hot topics

The two ties make for an interesting contrast. Niekro and Perry are very similar pitchers, but Saberhagen and Sutton are complete opposites. I guess their tie suggests that the peak and career votes were pretty evenly balanced in this election. It doesn't look to me like we have ended up with any surprising or counterintuitive placements here.

I knew going into this election that Seaver was the #1 pitcher from this era, but I hadn't realized before just how much better he was, overall, than anyone else.

Well, sheesh. I called this race before the first ballot had been cast. A simple click to B-Ref could have saved everyone a lot of time and trouble! ;-)

Biggest individual spreads: Wilhelm 3 - 18, Koufax 4 - 18, and Drysdale 6 - 20. And it still seems to me that one's ranking order is much more about "What constitutes merit?" than anything else.

Yeah, Sam, I don't really think any of the voters spent more than about 3 seconds deciding who to put #1 on their ballots. Now, if we go cross-era and compare Seaver to Johnson, Grove, and Clemens - then it's going to take more than 3 seconds.

Here are the results with point converted to average ranking. That's the first number after each name. The second number is the standard deviation of that placement, so you can see where our biggest disagreements lie.



And here are the consensus scores:

86 Chris Cobb
85 TomH
83 OCF
83 Esteban Rivera
82 Brent
81 bjhanke
80 Tiboreau
80 Rob Wood
79 Joe Dimino
79 AJM
78 Sunnyday2
78 DL from MN
78 Howie Menckel
78 Mark Donelson
78 Andy
76 EricC
75 karlmagnus
70 Sean Gilman
61 John Murphy

Andy: note that Sutton had an even larger standard deviation than any of the cases you cited - there was quite a profound disagreement about him, from seeing him as "Ryan without the no-hitters" (which is basically my view) to seeing him not deserving HoM election at all.

Thanks, OCF. My last post was just based on the gap between a pitcher's best and worst rankings, but I'm glad you've given us something a bit more elaborate than that.

I guess my biggest odd man out ranking is of Eckersley, where my views about the relatively low value of closers (other than Rivera and maybe Hoffman) and my view of Eck as a mediocre starter don't add up to anything special. If he'd taken his peak as a closer and stretched it out for a dozen or more years, then I'd see him in a much different light, or if he'd been as relatively effective as a starter for the same number of years as he was a closer, then he'd be Koufax and then some. But again, all this is based much more on how I measure "merit" than any great difference in how I see Eckersley's talent. We all agree that he was a so-so starter and a fabulous closer for a brief amount of time; we just don't agree on what all that amounts to. Same with Koufax or Palmer---it's not that we differ on what they did, only on what it says.

Most of these pitchers are in the Hall of Fame as well, and when it comes to Saberhagen and Stieb, our election can be seen mostly as a suggestion that maybe the HoF has been a bit too stingy overall. But Blyleven stands out in stark relief - not in the Hall of Fame (yet), and we have him at 6th out of 20, well ahead of a whole lot of household names and no-doubt Hall of Famers.

One of the things cited by those who attempt to make a case against Blyleven for the Hall of Fame is that he "didn't know how to win," that his actual W-L record should be taken into account. If you check the comments, I think you'll see that several voters explicitly said they were downrating Blyleven for exactly that - and yet he still finished that high in our rankings. (Without that, he might have been even higher.)

So the two pitchers with the highest SD are Koufax and Sutton: Mr. Peak and Mr. Career. The next highest SD scores are Wilhelm's and Eckersley's, suggesting that there is much disagreement about the relative value of the best relief pitchers.

Following up on Jolly Old St. Nick's comment about the question of "what constitutes merit?" as being the biggest factor in most disagreements, I'd agree completely with respect to the source of the biggest disagreements--peak vs. career. With the relief pitchers, I'd say that uncertainty about how to calculate the value of a relief pitcher relative to a starting pitcher remains high. Given that uncertainty, voters' choices become more influenced by subjective views of merit. I think that eventually there will be greater agreement about the questions of how to value what relief pitchers do, that may lead to a reduction in the extent of disagreement about their placements, though relief pitchers' usage patterns will probably mean that they will always fall fairly strongly to one side or the other of the peak vs. career debate.

I look forward to the "finals" - ranking a top-15 ballot of all-time pitchers.

The 4-part effort has finally made it digestible...

First off, we have to revote this. I never end up in the top third of consensus scores.

Second, OCF says, "Yeah, Sam, I don't really think any of the voters spent more than about 3 seconds deciding who to put #1 on their ballots. "

I spent over 5 hours doing Seaver vs. Carlton vs. Gibson. Admittedly, I never ended up with Seaver anywhere except first, but trying to balance his longevity against Gibson's little advantages and Carlton's monster season was more of a pain than I thought it could possibly be. I mean, you try to figure out how many career innings Bob's World Series performance advantage amounts to. The big thing that settled the matter for me was extracting Gibson's number of innings out of the best consecutive part of Seaver's career. I found out that if I do that, Seaver's Gibson-length OPS+ only goes up by 1 point, but the OPS+ for the extra innings only goes down to 123. Knowing that the bonus innings were still very very good was what stopped me from having any doubt. If those 900 extra IP had been junk, I might well have had Tom as low as third.

I think that Chris has the parameters of this discussion down pretty well, but I am coming around to the view that closer is a very serious job with serious consequences equal to an ace starter. The reason is watching what happens to teams whose closers fall apart for a year. Not just in St. Louis, where it has been obvious, but in general. Teams whose closers collapse really do seem to suffer just as much as teams whose ace starters collapse. That leverage seems to be worth a lot. True, some teams don't have many problems, but teams also win in spite of ace starter collapse, as well. One reason that closers seem to have so much impact seems to be the number of games they get into. They're in between starters and position players in that regard, and the games they do get into are the ones their team is winning late. Also, if a closer faces 4 batters, he has had the same impact, from at least one perspective, as a hitter's 4 PAs do.

- Brock

Oh, yeah. I promised you guys a Steve Carlton story, which has nothing to do with ranking him. Have I already posted one about the influence of Richard Nixon on Steve's trade to the Phillies? If so, let me know. If not, I have a wowzer for you, vetted by Bing Devine, who was the GM who had to make the trade against his wishes. I was not willing to repeat this one until Bing died, because there is some flame in there, but Bing is no longer with us. - Brock again

First of all, as a Met fan, I'm glad that Seaver was a unanimous choice.

