Of 37 pitchers with 10+ saves last year, it says 11 were exclusively relievers in the minors. That surprised me a bit, I thought more still began their pro careers as starters . Other than Storen, who are they?

Huston Street has to be one, he was also a closer in college. I see Kimbrel is another. Who else?

Instead, the real sea change will happen when managers begin using their closers in the highest-leverage situations late in the game—regardless of conventional baseball thinking or contract incentives that reward saves.

I tend to think that if this was going to happen, it would have by now, or at least we would be seeing significant evidence of an ongoing shift. It's not like managers and front offices are unaware of the advantages of such closer usage. It's just that they perceive disadvantages that, in their minds, outweigh the advantages. Routines, defined roles, warming up, bullpen harmony, yada yada yada.

Who else?

Robertson.

Kenley Jansen wasn't technically exclusively a reliever because he started 5 games and pitched just one inning apiece and he also played 5 minor league years as a catcher, but he fits the spirit of the question.

Edit: I don't think Melancon has ever started a game in the majors or minors.

2015 MLB Saves and Career non-MLB starts per bref:



Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 12/4/2015.

It must be these 11:

Street
Kimbrel
Storen
Robertson
K Jansen
Gregorson
Melancon
Tolleson
AJ Ramos
Carson Smith
Cody Allen

Ramos has one start in his minor league career, but it was a 1 inning appearance so for all intents he's been strictly a reliever as a pro.

You also have guys like Rodney who made 20 starts in 148 minor-league appearances before (essentially) making the majors for good. In his particular case those 20 starts do seem like a legit attempt to develop him as a starter but they were mixed in with relief and he'd been used exclusively in relief the year before. And I was looking at some reliever the other day who had started games at 18-19 in low-levels but got shifted to the pen for good at 20. That's not really any different than some college reliever who never relieved in the minors.

Clearly it's still the case that a majority of closers began as starters and I assume we see the same in the 7th/8th inning guys. But it's been shifting and (1) we're certainly going to see more guys starting their ML career as relievers and not with the intention of developing them as starters and (2) I suspect we'll see more of them trained as relievers in the minors too.

It's one of the reasons I don't think the saves record is particularly safe. Given we've seen many guys be effective relievers and/or pile up saves in their late 30s and even early 40s, it's only now possible for some guy to have a 20+ year career as a reliever/closer. Kimbrel, Chapman, Osuna (Tor) will have those careers if they can stay healthy; K-Rod probably would be having that career if he hadn't been such an a-hole.

Or, for #2, teams will change closer usage. I think we are seeing a few cracks but I won't call it a change yet. Tampa was doing this a bit -- they also had a freaky high number of team saves so maybe they only did it because Boxberger was getting over-used. Betances picked up 9 saves for the Yanks. The other crack is that closer salaries seem to be coming down. Robertson did well but (as an example) not as well as Soriano did just a few years before. Miller didn't do that well, but then he had never closed before. Papelbon's AAV was lower than Mariano's in raw dollars much less after inflation ... and lots of people called that contract a mistake when it was signed.

Maybe Kimbrel and Chapman will break that dam but while SP salaries have been cracking $25 then $30 M barriers, closer salaries have been stagnant. Meanwhile middle reliever salaries have been going up pretty dramatically. $5/6/7 per year is becoming reasonably commonplace -- Hochevar got 2/$10 coming off an injury and with just one relief season; Duke got 3/$15 after one good relief season.

I'll believe it's changing when I see it has changed but a key difference between now and even 5 years ago is that pointy-headed nerds are being listened to and are sometimes even in charge. We saw several teams at least experimenting with batting their pitcher 8th and the Cubs doing it pretty much the entire year -- I still can't get my head around that. So especially for any team that has a strong LH and a strong RH, if the analytics dept says you'll do better using the LH guy in the 9th if that's when the LHB come up, we'll start seeing more of that. Or if they say you should use the "closer" in the 8th if the inning starts with the #2, #3 or #4 spot in the order, we'll start seeing more of that. Or they might say "never, ever use a guy 3 straight days" and we'll stop seeing that usage (not that it comes up very often anyway).

