I’m proud to say that something I named is being discussed on MLB Network. Tom Verducci found what he called the Year After Effect and I found it, years later, calling it the “Rule of 30″, which lives on inside the Team Health Reports. After discovering I was repeating his work, I decided to start calling it the Verducci Effect. (He didn’t name it after himself.) It’s stuck and I’m glad. Verducci deserves the credit.
He was on MLB Network with Victor Rojas (FOP) discussing the reasons behind it and the pitchers to watch out for. Say what you will about MLB Network — and I do — but they’ve got the time and resources to do a lot more of this, educating people. Whether it’s injury management, WPA, or a million other things, there’s an opportunity, even if the market is small.
There’s a lot of intriguing looks at the Verducci Effect from around the web. There’s probably more I’ve missed, but I wanted to point to this one by kinda-BP Idol finalist Jeremy Greenhouse. If it’s there or not there at a deeper statistical level — and I’ll be honest, I don’t have a firm grasp on Jeremy’s methodology — it’s great that we’re looking at it. I think on a more superficial level, it’s still a nice rule of thumb. I know there are others looking at it, but I do want to point out that I think the stats are much less important than getting to the root of the problem: how do teams effectively manage the health and workload of starting pitchers?
Will/Kevin, can BP do an annual article on VE pitchers? The at-risks could be examined for the coming season, and those that were red-flagged the year before could be revisited.
Of the 10 Verducci highlighted, I'm particularly curious about Porcello, as he the Tigers seemed to measure him as much as possible. He also seemed to be asked to be a GB pitcher instead of power pitcher, which may be his more natural setting.
Can somebody please help me understand the following? (1) Verducci says that the player must be 25 or younger to be at risk...if so, why is Josh Johnson on his list? If the age is actually higher, then why isn't Edwin Jackson included?; (2) I've heard that minor league innings do not count...if so, why are Carillo, Norris, Latos and Wade Davis on his list?; (3) where are Brett Anderson and Trevor Cahill...don't they fit the criteria?; (4) I don't understand how Scherzer's innings increase is 42...when looking at his prior years, I would think that his innings have increased by around 61.
Thank you...I've been thoroughly confused about how the Verducci Effect is actually applied...there doesn't seem to be a consensus as to which players are included each year. Even Verducci doesn't seem to be applying his rules the same for everyone unless I'm missing something.
Just in time to be too late for 75% of fantasy managers. Not that I would depend on MLBN for fantasy analysis, but if you're going to have a fantasy segment, a weekly March 1-30 seems like the obvious answer.
To be honest, this is my nitpick about BP and many other baseball sites/newspapers. Many leagues hold their drafts by the time their annuals or fantasy analysis is published.
Heh, perhaps you should be nitpicking your leagues. I've never seen the point of drafting in mid-February before pitchers start blowing their arms out.
People spend March following NCAA/March Madness instead of spring training.
In keeper leagues, leagues ewquire keeper decisions mid-winter to allow time for a redraft.
November/December/January/February each have federal holidays which make live drafts/auctions easier to schedule.
Of course it is still a big juggling act since there are still free agent signings and trades as late as Opening Day, but as an example, note that the Annual wouldn't have been ready for the BP Kings league (or many other Scoresheet League drafts).
I'm excited that *this year*, the Annual arrived a few days before my (Strat-O-Matic perennial league) draft, instead of the day after. Of course that means a lot of NOT doing other things before Sunday morning...but I'll suffer that.
The Verducci Effect is a fancy way of saying that pitchers get hurt. It seems beneath the rigorous, outcome-oriented work BP has made their name on. Frankly, I am shocked that you continue your full-throated endorsement.
Verducci pitchers are, by definition, coming off healthy, injury-free seasons that garnered some measure of success. It is likely that, as a group, those pitchers will regress the next year. And taking out a healthful season makes makes it more likely that the next year will be something less than healthy.
And then there are the results. Last year Verducci red-flagged 10 pitchers: Tim Lincecum, Jon Lester, Clayton Kershaw, Jair Jurrjens, Cole Hamels, Chad Billingsley, Dana Eveland, Mike Pelfrey, John Danks, and Jon Niese. Even with a lone sample, I think it needs to be addressed when the results point so far in the opposite direction. If not by Verducci himself than by the guy who gave name to an "Effect" that just produced a Cy Young winner and the best pitcher in the AL over the last five months.