As for my consensus score, I was at the top for the last group's election. Easy come, easy go. :-)

I look forward to the "finals" - ranking a top-15 ballot of all-time pitchers.

The 4-part effort has finally made it digestible...

That's still going to be a nightmare, IMO.

As to where we go next, we'll figure that out after Joe comes back from his honeymoon.

Nice results. I wrote up most of a ballot earlier in the week, then never got back to posting it. Procastination is bad! One nice thing: I could come up with no justification for picking Niekro over Perry (or vice versa) in slots 4 and 5, so I'm rather pleased that I wouldn't have helped break a tie.

That's still going to be a nightmare, IMO.

For me, the top 15 will be fairly easy--the very top pitchers are spread out, so there are not too many doubtful cases, except insofar as period adjustments are called for. After that, things get murky fast. In my system, the #1 pitcher of all time, Walter Johnson, scores at just about twice what the #15 (strictly by numbers) all time scores. Johnson is at 340, while Bob Gibson is at 171. The next 15 pitchers are all within 20 points of one another, and the next 15 are even more closely packed.

If we had 200 ballots, we could assume that everybody's small errors in ranking a pitcher #20 instead of #40 would cancel out, so you'd get a finely grained calibration of where each pitcher was in the overall standings. As it is, with only 20 voters the random errors will still dominate at the middle. The top will be mildly interesting and not quite obvious, and the bottom 5 may well stand out as HOM errors or quasi-errors -- if everybody agrees that Joe Pitcher was 59th or 60th best, with a gap from the others, he's an error.

top 9 is easy IMHO. After that it's hard.

Johnson
........(Clemens)
Grove
Young
........(Maddux)
Seaver
Alexander
Spahn
Mathewson
Paige
SJ Williams

G Perry
Gibson
Nichols
Carlton

Roberts/Feller/Blyleven/Niekro

For me, the top 15 will be fairly easy

I wish I were that confident. I can't come up with but 11 no-brainers. In chronological order:

Cy Young
Matty
Walter Johnson
Alex
Grove
Spahn
Seaver
Clemens
Maddux
Randy Johnson
Pedro

After that I hopelessly bogged down in conflicting thoughts about peak vs career, era adjustments, and the many definitions of "merit" and "greatness." Maybe when I learn more about the 19th century stars I'll be able to add some more.

EDIT: And of course there's Paige and probably Williams, which would leave but two.

Here's my swag

W Johnson
(Clemens)
Grove
Young
Alexander
Williams
Seaver
(Maddux)
Spahn
(R Johnson)
Mathewson
(P Martinez)
Paige
Feller
Gibson
Nichols/Niekro/Carlton
Blyleven/Hubbell/Perry/Dihigo/Vance
Clarkson

The hard part for me is determining whether Clarkson deserves to squeeze in with Carlton and Niekro. That and there's still the possibility that Pedro throws one more good season and passes Matty. I suppose Randy Johnson could pitch through 2010 and pass Spahn also. The most uncertainty in the top 10 is probably around Joe Williams' placement.

I did not include the active guys, Pedro and Big Unit.

Quick rankings:

Johnson
(Clemens)
(Maddux)
Young
Alexander
Williams
Spahn
Grove
(Johnson)
Mathewson
Roberts
Carlton
Perry
Niekro
Brown
(Pedro)
(Galvine)
Gibson
Blyleven
Nichols

not sure why this one would be so hard

right now I have 19 pitchers to look at, for 15 slots, as opposed to say 20 for 20 before
(or we could do 20 slots if we prefer, then I'd have about 27 to look at)

plus in this case we've already ranked our subsets in order, it's just a matter of merging them for many/most voters

Grove below Spahn doesn't make sense.

not sure why this one would be so hard

Well, it's easy if you have no plans to look at any non-Young/Nichols 19th century pitchers, Howie. ;-)

Grove below Spahn doesn't make sense.

I'd rate Grove higher myself because his rate stats were so much better, and because I'd have to give a certain amount of credit for those years he was trapped in Baltimore through no fault of his own. But unless you're saying that people can't take the relative levels of competition as a factor at all, or Spahn's 1300 extra Major League innings, I can see how someone might see Spahn's value as being greater. There are so many different ways of approaching this that there's no one perspective that necessarily trumps all the others.

No, Clarkson to me also clearly deserves consideration

Keefe? Not in an all-time top 15, imo

We already ranked Grove unanimously ahead of Spahn. What changed?

No, Clarkson to me also clearly deserves consideration

Keefe? Not in an all-time top 15, imo

I agree, Howie. It might not be as hard as I first thought. :-)

My top pitchers ballot deliberations are at about this point:

1) Walter Johnson
2) Cy Young
3) Pete Alexander

No period adjustment can be large enough to put any eligible modern pitcher into this group. Clemens and maybe Maddux will break into this group when their eligibility arrives.

4-10 Some order of Christy Mathewson, Satchel Paige, Tom Seaver, Lefty Grove, Joe Williams, Warren Spahn, and Kid Nichols

I'm not sure of the order of this group--the above is my current thinking--but I know no one else has a chance to break into the Top 10.

11-15

Definitely In -- Steve Carlton, Bob Gibson
Under consideration -- Phil Niekro, Gaylord Perry, Robin Roberts, John Clarkson, Carl Hubbell, Bob Feller, Eddie Plank

My preliminary top 15, now in order. Since IMO actually trying to rank these 15 with any great precision is a fool's errand, I'll only say that I purposely tried to scatter representatives from each era among each grouping, even though that likely slights the modern era (meaning today's pitchers) somewhat. Greg Maddux with a spitball and a emery board might well have won 600 games in Cy Young's era.

When I start reading the arguments for some of the other 19th century pitchers, I may very well revise the list. And hell, I may change my mind on the order 15 minutes from now.