OK, bothered to read the article -- nice and short. Nothing particularly interesting in there and he made the boo-boo of not looking at number of appearances. I believe "closer" innings have been coming down. The guys at the top of the table in 2015 -- Melancon, Rosenthal, Familia -- all threw quite a few innings by modern standards. But Kimbrel, Casilla, K-Rod, Jansen, Perkins, Storen were all under 60 IP; Holland had 32 saves in just 45 IP.

I'm not sure 2005 looks much different but there were 11 closers who topped 70 IP with IronMan Ryan Dempster leading the way with 92!

I suppose if somebody wants to look at #appearances/save (for guys with ... 25+ saves in a season?) over time, I'd be happy.

Apologies for the self linking, and feel free to ban me or whatever if you like, but I took his lead and went a bit beyond it.


http://www.boyofsummer.net/2015/12/a-closer-look-at-closer-usage.html

I see some of what he's saying about the number of innings and lengths of appearances dropping, though I make a bit bigger deal about it than the original author does. But I also wonder if, given some of the tendency to (somewhat) split closer duties, we're not already seeing that trend starting, using closers in higher leverage situations, not just in "Save" situations.

closer salaries have been stagnant

I think GMs must look at cases like Shawn Tolleson (30th-round pick, waiver acquisition, suddenly has 35 saves with league-average stats). It's not that absolutely any random pitcher lying around can become a league-average closer, but that relatively few random pitchers lying around can become league-average starters.

I took his lead and went a bit beyond it.

I think you're conflating two different phenomena. The shift in IP/appearance is indeed pretty trivial. (That was the only thing the other author looked at). The shift in total innings is mainly a reduction in appearances.

Take the two extremes. This year they averaged 62.3 IP at just under 1 IP/appearance -- I get an average of 63.6 appearances. In 2004, they averaged 73.2 IP at 1.06 IP/appearance, an average of 69.1 appearances. OK, so it's about half and half between appearances and IP/appearance.

I'd still argue the IP/appearance shift is pretty trivial. We're mainly talking about a small handful of times that a closer entered in the 8th or a couple of times he entered top 9 of a tie game and then pitched the 10th since he hadn't thrown many pitches yet or the pen was tired. The "job" was still defined in the same way -- you pitch the 9th.

The appearances shift may be equally trivial -- these might all be relatively low-leverage appearances. Teams don't like leaving their closers unused for extended periods but maybe 10 years ago wisdom was never let him sit more than 4 days, maybe not the wisdom is 5 or 6 days. You might get at this by looking at the ratio of 9th inning save opps per appearance.

Similarly on whether there's a shift in closers being used more in non-closing opportunity, you can track this by looking at the mix of holds and saves, being careful to make not of guys who won/lost the job during the season. Probably better than GF. I raised Andrew Miller as a possibility but it looks like that wasn't the case. 60 appearances, 53 GF, 38 save opps, 36 saves, 2 BS, 0 holds.

loss 1: 10th inning of a tie game
loss 2: b9, 1--, tie game ... non-standard
BS1: b10, a2
W 1: t10, tied
W 2: t9, tied
BS2, L3: b9, a1 ... gave up 1 then pitched 10th gave up another

His other appearances seem to have been a mix of top 9/10 tied game, hasn't pitched in a few days and maybe a few where he was warming up to close when the Yanks added a few and made it a non-save. He did have 3 saves >1 IP but none after June 3. Those three and a game where he hadn't pitched in 5 days were the only games he entered before the 9th.

Somebody else will have to take the time to map the connection between Betances' 9 saves and Miller's usage. Possibly there was something different going on there or maybe it was just games where Miller was tired or extra-inning games where Miller had been used or whatever.

Mapping is not necessary. Most of Betances' saves occurred during the month that Miller was on the DL with a forearm injury.