Holy peak and prime, Batman!

Moving three of my own posts over from the Spahn thread:

The contemporary comparison [to Spahn] is to Roberts. I've got Roberts at a career 295-226. Roberts has a three-year stretch (1952-54) with equivalent records of 24-13, 26-12, and 25-13, with two very good years just before that. Spahn never had three consecutive years like that, or 5 consecutive years like that, any maybe not even a three best or five best years like that. So I'm willing to say that Roberts had a better peak than Spahn. But for career value, the difference between the two career equivalent records is 45-16, and 45-16 is a lot of value.

---

Looking again at what I wrote [above] about Robin Roberts: "only" 45-16 short of Spahn's career value; a better peak than Spahn, comes on a ballot with no overwhelming holdovers; same year as Koufax, but I'm kind of a career guy with pitchers, and Koufax isn't that.

Roberts will be #1 on my 1972 ballot. If he were eligible in 1971, he'd be #2 behind Spahn.

OK, I shouldn't say "only" 45-16; that's two seasons each better than a typical Cy Young season.

---

n my RA+ equivalent system:

Koufax, 1962-1966: 107-46, 137 FWP
Roberts, 1950-1954: 118-63, 132 FWP

Another way to put that: the difference between them for those 5 years, 11-17, is close enough to replacement as to have fairly small value.

But even there: I haven't corrected for the pitcher's own offense. Roberts certainly wasn't a good hitter (yeah, there was a 120 OPS+ one year but that was a fluke), but for a pitcher, he wasn't completely hopeless with a bat in his hand. Koufax was completely hopeless with a bat in his hand; you have to discount him a little for that.

Outside of those 5 years - well, Roberts had plenty of years in which he was pretty ordinary. But still:

Koufax: 56-49, 34 FWP
Roberts: 177-163, 105 FWP

The difference is mostly just bulk as a slightly-better-than-average pitcher. But slightly-better-than-average bulk has value.

I have Roberts as significantly below Spahn but a good chunk of that is Spahn's war credit. Take away war credit and Spahn's still better but there's an argument.

Anyone else surprised to see a 113 career ERA+? Not as dominant as I thought in his peak, either. Durability is great but the peak is in Wins, not ERA. His best seasons in ERA+ are typical Addie Joss seasons, and indeed he has a lot of years in which he is pretty ordinary. He's on the good side of my borderline, but he's start off around 14 on my ballot. Just ahead of Koufax I think...

Anyone else surprised to see a 113 career ERA+?

1956-57 and 1959-61 are weighing down his career numbers. His peak is better characterized by the absurdly high IP levels.

It would be cool if baseball-reference had by-age similarity pages for pitchers as well. I often use those to get "pre-decline" career numbers.

Did the lower offense of the era lower the OPS+ and ERA+ standard deviation during the 60's?

Did the lower offense of the era lower the OPS+ and ERA+ standard deviation during the 60's?

This is a subject that has been argued over repeatedly on this site, with no consensus reached.

The only matter on which there is consensus is that the win-value of higher ERA+ values is lower in a low-scoring era than in a high scoring era.

It is my opinion that higher ERA+ scores tend to appear in lower-scoring eras, because of the lower baseline in the denominator: If a pitcher has an ERA 1 run better than league average, that will give a higher ERA+ in a 3.5 r/g environment than in a 4 r/g environment.

But not everyone agrees with that case, and that's not really what you are asking about, but I think it suggests that the sd for ERA+ would not drop during the 60s.

His best seasons in ERA+ are typical Addie Joss seasons,

Except that Roberts flat-out destroys Joss in IP. How many pitchers throughout history have been in the top-two in IP 7 times in a row or #1 5 times in a row? That had to affect his ERA+, too.

Except that Roberts flat-out destroys Joss in IP. How many pitchers throughout history have been in the top-two in IP 7 times in a row or #1 5 times in a row? That had to affect his ERA+, too.

That workload thing is what Roberts has in aces. Let's look at 1950-55.

1950 2nd place in innings by 7 IP
1951 1st place in innings by 5 IP
1952 1st place in innings by 40 IP
1953 1st place in innings by 81 IP!
1954 1st place in innings by 52 IP
1955 1st place in innings by 48 IP

Over this time his ERA+ was 135, 127, 141, 152, 136, 121. So the median of those is about 135.

Due to his extreme innings pitched, he actually had about 7 years of workload compressed into 6 years. If you lower his innings pitched in 50-55 you still get a full-time starter but you could then add another ERA+ of 135 in a hypothetical second 1955 season.

A 7 year prime with an ERA+ of 135 is worthy of notice when you consider he did some other pretty good things along the way.

IIRC, he once tossed 28 straight complete games. Not even Rick Langford can match that.

How many pitchers throughout history have been in the top-two in IP 7 times in a row or #1 5 times in a row?