1 - Clemens - Yes, steroids, and he won't make my HoF because of that, but this is different. And the competition level he faced puts him at the top.
2 - Grove - In a variant of what I said about Clemens, I can't put Grove below any dead ball era pitcher.
3 - Walter Johnson - Nolan Ryan with control
4 - Williams - Might well be # 2, or even # 1, if only....
5 - Randy Johnson - Not as consistent as Clemens, but he combined a Koufax-like mountaintop with a lot more extras
6 - Cy Young - Hard to rate him, but I can't ignore 511 wins, 7092 innings, and a very respectable 136 ERA+
7 - Seaver - Nobody needs to spell out Seaver's greatness
8 - Paige - See comment about Williams
9 - Alex - below Johnson and Young, but above Matty. Matched Mathewson's wins and rate stats, but gets extra credit for doing it in part in the lively ball era and also for doing it with lesser teams.
10- Maddux - Probably the best pure pitcher ever, considering his era. He treated the entire National League the way Hub Pruett treated Babe Ruth.
11- Spahn - Only mark against him is his lack of a record against his team's main opponent (Brooklyn)
12- Pedro - Only longevity prevents him from being # 1 on this list. The best pitcher I ever saw.
13- Matty - See comments about Alexander
14- Gibson - Gets knocked down slightly due to era adjustment and lack of all that many truly dominant seasons. But for a single game I'd take him over anyone on this list with the possible exception of the guy below him.
15- Koufax - Yeah, I know, but no list of "greatest pitchers" is complete without Sandy Koufax. And I never rooted for him in a single game in my life, so forget the "fanboy" stuff.

Andy - if this is what we do next (I suppose it's still not fully decided), we'll probably limit it to pitchers that have already been elected to the HoM. If that's the case, you'll have to pull Clemens, R. Johnson, Maddux, and Pedro from your list and find four more to replace them with.

# 21 -- You left out a certain vintner from Fresno, AJMac. Might want to check out the top of this thread . . . . ;-)

Andy - if this is what we do next (I suppose it's still not fully decided), we'll probably limit it to pitchers that have already been elected to the HoM. If that's the case, you'll have to pull Clemens, R. Johnson, Maddux, and Pedro from your list and find four more to replace them with.

Ouch! But rules are rules, and thanks for clueing me in. It now hits me why others put the non-HoM names in ( ).

I assume "anything goes" in the top 15. I mean, I rank my guys anyway I want, right? IOW the results from the four time periods don't obligate anybody to do anything, right? So I don't see how the four time periods help us to digest anything, not that I object to voting for the four time periods, too. But as for Clemens, et al, I sure don't want them in the mix. I'm with John. This is still gonna be hard.

First pass. This is going to change. This is not easy.

1. Walter Johnson
2. Lefty Grove
3. Cy Young
4. Tom Seaver
5. Pete Alexander
6. Steve Carlton or Bob Gibson
7. Kid Nichols
8. Gibson or Carlton
9. Christy Mathewson
10. Sandy Koufax

11. John Clarkson
12. Warren Spahn
13. Hoss Radbourne
14. Satchel Paige
15. Joe Williams

16. Carl Hubbell
17. Ed Walsh
18. Bob Feller
19. Gaylord Perry
20. Amos Rusie

6. Steve Carlton or Bob Gibson
7. Kid Nichols
8. Gibson or Carlton

This seems very odd. You're sure that exactly one of Carlton or Gibson is better than Nichols, you're just not sure which one?

wow, Chris C, I suspect you'll catch some flack for Lefty Grove being no-better-than-4th. I look forward to your defense tho, even if it's me handing out the flack :)

Hm. I think the only way to argue that Grove is any better than #4 is to argue that pitchers who pitched pre-1920 don't really count.

Let's look at some very basic stats and a complex one

Walter Johnson, 5914.7 IP, 147 ERA+, 169.3 WARP3
Cy Young, 7354.7 IP, 138 ERA+, 135.5 WARP3
Pete Alexander, 5190 IP, 135 ERA+, 132.2 WARP3

Lefty Grove, 3940.7 IP, 148 ERA+, 110.6 WARP3

For the statistically minded, I don't see any evidence that could possibly justify Grove ahead of Johnson, who has 2000 more IP at the same level of effectiveness. In addition (if anything more were needed), Johnson was an excellent hitter while Grove was a historically terrible one.

Young has 3400 innings on Grove, and his effectiveness was not much less than Grove's. How do you write off 3400 IP at 138 ERA+? That's a Hall-of-Fame career right there, in any era.

Alexander has the narrowest edge, "only" 1200 IP (4 years as a top starter), and his career ERA+ was only 13 points less than Grove's. If the two ERA+ totals are compared in a normalized 4.5 r/g environment, Alexander saves 673 runs above average in his career, Grove 639, and in runs saved above replacement Alexander will be far ahead. And Alexander pitched the second half of his career in the lively ball era, and he deserves some war credit. Grove probably deserves some MiL credit, but it would bring his career ERA+ down quite a bit. He had an 98 ERA+ in his rookie season, remember, leading the league with 131 walks. Now, I haven't looked at his minor-league stats, so I don't want to claim too much about how the MiL credit case is overblown, but it's clear there was a significant learning curve for Lefty once he reached the majors.

WARP3 makes an effort to do all the adjusting for competition, etc. that one needs to do to put players into an all-time context, and it doesn't see Grove as close to the other three. I am not saying that it's right, but I am saying that, given that WARP3 is doing a lot of adjusting for context, you'd have to do a whole lot more adjusting for context to put Grove ahead. I see three ways you could do that:

1) Saying that Grove was the best pitcher of his era by a lot and so he has to be better than pitchers who weren't the best pitchers of their era, regardless of what the stats show;
2) Going with your gut and concluding that there's _no way_ a deadball era pitcher could rank ahead of Grove, regardless of what the numbers say (as Jolly Old St. Nick does, I infer);
3) Making a peak argument for Grove, which I haven't attempted to address here. But Johnson, Young, and Alexander have pretty good peaks, too.

Anyway, it appears to me that the numbers are pretty solidly on my side. You can give me flack for going with the numbers, if you like, but I think I can take it. And if you show me a better way to interpret the numbers, that's all to the good!

I'll likely argue a version of #1; ease of domination argument, which is why I like Seaver as much as Maddux, even if the WARPy/WinSharey/ERA+ ##s don't agree. Ya know, that nine ERA titles thing, do we really think the 3 best pitchers existed pre-1925, etc.

Kid Nichols was as good as Cy thru 1900ish. Cy is more like a small uprgade to Warren Spahn, with a tremendous career.

