 3.978340707 4.8668936 3.978340707 4.8668936
Total ER 0.000505842 3.35709E-05 15.0678679 4.15393E-28 0.000439284 0.0005724 0.000439284 0.0005724
   99. Jeff K. Posted: December 16, 2008 at 09:05 PM (#3030960)
Hmm, now I'm wondering if I shouldn't normalize total runs for size of league to make sure that doesn't affect it. One sec...

Okay, new totals after normalizing for # of games. I did that by creating "New Total ER" = TotalER * (Max # of games in current league size / Actual GP) which fixes strike years as well.

NL 1901-2008

SUMMARY OUTPUT

Regression Statistics
Multiple R 0.930
R Square 0.864
Adjusted R Square 0.863
Standard Error 0.410
Observations 108.000

ANOVA
df SS MS F Significance F
Regression 1.000 113.607 113.607 674.762 0.000
Residual 106.000 17.847 0.168
Total 107.000 131.454

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%
Intercept -0.362 0.288 -1.255 0.212 -0.934 0.210 -0.934 0.210
New Toter 0.001 0.000 25.976 0.000 0.001 0.001 0.001 0.001

-------

AL 1901-2008

SUMMARY OUTPUT

Regression Statistics
Multiple R 0.884
R Square 0.781
Adjusted R Square 0.779
Standard Error 0.683
Observations 108.000

ANOVA
df SS MS F Significance F
Regression 1.000 176.338 176.338 377.653 0.000
Residual 106.000 49.495 0.467
Total 107.000 225.833

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0%
Intercept -0.394 0.416 -0.947 0.346 -1.218 0.430 -1.218 0.430
New Toter 0.001 0.000 19.433 0.000 0.001 0.001 0.001 0.001
   100. Chris Dial Posted: December 17, 2008 at 12:40 AM (#3031189)
Sure but what does any of that gobbledygook mean?
Page 1 of 2 pages  1 2 > 

You must be Registered and Logged In to post comments.

 

 

<< Back to main

News

All News | Prime News

Old-School Newsstand


BBTF Partner

Dynasty League Baseball

Support BBTF

donate

Thanks to
greenback slays lewks
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogThe moral argument for keeping Barry Bonds out of Cooperstown doesn’t hold up
(12 - 5:55am, Jul 20)
Last: MuttsIdolCochrane

Newsblog2019 Trade Value:
(7 - 5:41am, Jul 20)
Last: bookbook

NewsblogDeadspin: Baseball Writer Jonah Keri Arrested, Charged With Assault On His Wife
(43 - 2:45am, Jul 20)
Last: the Hugh Jorgan returns

NewsblogHey, $20 bucks is $20 bucks to OMNICHATTER! for July 18, 2019
(92 - 1:34am, Jul 20)
Last: Howie Menckel

NewsblogOT - Catch-All Pop Culture Extravaganza (July 2019)
(611 - 12:54am, Jul 20)
Last: Howie Menckel

NewsblogThe Five Trends That Could Define Baseball’s Future - The Ringer
(3 - 12:47am, Jul 20)
Last: John Northey

NewsblogPrice responds as Eckersley dispute resurfaces
(12 - 10:34pm, Jul 19)
Last: Slivers of Maranville descends into chaos (SdeB)

NewsblogOT - NBA thread (Playoffs through off-season)
(6213 - 10:24pm, Jul 19)
Last: PASTE, Now with Extra Pitch and Extra Stamina

NewsblogOT Soccer Thread - Baldrick Reports Live
(1191 - 9:20pm, Jul 19)
Last: It was something about the man-spider and sodomy,

NewsblogThe Transformation of Alex Rodriguez (SI)
(55 - 8:26pm, Jul 19)
Last: ERROR---Jolly Old St. Nick

NewsblogPrimer Dugout (and link of the day) 7-19-2019
(23 - 7:40pm, Jul 19)
Last: The Run Fairy

NewsblogOT - College Football Bowl Spectacular (December 2016 - January 2017)
(412 - 7:24pm, Jul 19)
Last: shout-out to 57i66135; that shit's working now

NewsblogWhy gambling used to scare baseball and why it doesn’t anymore
(98 - 6:43pm, Jul 19)
Last: Howie Menckel

NewsblogLet’s make a deal? Assessing the Mariners’ likely trade chips as deadline looms | The Seattle Times
(1 - 6:21pm, Jul 19)
Last: Walt Davis

Gonfalon CubsThat was fun
(427 - 4:01pm, Jul 19)
Last: Moses Taylor, aka Hambone Fakenameington

Page rendered in 0.6912 seconds
46 querie(s) executed