Forty years later, Greg Maddux for one:
2-1-1-1-1-1-2 -8-3-9 in the 1990s (NL rank, innings).
If you wish, extend that to 14 seasons with a 4-6 before and a 2-4 after. And now thru 2005 three more seasons of grey ink after missing in 2002. He led the league (what's a league?) in starts for the seventh time last season.

Forty years earlier, Grover Alexander deserves honorable mention:
1-1-3-1-1-1-1 in his Phillies career (NL rank, innings)

Forty years later, Greg Maddux for one:
2-1-1-1-1-1-2 -8-3-9 in the 1990s (NL rank, innings).
If you wish, extend that to 14 seasons with a 4-6 before and a 2-4 after. And now thru 2005 three more seasons of grey ink after missing in 2002. He led the league (what's a league?) in starts for the seventh time last season.

Forty years earlier, Grover Alexander deserves honorable mention:
1-1-3-1-1-1-1 in his Phillies career (NL rank, innings)

Not too shabby of a group, I must say. :-)

>His best seasons in ERA+ are typical Addie Joss seasons, and indeed he has a lot of years in which he is pretty ordinary.

Well, Roberts was certainly no Early Wynn.

Well, Roberts was certainly no Early Wynn.

Ain't that the truth, Marc.

Well, Roberts was certainly no Early Wynn.

I had my own fears. There are similarities. They show up on each other's similarity lists. Neither won an ERA+ title (though Wynn has one unadjusted-ERA title) and relied on being a 'workhorse' to rack up high value seasons. Also, both had some crummy seasons to lower the career rate stats (Wynn shuffled his good and bad seasons more, whereas Roberts had a much more consecutive peak). I did their sorted WS lines as a sanity check:



Indeed, Roberts' peak simply blows him away. Wynn's non-consective prime is a tad longer -- taking back some value in seasons 7-11 (especially 8-9) and then Roberts' weak years were not nearly as bad as Wynn's.

So, yes, slam dunk choice of Roberts over Wynn. :-)

I missed this discussion while traveling...so, catching up...

I have Roberts at #1 on my prelim ballot but I thought before I punch “submit” I would look at him versus the other long career pitchers still in my top 25-30 (including HoM/not PHoM pitchers). The candidates are listed under each category inn the order that that particular part of the analysis seems to suggest.


ERA+
Rixey 116/145-43-43-38-37-29-24-15-15-13-10-9-9-7-(98)-(95)-(91) in 4495 IP
Roberts 113/153-41-36-34-33-27-23-22-21-7-4-(97)-(96)-(93)-(84) in 4689 IP
(Ruffing) 109/149-49-37-32-30-21-19-11-7-6-3-(99)-(99)-(93)-(91)-(91)-(90)-(88) in 4344 IP
(Wynn) 106/154-42-35-35-26-19-15-10-8-2-(96)-(96)-(95)-(88)-(88)-(86)-(75)-(71) in 4564 IP

For ERA+ it seems that the career total positions each of the four pitchers accurately. None had a huge peak, all in fact about equal for their best 4-5-6 seasons. Where they separate is on their seasons number 6 through 10, not a particularly compelled basis for differentiation, but basically it’s what you’ve got. Of course, Roberts has that edge in IP which maybe moves him up to equal with Rixey on this measure.

Win Shares

16. Roberts 339/35-32-31/153/27.0
75. Rixey 315/26-26-24/118/26.5
51. Ruffing 322/27-25-24/116/25.5
47. Wynn 308/28-25-24/110/24.6

The only possible basis for putting Wynn ahead of Rixey and Ruffing appears to me to be a timeline and/or the BS dump. My reading of James’ own numbers suggests the above order. Roberts pulls ahead quite clearly by virtue of his peak which is more based (from the above) on IP than on being any more effective. Still, there it is.

Roberts 339/35-32-31-28-27-26-20-16-16-15-15-13-13-13-12-12-12
Rixey 315/26-26-24-23-22-22-21-20-18-15-14-14-14-12-12-10-10
Ruffing 322/27-25-24-23-22-22-21-20-19-17-16-15-15-11
Wynn 308/28-25-24-23-21-21-21-20-19-16-16-11-10

Extend it out, and Roberts has 7 20 WS seasons, the others have 8 but I have no problem seeing Roberts as #1 on this measure by virture, again, of that peak. His 10 additional 10 WS seasons also help, while Rixey moves up with his 9 additional 10 WS seasons versus Ruffing with 6 and Wynn with 5.

Offensive Win Shares

Ruffing 32
Wynn 15
Roberts 7
Rixey approaching 0

As pitchers Roberts is closer to this pack than I expected and the batting just tightens it up further. Still Roberts remains the best of the 4, and Wynn I am quite sure remains the weakest. Ruffing and Rixey remain very close.

My dilemma is Roberts #1 on 1972 ballot, when the others are all around #20 or so. Seems like to big of a spread. I have had Wynn ahead of Ruffing and Rixey and will have to fix that. But is 1-20 an appropriate spread between Roberts and the others??? Well, I am a peak/prime voter and those 3 30 WS seasons a pretty impressive from a guy who is basically in the Rixey-Ruffing-Wynn camp. As I said before (based more impressionistically) Roberts is closer to Spahn than he is to Wynn. And considering I am no friend at all of Red Ruffing, I am now satisfied til further review that Wynn will never make my PHoM.