Walter Johnson, once the live ball came, outside of his huge home stadium was a somewhat-above-average pitcher. His road stats 1920ff are merely nice. I'll still take him first, but it's close.

Pete Alexander - underrated by the masses 'tis true.

More substantial arguments, maybe some other day.

> Lefty Grove, 3940.7 IP

That doesn't include the _five_ years in Baltimore. That takes care of the innings. Take a look at his minor league stats, they're very impressive.

As for Young, he pitched in an era where starting 50+ games was common and led the league once with 43. Grove led the league once also, but with 37 starts. I'm not going to say what Young did was less impressive but it was more commonplace to start that many games. Young also completed a much higher % of games - they rarely relieved pitchers then. Yes he pitched those innings - approximately 2000 more - but his opportunity to pitch that much was better.

"Making a peak argument for Grove, which I haven't attempted to address here. But Johnson, Young, and Alexander have pretty good peaks, too."

Young's peak doesn't rate with Grove's, even just going relative to league and no timelining.

I could wind up with Young over Grove as well, but Cy would have to leap that hurdle first.

I also like Smokey Joe a tiny bit better than GC-Alex.
I guess he'll be underrated by some. I like him better than the great Paige as well.

DL, exactamente.

Tom H, bingo.

Grove's innings as compared to Johnson, Young, et al, have everything to do with the environment he pitched in. But even then, add in some recognition for his getting stuck in Bollimer and the innings practically go away anyway.

From Joe's research:

Lefty Grove's numbers include his time with Jack Dunn's Orioles; Pete Alexander's include WWI credit.

Lefty Grove, 3940.7 IP

That doesn't include the _five_ years in Baltimore. That takes care of the innings. Take a look at his minor league stats, they're very impressive.

Well, I've looked at what BB-ref has, and they're good, but Grove's walk rates were astronomical--way above what any effective pitcher in the majors was putting up--and his H/9 were way below what he ever achieved in the majors. I'd need to see a more thorough study to accept that Grove would project to an above-average pitcher during those seasons. Maybe he was better then than he was in 1925, and maybe he wasn't.

But if you give him 1922-24 as major-league equivalents, cut his IP by 10%, and assume a league average ERA+ for the 3 seasons, Grove's career line becomes

4613.7 IP, 139 ERA+

That might be enough to lock him into 4th place in my rankings or even challenge Pete Alexander, but it's still a ways up to Young and Johnson.

Incidentally, if you map Young's career onto Grove's career, starting him in the majors in 1922, ending in 1923 and giving him exactly the same number of IP per season as the AL pitcher who had the same ranking among IP leaders as Young did for the corresponding season of his career (he was in the top 10 in IP 19 consecutive seasons), he ends up with 5487.3 IP, at what would still be a 138 ERA+.

Unless Grove would have been a well-above average pitcher in the majors during the 1922-24 seasons, minor-league credit won't give him the bulk he needs without costing him his edge in career effectiveness. To be convinced that he really was that good in those years, I would need to see the analysis.

I agree with Howie that Young never had a five- or six-year stretch as good as Grove's 1928-33, which is one the five or so top pitching peaks of all time. Young's peak is merely excellent in height. It is remarkable in that it is about 14 years long.

Grove below Spahn doesn't make sense.

You left out a certain vintner from Fresno, AJMac. Might want to check out the top of this thread

Hey, what part of "quick rankings" don't you people understand! ;-)

Shoot, I wasn't giving Alexander WWI credit. Thanks for pointing that out.

As with most discussions with Chris Cobb I end up learning more than he does. Alexander/Grove/Young is a complicated mess for 2/3/4.

Walter Johnson, once the live ball came, outside of his huge home stadium was a somewhat-above-average pitcher.

You've listed two enormous factors** right there which make it very hard for me to rank him at the top of my list---even though I still rank him # 3 below (Clemens) and Grove. And yet it's still tough to imagine that "Nolan Ryan with control" wouldn't dominate in any era he played in.

**a third factor would be the added advantage of being a fast ball pitcher in an era of day games only, when you had many games played with tobacco stained baseballs in late afternoon twilight, making it virtually impossible to gauge a pitcher like Johnson. And he wouldn't have had those advantages today.

Walter Johnson

career home IP home ER home ERA
........ 3050 ......... 660 ....... 1.95
career .road IP . road ER road ERA
........ 2874 ......... 768 ....... 2.41

pre-1920

home IP home ER home ERA
.. 2054 .... 349 ....... 1.53
road IP . road ER road ERA
.. 2046 .... 404 ....... 1.78
2.81

1920ff

home IP home ER home ERA
... 996 ..... 311 ...... 2.81
road IP . road ER road ERA
... 828 ..... 364 ...... 3.96

I can see why some people would care, but I think it is completely legitimate to ignore home/road splits. If someone can leverage their home park to improve their performance by more than a simple park factor would indicate, that counts for real wins, and they should be rewarded for doing so. Of course, these are very well-trodden grounds around here...

Hm. I think the only way to argue that Grove is any better than #4 is to argue that pitchers who pitched pre-1920 don't really count.

I think the only way to drop Grove below Young and Alexander no questions asked is to completely ignore the fact that the era in which Grove pitched was much harder for pitchers to accumulate starts and IP than the eras before 1920.

I'd probably put Young ahead of Grove, but not Alexander

TomH: If you do those splits in R rather than ER, does the difference hold at about the same magnitude? Or it is a little larger or a little smaller?

The analysis of Johnson pre-1920 vs. post-1920 in order to make a case that deadball era stats should be massively discounted is not sensible, for two big reasons.

First and foremost, Johnson had an arm injury in 1920, and he was simply not the same pitcher before as he was after. It's an interesting coincidence that it happened in 1920, but he was 32 at the time, an age at which a lot of pitchers carrying deadball era workloads burned out. Mathewson had his last good year at 32, Walsh at 31, Waddell at 32. Johnson was at a point in his career when we should expect to see him decline.

Re home-road splits: if the league average ERA for Johnson is park-adjusted, why should it matter what his home-road splits in raw ERA were?

I think the only way to drop Grove below Young and Alexander no questions asked is to completely ignore the fact that the era in which Grove pitched was much harder for pitchers to accumulate starts and IP than the eras before 1920.