I was looking into Brett Anderson last week and found that 12 SP's in the last 20 years have thrown a 175+ ip season by his age (21). They are: Anderson, Cahill, Felix (x2), Cain, Kazmir, Greinke, Bonderman, Sabathia (x2), Ankiel, Valdez, A.Fernandez, Avery. Of the 10 we know, only two who had issues the next year--Greinke and Ankiel. And both of those were psychological. Ten pitchers at age 20-21 putting up huge innings increases, and the Effect--strictly speaking--has applied to zero of them.
I'm sure other commenters will have more insightful critiques, but this is something that deserves to be fleshed out.
Has no one yet studied the Verducci Effect in a rigorous way? For something that's been around for a few years, I'm kind of shocked that no one has gone through all the data. It seems like it wouldn't take very long to do.
The consensus is generally that this is, as Tuck says, regression. In the very least, it would be nice if this kind of stuff got addressed before Will gave whole hearted approval.
I think the Verducci Effect could use some fleshing out and some rigorous study. That being said, I think you paint BP a bit unfairly under the assumption that there is some analytical way to predict pitching injury/fatigue. One pitcher's sore toe becomes another pitcher's back and arm problem.
But Carrol claims there is an analytical method, and he claims to have found evidence to support the "Verducci Effect."
If he has evidence that these pitchers are at risk beyond normal regression and injury concerns, he should publish. Until then, he's opened himself up to fair criticism--as has BP.
Could you show me where I said that? A full study of this is something Matt Swartz talked about late last year and I hope it's on his (or someone's) to do list. I simply don't have the statistical chops for that. Back in '02, I found that 20% increases were a negative indicator and ran with it.
Will,
You wrote in this unfiltered post "I found it, years later." It sounded to me like you said you independently did an analysis that led to the same Verducci Effect. Apparently that's not what you meant?
Not in any rigorous way. It's a very superficial thing. "Oh look, pitchers with X increase tend to get hurt a lot." I didn't do regressions or any of the sharp knife kind of stuff. It's akin to saying "players that wear white shoes tend to get hurt." It makes sense, we can watch each year, but I can't tell you WHY it works. It's a very simple, rule-of-thumb kind of thing, not a statistically vetted formula.
You say you found it in this very unfiltered post.
And then there's the BP Glossary entry for the Verducci Effect:
--
Named for Tom Verducci of Sports Illustrated, this is a negative forward indicator for pitcher workload. Verducci, who called this the 'Year After Effect,' found that pitchers under the age of 25 who have 30-inning increases year over year tend to underperform. Will Carroll independently found that pitchers who break the "Rule of 30" tend to get injured. Carroll renamed this 'rule' the Verducci Effect in honor of the man who initially found the evidence.
I think that what Will is saying is that he found a simple relationship between pitchers who's major league innings increase by 30 in one year and injuries the following year but that he hasn't done rigorous analysis to rule out potential confounding explanations or to try to determine exactly when and where this effect is likely to be largest, etc.
Is that bad though? The effect itself is a suggestion that these guys are more prone to major injury than other pitchers. 12% is only bad if the rest of pitchers get injured at a statistically similar or greater rate. If it was found that other pitchers have major injury at a rate of say 8%, then that 12% means that they're on to something.
It's not like I've done the work or anything, but if you're going to call a number bad you need to consider it in the context it's supposed to be considered in. To the best of my understanding, the Verducci Effect is saying that these guys are riskier assets than the average Joe the Pitcher. The injury rate of VE eligible pitchers alone tells us jackshit.
That Razzball article didn't include minor league pitches and the Verducci effect does include them, so not sure how helpful it is.
It found 33.9% did dropoff which is quite a bit, then isolated out some other factors from that 33.9% and was left with a number that has little meaning left and that was still 27%. It is also using xFIP as the stat of choice which might not actually show the regression accurately since a spike in HR/FB could be caused by a tired arm.
I'd like to see a really good study done on this but I wouldn't say it isn't for real based on that Razzball one.
What is the success rate based on? Because besides injury, there's also general ineffectiveness to watch out for among Verducci pitchers. Hamels may have been an example of this, along with Verlander, off the top of my head.
Just like one stat doesn't tell the whole story for every player, the Verducci Effect is not the end-all-be-all theory on injuries. It's just one more thing to keep in mind when projecting young pitchers.