I don't see the Roberts-Rixey comparison. Roberts peak, compared to his peers, was far greater than Rixey's. Eppa Jeptha also had the good fortune of starting his career during the much easier teens, while Roberts experienced the pitcher-friendly '60s when he was on his way out.

Rixey was real good, but Roberts was at a higher level.

I don't know the exact count, but Robin Roberts has about 5 WS Cy Young awards. Eppa Rixey doesn't have any.

I don't know the exact count, but Robin Roberts has about 5 WS Cy Young awards. Eppa Rixey doesn't have any.

Exactly, David. Roberts was superior to Rixey.

>Robin Roberts has about 5 WS Cy Young awards. Eppa Rixey doesn't have any

Spoken like a good peak voter!

But seriously, you can't see any comparison between these two pitchers?

Rixey 116/145-43-43-38-37-29-24-15-15-13-10-9-9-7-(98)-(95)-(91) in 4495 IP
Roberts 113/153-41-36-34-33-27-23-22-21-7-4-(97)-(96)-(93)-(84) in 4689 IP

But seriously, you can't see any comparison between these two pitchers?

Rixey 116/145-43-43-38-37-29-24-15-15-13-10-9-9-7-(98)-(95)-(91) in 4495 IP
Roberts 113/153-41-36-34-33-27-23-22-21-7-4-(97)-(96)-(93)-(84) in 4689 IP

The difference is that Roberts packed in more IP in less years than Rixey did. Rixey never dominated the IP list the way Roberts did (not many other pitchers did, for that matter). Roberts also pitched in a far more difficult environment in his twenties than Rixey did. Pitching 200 IP during the Deadball Era is not the same thing as 200 IP during the first home run explosion.

Not to mention the fact that the NL was far stronger during the 1950s than it was during the 1910s and 1920s.

>But seriously, you can't see *any* comparison between these two pitchers?

** added

** added

Yeah, sure. I myself get used to scanning OPS+ lines and assuming that for two full time players, then OPS+ will determine (roughly) who was the better hitter. For full-time players, PA is often taken for granted.

For ERA+, the playing time component (IP) has much more variance among full time players.

>But seriously, you can't see *any* comparison between these two pitchers?

** added

I understand, Marc, but there's a big difference between their peaks. As David alludes to, that doesn't show up when you just look at their career ERA+ or career IP. One also gets the wrong impression just looking at their seasons of peak ERA+ without comparing their respective IP among their peers.

Also from #16:

>Roberts pulls ahead quite clearly by virtue of his peak which is more based (from the above) on IP than on being any more effective. Still, there it is.

EpRixey Top 10 IP: 1 3 3 3 4 7 8 8 9 9
Roberts Top 10 IP: 1 1 1 1 1 2 2 4 5 5 7 9

That's a major edge for Roberts, indeed, but some of the comments to me almost seem to imply that it's even greater than this.

Rixey does not have Griffith/Pierce issues regarding IP.
We're talking about a couple of extra seasons of Roberts leading in IP. Quite significant, but Rixey's no slouch on that front.

Marc,

Just wondering where you have Burleigh Grimes. He has a higher peak than Rixey, and almost as many IP. As a peak/prime voter, I would think he would be ranked higher than Rixey.

Grimes dropped out of my consideration set a long time ago. It might be because of ERA+ issues...? Somebody I will re-eval some more pitchers.

Top MLB pitchers by WS (numerical ranking includes all players):




When compared to their peers, not quite the same.

WYSIWYG it's not. Looked good in preview.

(Results of the recent Combined Pitchers election (results), with 20-deep rank-order ballots, have provoked more comment on consensus #16 Robin Roberts than anyone else.)

Beside providing that notice and link, let me extend one of my remarks there (italic).

The outstanding feature of Robin Roberts at his peak was his heavy workload in combination with excellent rates. By innings pitched his 1950-56 margins over the third-leading NL pitcher were roughly 5%, 5%, 20%, 35%, 30%, 20%, 5% (his ranks <2 1 1 1 1 1 2>). "Poor" Warren Spahn ranked <3 2 2 2 2 3 3> followed by another 2 between his league leaderships of the late forties and late fifties.

Ten years later, 1960-66 Sandy Koufax ranked < - 4 - 3 - 1 1> in the NL by innings pitched --behind leaders Burdette, Drysdale, Marichal, Drysdale before closing in 1965-66 about 10% and 5% ahead of the third-leading NL pitcher.

Forty years after Roberts, 1990-96 Greg Maddux ranked <2 1 1 1 1 1 2> by innings pitched, precisely the same rank-order record. His margins over the third-leading NL pitcher were roughly 1% 10% 10% 10% 15% 1% 2%, rare dominance.