It was harder to accumulate starts and IP after 1920 than before 1920, but it wasn't _much_ harder. People often exaggerate the changes in seasonal IP across this divide. In the AL of 1908-12, the starter placing #5 in the IP rankings averaged 290.3 IP. In the AL of 1925-29, the starter placing #5 in the IP rankings averaged 259.7 IP, a decline of 10.5%.

Earlier I made two rough adjustments to find out (1) the approximate number of innings a pitcher with Young's durability profile would have been able to throw if he had broken into the majors in the season that Grove's MLE credit seems likely to begin--1922--and to see what Grove's career IP/ERA+ line would have been like with his 1922-24 seasons added to it, at 90% of minor league innings and a league average ERA+. I made no adjustment to Young's career ERA+.

Young's line was 5487.3 IP, 138 ERA+
Grove's line was 4613.7 IP, 139 ERA+

I've now performed the same exercise on Pete Alexander that I performed on Young, translating his 1911 NL rookie season into the 1922 AL, and so on, leaving ERA+ unchanged.

Old Pete's line was 4695 IP, 135 ERA+

Very close to Grove's adjusted line indeed! However, that does not include a season's worth of peak years war credit for Alexander.

In reviewing Alexander's record, I also noted a couple of interesting facts I had not previously considered.

1) His rookie season, which was possibly the greatest rookie pitching performance ever, came at age 24. He got his break in the majors only one life-year earlier than Grove. His minor-league line for 1910 in the New York State League was 29-11, 1.85 ERA, 345 IP, 6.6 H/9, 1.9 BB/9.

2) He has a good argument to have had a better peak than Grove, relative to his league. Here's Grove's IP/ERA+ placements during each season of his 1928-33 peak:

1928 6/2
1929 3/1
1930 4/1
1931 2/1
1932 2/1
1933 3/3

Here's Alexander's, 1914-20, skipping 1918, b/c Alexander was in the army

1914 1/7
1915 1/1
1916 1/1
1917 1/2
1919 x/1 (missed the first part of the season due to military service, hence the lower IP)
1920 1/1

That's not half bad. So I'll agree that giving Grove minor-league credit to boost his innings and translating Alexander into Grove's context narrows the career gap between them significantly. But I think war credit and consideration of peak probably give the edge to Alexander, though a more fine-grained assessment of Grove's minor league performance than I have done might bolster Grove's case.

Just to complete the study, here are Johnson's translated numbers, added to set:

Young 5487.3 IP, 138 ERA+
Johnson 5198.7 IP, 147 ERA+
Alexander 4695 IP, 135 ERA+
Grove 4613.7 IP, 139 ERA+

Johnson's 6-yr consecutive peak, in IP/ERA+ rankings by season:

1910 1/2
1911 3/2
1912 2/1
1913 1/1
1914 1/2
1915 1/1

He misses Alexander's three 1/1 sweeps, but he's first or second in every ranking but one.

Johnson started in the majors at age 19, so I don't think there's a need to consider minor-league credit for him!

Back when Grove became eligible, I calculated rough MLEs for him. One thing I found is that Baltimore played in a fairly extreme hitters' park, so his actual minor league ERA statistics aren't too different from his major league projections for those seasons. I calculated his MLE ERAs (expressed in ERA+) as:

1920 89
1921 148
1922 120
1923 119
1924 121

Adding these seasons would add bulk to his career, but would also hurt his career rate statistics somewhat.

re: 52 & 53: I do not have access to R allowed, OCF, only ER allowed. Courtesy of Bill Deane. CC asks a good Q about park effects; maybe someone with time could account for the Senators' field and how that affected Walter's stats. I simply used UNadjusted ##s because I thin the general perception of Johnson (utterly awesome in his 20s, very good in his 30s) is affected by lack of understanding how the big park helped him once the ball became livelier. Somewhat like the general perception of Koufax that we overall have dismissed.

Here's one mini-fact about Griffith Stadium in the early lively ball era: In 1924, the year the Senators won their only World Series, they hit exactly one home run in their own home park.

I don't wish to diminish Walter J, either, I really believe he WAS the best pitcher ever. In fact, his dominance over even other great pitchers like Pete A shows to me that it's a challenge for me to rank Alexander as one of the top five pitchers all-time. They were born the same year, and if you line up their careers by either Win Shares or WARP3, Walter beats him by 1.5 to 2 wins Per Year. The reason for Alexander's dominance on the league leader board (see post 54) is that Johnson was in the other league.. doing a lot better! (to be fair, some of Johnson's value is in his hitting, which will not show up in IP/ERA+).

With respect to Alexander, another consideration, IMO, is league quality. There are three periods since 1900 when I think there is compelling evidence from several sources of a gap in quality between the two leagues: the 1910s, when the AL had all the first tier stars--Cobb, Johnson, Collins, Speaker, and later, Ruth--except for Alexander (and also had the advantage in second tier stars like Jackson, Baker, etc.); the late 1950s and 1960s, when faster integration by the NL allowed them to dominate the AL; and the last few years [since when? 2002/3/4?].

The 1910s league quality gap is not just an artifact of modern sabermetrics. It was recognized and discussed in the baseball guides of the time. I recall that one of the guides kept track of all the inter-league games. (Besides the World Series, there were various inter-league games and series, such as "city series" where teams from the same city--like the White Sox and Cubs--challenged each other.) My recollection is that for a period of several years, the AL dominated not only the World Series, but also the non-World Series interleague games.

Consequently, in making comparisons I shave a bit off Alexander's raw statistics.

Brent, do you know anything about whether the AL was superior in both hitting and pitching? On the position player side, it's clear the AL was the stronger league, but I've never seen commentary addressed directly to pitching strength. Alexander's raw stats, would, as you note, be affected by the paucity of superstar hitters in his league, but would his leaderboard placements and ERA+ have benefited from weaker competition on the pitching side, or not? Anyone know about that?

Chris asks, "Brent, do you know anything about whether the AL was superior in both hitting and pitching? On the position player side, it's clear the AL was the stronger league, but I've never seen commentary addressed directly to pitching strength. Alexander's raw stats, would, as you note, be affected by the paucity of superstar hitters in his league, but would his leaderboard placements and ERA+ have benefited from weaker competition on the pitching side, or not? Anyone know about that?"

