This is very interesting and obviously reflects a lot of work.  What implications do you see for teams today?

1)  Draft young fastballers and let them go via FA?
2)  Draft HS fastballers in the mid rounds?
3)  Sign old junkballers as FA?
4)  Trade productive young starters at age 27?

Which teams do you think most profit from these implications?  How do you think these conclusions translate to the minor leagues or to relief pitchers?

Chris,

A few minor points.  (Too much material to read in one sitting)

Your “aging” patterns is not really an aging pattern.  I would guess that you would have gotten similar results if you simply used IP.  What your data is telling you is how many pitchers at each age class are pitching.  When you have things like expansion, it’s easy to see that this opens up the door for the older guys to stay in the game.  To do a true aging pattern, you need to have the same pitchers at the same weights in two or more age classes (which by itself will bring selective sampling issues).

Technical notes: to sort in Excel by row, click data/sort, and select “options”.  Toggle your orientation as you need it.  Your other choice is to cut/PasteSpecial, and select “transpose”.  This turns rows into cols and cols into rows.

Did you actually go through each pitcher in the (excellent) James/Neyer guide manually, and noted his pitch type style?  This is one of those times that I wished the book came with a CD.  I’d love to get that file from you, if you can.

Tom

Looks great.  I’m going to have to print it out to read, but I can’t print.  Are Primer articles set up so that they won’t configure to a printer somehow?  Just call me clueless…

Never mind.  Apparently I had a problem with Mozilla—it prints from Explorer.

Hey Chris,

This is phenomenal, one of the best articles I’ve read on this site, if not the best.

Thank you,

Derrick

Sorry about some of the columns being off. 

Technical notes: to sort in Excel by row, click data/sort, and select “options”. Toggle your orientation as you need it. Your other choice is to cut/PasteSpecial, and select “transpose”. This turns rows into cols and cols into rows.

I’ll check that out.

Did you actually go through each pitcher in the (excellent) James/Neyer guide manually, and noted his pitch type style?

When possible, yea.  Some weren’t listed, some had no clear #1 pitch. 

This is one of those times that I wished the book came with a CD. I’d love to get that file from you, if you can.

Check your e-mail.  If anyone else wants the database, just e-mail me and I’ll send it to you when I get the chance (note: may not be done right away).

What implications do you see for teams today?

Well, that’s a trickier one.  Any implications taken from this would have a giant qualifier on it - just because it’s some sort of trend doesn’t mean its true for all pitchers. 

Derrick - thanks for your kind words.

A few notes . . . some of the conclusions differ from what I said in Toronto.  There, I was struck by how close the 1890s and 1960s were - in part because I never looked at spitters.  In whipping this article up it seemed clearer that 1960s-born pitchers are retaining more value later. 

Really, in some ways the real story are the 1930s/40s/50s/ guys.  They’re the great step backwards.  Deadballers age better than box’rs.  Early live ballers age better than deadballers.  Current age better than early liveballers, but the mid-live ballers were the only major step backwards. 

Also, the bit at the end in comparing knucklers to late arrvials comes from a talk I had with Dial.  He wondered how much of knucklers aging was simply due to their getting a later start than most.

Haven’t read the whole thing yet.  And sorry if I missed you addressing this point.

Generally you’re measuring aging patterns as % of peak.  This would seem to guarantee that pitchers who pitch during a transition in starter usage will “age poorly”.  For example, the young guys starting in the 60s/70s (roughly 1945-1955 birthdates) were making 35-38 starts a year, then baseball transitioned to the 4.5-5 man rotation and these guys were making 31-34 starts a year later in their careers (of those who had late careers).  Four fewer starts a year means a lot fewer win shares.

To take one example, Rick Reuschel (born May 1949) made 37-38 starts almost every year through age 31.  In 74, he made 38 (and 3 relief appearances god love him) which was only 6th in the league.  Late in his career, he was making 32 to 36 starts a year, but he tied for the league lead with those 36.  Reuschel, one of the best-aging pitchers of his generation, did as much in his late career as most of today’s best old pitchers (158, 131, and 116 ERA+s at ages 36, 38, and 40 and always 200+ IP).  But he, and every other starter in baseball, was making 10% fewer starts and pitching about 15% fewer innings than he did in his prime.  He aged better than Tom Glavine, probably Maddux, and maybe even nearly as well as Clemens (given Reuschel’s lower talent level to begin with), but he may look worse compared to peak just because pitcher usage changed.

Put most simply, in raw win shares, pitchers of earlier eras had much higher peaks than today’s.  They could have posted the same number of late-career win shares in the same number of starts as today’s pitchers, but measured against their own higher peak, they would appear to be much worse.  That wouldn’t appear to be enough to explain the whole difference, but some.

Not sure there’s an easy solution.  You could look at raw win shares.  Or you could try to normalize in some way—maybe looking at % of win shares to % of innings pitched at each age.

I think Tango’s comment brushes up against this too.  It’s not clear to me whether you’re getting at different aging patterns or changes in pitcher usage.

For example, the young guys starting in the 60s/70s (roughly 1945-1955 birthdates) were making 35-38 starts a year, then baseball transitioned to the 4.5-5 man rotation and these guys were making 31-34 starts a year later in their careers (of those who had late careers). Four fewer starts a year means a lot fewer win shares.

Yea, but those guys were part of a 30 year swath where pitchers kept aging like that.  I don’t think there was a continual 30 year revolution in pitcher usage.

Your larger point stands that changes in usage would influence this.

Uh, some (many?) current pitchers have the benefit of performance enhancement drugs.  Just as the likes of Bonds shattered 100 years of aging statistical data, we see similar, goofy aging trends for contemporary pitchers.  The answer swirls in the juice.

The answer swirls in the juice.

First, Catfish, you have to know if you even have a question the juice can answer.

Then came the 1980s. Now absolutely none of that is true. The exact opposite is going on. Sure. OK. Why not? What the hell.

I think that you switched from birth decade to pitching decade somewhere around this quote, but I couldn’t tell when.  Did the effect start happening in the 1980s or did it start affecting pitchers born in the 1980s?