I have studied this issue a little. Someone who has studied it in more depth may know more. As far as I was able to tell, the AL, when it went for major league status, did more than just raid existing major league teams for players. The AL started recruiting in the deep South, which had been almost unmined until that time. Cobb, Speaker and Jackson are from that area, as are pitchers like Eddie Cicotte. There do seem to be more hitters, and more superstar hitters, than pitchers, but yes, there are pitchers involved. For a decade or so, that put the AL ahead of the NL. Again, I don't have comprehensive lists of who came from where, but I have enough southerners in the AL and so few in the NL that it has to have had some serious effect. The first crop of southerners, like almost all groups coming from new recruiting territories (like the Negro Leagues), tended to be stars and better, so they had a disproportionate impact on league difference. When the NL got around to recruiting people like Hornsby from Texas, that AL advantage started to go away. - Brock

Chris, No, I don't have any specific evidence on hitting versus pitching. I agree with Brock that the difference seems to have been concentrated in position players, but the AL also seems to have had some advantage in pitchers.

The obvious way to measure pitchers separately from position players would be to look at players who switched leagues, but as you know, that's an extremely thin sample during that period.

Brock and Brent,

Thanks! I imagine we'll be looking into this issue some more if/when the combined pitcher ranking project goes forward. . . .

Relatedly, I have started working on a more detailed, component-stat based set of MLEs for Grove's Baltimore years. I have a method worked out and some initial results, but I won't have time to post them before I leave for a long weekend. But they should be up for vetting by the middle of next week. I haven't had time to do any serious work on MLEs for several years, and it's been fun to re-open the topic.

Chris says, "Relatedly, I have started working on a more detailed, component-stat based set of MLEs for Grove's Baltimore years."

YAY! Grove is the very hardest of the absolute top group to place, because of the minor league thing. Good MLEs (and yours have been very very good in general) would be a real help.

Also, just so no one wonders why I didn't finish the southerners study, the problem comes from defining where a player actually "came from." Take Walter Johnson, for example. In baseball hagiography, he is considered to come from Kansas, where he grew up. But when he was discovered by a major league scout, he had moved to and was pitching for a team in, I think, Idaho. So, for purposes of my study, I would consider him to have come from Idaho, not Kansas. I can find that kind of info easily for the superstars, because they get biographies written about them. But for average or even just good players, it's hard to find anything beyond the birthplace, which is not what I am looking for. So I am sure about the few superstars, but not enough about anyone else to do a proper study. I need to know where the player was living when he was "discovered" by organized baseball, and that info is hard to come by. - Brock

Doh. Posted my ballot in the discussion thread but never in the voting thread. No major disagreements with my prelim ballot, except I had Wilhelm lower. For posterity, my prelim ballot:

1. Tom Seaver
2. Steve Carlton
3. Bob Gibson
4. Phil Niekro
5. Bert Blyleven
6. Gaylord Perry
7. Nolan Ryan
8. Fergie Jenkins
9. Don Sutton
10. Sandy Koufax
11. Jim Palmer
12. Jim Bunning
13. Don Drysdale
14. Juan Marichal
15. Bret Saberhagen
16. Hoyt Wilhelm
17. Goose Gossage
18. Dennis Eckersly
19. Dave Stieb
20. Rollie Fingers

Had stax voted, his ballot would have:

- Broken the Niekro/Perry tie in favor of Niekro.

- Flipped the Drysdale/Eckersley order, moving Drysdale up into 13th.

- Broken the Saberhagen/Sutton tie in favor of Sutton.

The standard deviation in Sutton's scores, which I reported in #5 as 4.23, would have gone up to 4.36. stax's consensus score would have been an above-average 80.

Thanks for running the numbers OCF, nice to know I'm maintaining a good consensus score.

I look forward to the "finals" - ranking a top-15 ballot of all-time pitchers.

I suppose next is taking the top 5 from the 3 large groups.
However, there might be some arguments about e.g. Nichols (6) vs Hubbell (5),
or whether to include Clarkson, etc.

So, a suggestion for the final ranking.

A two-part ballot.

Part I. Each voter casts a ballot of that voter's top 15 HOM pitchers.
Scoring is the same as for our annual ballot. (No elect-me bonus.)
This is the "primary". The top-15 qualifiers go into the "runoff".

Part II. Each voter casts a final ballot ranking the group's top-15 pitchers.
Scoring is the same as for the other ranking exercises.

Little discussion would be needed between the "primary" and the "runoff".
Voters would remove their personal favorites that didn't make the cut,
and add the group's choices to the bottom of their ballot, in order.
Maybe even change one's mind about the order of the other selections.

It will probably be pretty similar to just taking the top 5 from the 3 big groups,
but it would be completely fair to the number 6 pitchers and the early group.

I could live with that.
I agree that it's not about "top 5 from 3 groups."

It's top 15.
I'd probably either go this route, or just have everyone rank top 20 or 25 in a one-time vote (there are 25 SSs, might be a good comparison).

That way you likely see where the 'rejected' top 10 or top 15 candidate ranks with that voter....

A bit of analysis on L Grove vs P Alexander.

Since I had already convinced myself based on "what was obviously true", I decided I oughta look for what the evidence said. Was it so much eaiser to put up great ##s in Alex's day than Grove's?

I have a big Win Shares database, so I used WS as my primary tool. I tried to compare each pitcher's career with their best contemps.

Alex: I found the years pitched, added one to each end, and so used 1910-1930. I found career WS leaders (top 8, excluding Alex) among all MLB pitchers in this time.
Lefty: Same, but I used 1922-42. This kept the # of yrs the same, and even though Grove didn't reach the majors until 25, I was inclined to give him some pre-MLB credit.

Top 8 in WS 1910-30:
Johnson, Walter / Rixey, Eppa / Faber, Red / Cooper, Wilbur
Quinn, Jack / Grimes, Burleigh / Mays, Carl / Coveleski, Stan
avg WS 298
Alex WS 476

Top 8 in WS 1922-42:
Lyons, Ted / Ruffing, Red / Hubbell, Carl / Vance, Dazzy
Ferrell, Wes / Hoyt, Waite / Root, Charlie / Fitzsimmons, Freddie
avg WS 258
Lefty WS 391

I had decided aheaad of time to give Alex 20 WS WWI credit, and Grove 40 WS Minors credit. This makes the ratio of Alexander's WS to the other MLB greats = 1.67 (496/298). And, Grove's ratio of WS to other greats is 431/258 = ... 1.67!

Of course, the normal WS caveats apply, and this does not account for league strength, etc.

*poke You still alive, HoM?

*poke You still alive, HoM?

I've always thought the Hall of Merit should operate like the proverbial task of painting the Forth Bridge.

Once they finish, they should start again. Those electors who return can share the wisdom of their experiences with newcomers. Heck, getting a sabermetric hero like Brock Hanke to be around for the whole exercise would be worth doing an 'HoM 2.0' in itself.

We're all off for founder Joe Dimino's honeymoon - though he is the only one partaking in said event, of course.
:)

By the time we finish ranking the pitchers, we'll be ramping up to talk further about an interesting 2010 ballot.

fra paolo says, "sabermetric hero like Brock Hanke"

What?! Someone thinks THAT about ME?!!! Please tell me you weren't being sarcastic. If you weren't, you've just Made My Month!

Unlimited Thank Yous - Brock

Well, seriously, it is kind of sad to see this thing die.

I always thought the one shot to keep it going was the MVP Project that somebody suggested, but the consensus seemed to be that would be too hard.

Please tell me you weren't being sarcastic.

I am in earnest. The Baseball Sabermetrics were an important bridge between the original Abstracts and the Usenet era.

I think without them, our experience of sabermetrics in the 21st century would be very different, and worse.

So I await the ballot for the sabermetric wing of the Hall of Merit, so I can vote for Brock!

I think you need something like an MVP project to inject new information into the process. At this point we've all created our lists and have created a certain level of detail. An MVP project would be diving in one level deeper. I think an MVP project should run _backwards_ though. Start with the easy ones to work out the process and work your way back. After all, there are no 2nd chances like there are with HoM inductions.

I still think you guys should determine the greatest 50 teams of all time. That would be a challenge.

Well, seriously, it is kind of sad to see this thing die.

But it's inherent in the process, unless you have the 'Forth Bridge' approach. Eventually, you reach the point where you have your annual ballots, then that's it until next year. Just like the Hall of Fame. The question is, will it keep attracting a sufficient number of electors on an annual basis?

I wouldn't call it dead, though. It's resting.

I'm one of the HOMies that was in favor of doing the MVP project and would still like the group to do it. I think it's a natural extension of what we've done so far and, as DL says, it's taking it one level deeper. We've looked at the forest, let's look at the trees. It may yet give us new appreciation for certain players when we actually immerse ourselves in the year to year context instead of a general overview with a slight disconnect to what else was going on during the player's career. Who would be up for it?

Sure, I'd go along, if we had the infrastructure. Is there someone (not me) who would be willing to post the top dozen or so WS/WARP/whatever other systems you want for each season? We'd also have to decide on a voting structure.

I'd suggest having two simultaneous elections each cycle, one for each league, same year. Just the MVP - pitchers should be eligible for that, and whoever posts the WS/WARP/etc. should be sure to always include a pitcher or two.

Or would it make any sense to cross-compare leagues and elect just one for each year?

That raises the separate issue of how to handle the Negro Leagues - elect an MVP (or two) for each pertinent year? Back away and say that our data isn't detailed enough at the single season level? Try to cross-compare with the MLers of the same year? Any of these would seem to require more research and more data than we have used so far.

With or without the Negro Leagues, this would require more data. It's only worthwhile if there are people willing to sink that kind of effort into things - Chris Cobb, KJOK, Dan R, et al. If we don't have enough people of that ilk, than it won't be worth it.

I still think you guys should determine the greatest 50 teams of all time. That would be a challenge.

You just want to start another doomed project.

> how to handle the Negro Leagues

If you work backwards that gives us plenty of time to figure that out while we determine whether our process is giving good results.

I don't believe we have the data to do meaningful year-by-year NeL picks. If the HoF were ever to release its full data set, then we could do it. But until then, not really.

Even if we had the data, I think putting NeL seasons and ML seasons in the same pool wouldn't work well, because of the difference in season length.

- I agree on working backwards re MVPs.

- I think it's workable, as noted it's not like we all have to spend 8 hours figuring out who the top 25 guys are. And some years, it's no debate at all. But still would be fun. I could be wrong, but I'd think a lot of HOM voters who got a little worn down might find this much more manageable.

- With the Negro Leaguers, I do think we can consider them alongside the MLBers of that year. We did it with their careers, I think we can do it with their seasons. We reached a consensus on Smokey Joe Williams being in the same league, so to speak, as Pete Alexander. We can get a sense of which years where Williams was at his peak, and look at the MLB pitching seasons as well.

What Chris Cobb says about "meaningful year-by-year NeL picks" fits what we think we know (quantitatively) about many NeL careers, commonly by heavy reliance on his estimates. Recall the discussions of regression and flattened peaks. Because the data are incomplete and the championship seasons in the formal leagues were short, we probably have something half-way between true season records and moving averages, with a lot of statistical variance too.

What would we do with Pete Alexander in a season MVP project if we had the familiar batting and fielding records without the pitching records from his leagues? Suppose we had judged, based on incomplete game records and anecdotes, that Old Pete was a "career great" comparable to Matty, Seaver, and Maddux. We judged that he enjoyed a dominating peak from 1915 to 1920, including four full seasons, we knew, but missing most of 1918 and some of 1919. He would seem to be the MVP, in fact as a kind of career award, for zero, one, or two of those four seasons when the leading batter-fielders were relatively mediocre.

"we probably have something half-way between true season records and moving averages, with a lot of statistical variance too."

That is, we probably have something
on average half-way between true season records and true moving averages of consecutive seasons, with noise in each one.

Paul has accurately described one of my reservations about putting NeL players into the same pool as ML players for seasonal MVP rankings.

The other reservations are that, first, our player-season data is highly incomplete. We have fairly complete season-by-season data for the top players, but we have very little data about other players who might have had a couple of great years but who did not, on the whole, have HoM careers. Second, even in cases where we have data, many, many more seasons of MLEs would need to be generated to make comparisons. And then the questions of what to do with season size and regression would arise. And the fact that our tools for evaluating NeL pitchers (to say nothing of fielders!) are much, much cruder than our tools for evaluating hitters.

I suppose if we were willing to accept that what we were in fact doing was throwing some of the NeL greats in with major leaguers, rather than making a rigorous effort to identify who was really the most valuable player in the NeL for a given season, we could do that. But I think that approach would empty out the meaning of "most valuable player" quite a bit.

ERA+ and IP are not as useful tools as they would usually be for comparing Johnson or Alexander with Grove or Clemens or Maddux, say. The environment in which Johnson and Alexander thrived was extreme. The absence of the threat of the home run made it possible for the best pitcher of the time to throw many more innings by saving stuff. The fact that both Johnson and Alexander took huge steps down to earth once the lively ball era arrived is suggestive of the importance of the extreme conditions. Part of this was due to age and health, but part was, I think, a more realistic reflection of actual ability. I do have Johnson at the head of the list, but the differences aren't as large as one might think. It's a lot like the Hornsby/Morgan comparison.

"I suppose if we were willing to accept that what we were in fact doing was throwing some of the NeL greats in with major leaguers, rather than making a rigorous effort to identify who was really the most valuable player in the NeL for a given season, we could do that."

Right, I am conceding that we may have a very hard time identifying the "fluky NeL MVP" - the guy who had the one spectacular year, and otherwise wasn't that great. But I hope we can do some justice to the all-time NeL greats, even if to some extent they tend to win an MVP when there's no "definite" regular AL or NL MVP, and likely won't in other cases even if hypothetically it is deserved.

We can't be definitive, but we can try our best. As with the HOM elections, I think these ballplayers are owed that much.

I think an MVP project is a GREAT idea, though the NeL issue would need to be dealt with.

What format are we talking about here for a "MVP Project"? There was a lot of discussion of this, what, 18 months back?

And IMHO, the logical starting point is the first complete retrosheet season, 1954, and working to the present, 2011, 58 elections. If there is sufficient interest after that, the project would continue to the earlier years. Perhaps the HOF would have released the NeL data by then.

I think this project should be done by individual leagues, not by taking the entirety of MLB for each year and doing one vote for all players. Each league has its own set teams and different circumstances. A combined vote could be done also but I think that it should at least be by league.

It could be as simple as submitting your top 15 MVP ballot for each league and have it tabulated either MVP ballot style or HOM vote style. From there you get the league MVP but also the league team for each year based on the voting results (highest scoring pitcher is the league's best pitcher, highest scoring second baseman is the league's first team at that position, etc.).

As for the NeL, if we had enough data for a year we could vote for them in their own league context and in an overall vote, but that is difficult with the incomplete information we currently have. Unlike the HOM or other career based lists, this one depends exclusively on what was done in the specific year. Lack of statistical information for the specific year being voted on is a significant problem for inclusion in this case in an overall vote.

I would love it if I could honor players like Clarence Williams, Bill Monroe, Pedro Cepeda, et. al., by acknowledging their superlative yearly performances, not only in their league context but in an overall professional season ranking. I just don't see how it can be done in an overall scheme at this moment in time with the information at hand. Career wise you can attempt it (as we have done). Break it down to one year..., I don't know. I'm willing to try if someone has a good idea on how it could be done.

Even if we had the data, I think putting NeL seasons and ML seasons in the same pool wouldn't work well, because of the difference in season length.

C'mon, where's the courage we had in picking the first class of HoMers in 1898 and of picking Dickey Pearce and Lip Pike? So we consider qualitative data. I mean I know the world has changed since 1898 (our version of 1898) and numbers have long since driven qualitative analysis out. All the more reason to get on to something new where some more creativity is required. We were up to that challenge. We're up to this one.

But of course MVPs have always been league-specific. That's the real issue. If all we have to do is pick a NeL MVP, what's the problem. If we have to combine AL-NL-NeL, that's harder and it's not what "MVP" traditionally means. Still I'd be up for it.

Making it easy for ourselves is not what made this project great.

C'mon, where's the courage we had in picking the first class of HoMers in 1898 and of picking Dickey Pearce and Lip Pike?
Making it easy for ourselves is not what made this project great.

...........

BINGO!

Plus the Negro Leagues themselves for many years had their own "American" and a "National" League to make more convenient for us.

Maybe a great NeLer only winds up 5th in the voting when he was really the best.

Isn't that better than ignoring him completely?

I think everyone but us tried that already, and it sucked.

Hey guys, realize I've been gone awhile, getting back into the flow here now . . . kind of a long honeymoon. Actually the honeymoon was a week, and then it took me a month to catch up on everything.

"I wouldn't call it dead, though. It's resting."

I agree!

I like the MVP idea . . . but I don't understand why we'd go backwards and not forwards?

Going forward you get the 'will Honus Wagner win his Xth MVP, etc.'. I like the 'walking through time' idea better, personally.

I guess I can see the case for more data in the future. But at the same point, we already have MVP votes from 1931 onward, Cy Young votes from the 50s onward. The real value is in figuring out the old seasons, right?

I would think we should stick to coming up with an MVP/Cy Young for each league.

Couldn't we do both? Use the league votes as a nominating vote for the overall vote? Say only top 10 in each league are eligible for the overall ballot or something?

Also, would it be an MVP or a MOP? That's another can of worms I suppose . . .

I definitely agree on not taking the easy way out, if you challenge people to do something difficult, generally they respond well.

We also want to do an overall pitcher's vote right? Would that be say everyone ranks their top 20 (which makes it ballot counter friendly)? Going to take me awhile to catch up (reading everything posted in the last month) . . . so if I'm bringing things up that were discussion in June or something, I apologize!

Also one thing, regarding the 'do it again' idea . . . I had an idea . . .

Maybe if we can get the website going (I'm working on the table I owe tange from 5 months ago now), we could start over doing our personal Halls of Merit. If we could set up a form on the website, each voter could update his year by year. We could discuss each year for a week, update, etc.. They would all be easily accessible with a link to the voter.

I have no idea how hard that would be to set up . . . I'm going to cross post this to the website page.

I was for the MVP project a year+ ago, and I'm for it now. I'd say break it down by league (AL, NL, NeL), with the same ballot structure as the HOM, no separate ballot for pitchers. I'd also prefer going forward through time.