"I was looking into Brett Anderson last week and found that 12 SP's in the last 20 years have thrown a 175+ ip season by his age (21). They are: Anderson, Cahill, Felix (x2), Cain, Kazmir, Greinke, Bonderman, Sabathia (x2), Ankiel, Valdez, A.Fernandez, Avery. Of the 10 we know, only two who had issues the next year--Greinke and Ankiel. And both of those were psychological. Ten pitchers at age 20-21 putting up huge innings increases, and the Effect--strictly speaking--has applied to zero of them. "
Tuck, the Verducci Effect has nothing to do with actual IP count. It has to do with season-by-season IP increase.
Verducci himself thought Anderson was a serious red flag until it was pointed out to him (by Billy Beane) that Anderson and Cahill both pitched on the Olympic Team in 2008, and that those innings aren't accounted for.
I'm not going to look at every pitcher you mentioned, but I'll just take the first two, other than Anderson and Cahill for whom there is little historical data:
Matt Cain's largest inning increase between seasons, other than his promotion from low A, is a whopping 17.2 inning increase between 2007 and 2008. Matt Cain has remained healthy, an example of a pitcher who did NOT surpass the 30 inning increase "rule" of the VE.
Scott Kazmir had an inning increase from 134 to 186 in 2005, or an increase of 52 innings. He only pitched 144.2 innings in 2006. I don't recall if he was injured or not, but I suspect injury is the reason for the 42 inning decrease. Then again from 2006 to 2007 he increased 62 innings pitched. In 2008, he was injured, leading to only 152 IP, a decrease of 54.1. Last season he only managed 147.1 IP due to injury.
Just because Kazmir managed the same IP as Anderson by the same age means nothing. It's the increase of usage that counts, and I believe on that count, the VE is somewhere in the neighborhood of being a 75% effective predictor of injury in starting pitchers under the age of 25.
So in other words, your criticism is complete irrelevant to the the effectiveness of the VE.
Oh yes, and in contract to your point anyway, I believe Kazmir DID get injured the year after his Age-21 186 IP season, leading to his only 144.2 IPAge-22 season.
Just occurred to me...are innings pitched during the Olympic season and the WBC taken into account in your THR's? If not, how much does this affect the 'color' of young pitchers on your spreadsheet?
It's not supposed to be scripture, it's just a rule of thumb sort of way to identify the risk groups. Over the years Verducci's record of identifying risky pitchers has been prescient. However, to state that because a pitcher has a big innings increase he will get hurt isn't the point, it's simply a way of identifying risk.
But does it even do this? How does the list of Verducci Effect (or rule of 30 or whatever you want to call it) pitchers compare to any other list of starting pitchers? Or young starting pitchers? Or young starting pitchers who were born in the winter?
I've been following this list for a while. I have read one article from Hardball Times addressing this issue:
The bottom line: a dramatic increase in innings on a young pitcher elevates the risk of injury or a setback to their development.
But the evidence points to the opposite. Pitchers who see a large increase in workload are more likely to continue to be successful than those who don’t. It’s important to remember that correlation does not mean causation—just because throwing a lot more innings than a pitcher ever has before is correlated with future success does not mean that managers should be riding their young pitchers hard—but it does imply that Verducci’s argument is incorrect, and there is absolutely no reason that we should expect these YAE candidates to do worse because they’ve overworked."
I'd like to see a real study that addresses this issue. I hope you guys give Matt a shot at it soon.
I don't think Verducci did the hardcore analytic work, and I bet with close analysis you could refine the risk factors. PAP is better in the short term. However, in the macro sense, Verducci is absolutely and positively right, and it's good he's spreading awareness about this phenomenon. The more the knowledge is out there, the more arms can be saved, especially at all the levels below MLB.
This is something MLB Network should be doing rather than rolling their seemingly endless supply of has-been players who are absolutely clueless when it comes to evaluating players using modern tools.
MLB Network, like ESPN, is in the entertainment business, which is different than the baseball analysis business, apparently. Its why I come here and haven't been to ESPN's site in years (other than for fantasy football).
I've said it before and I'll say it again - especially to people arguing that sabermetrics is mainstream - that you are the minority here. BP's audience isn't big enough to be a rounding error to ESPN. They made $8 BILLION in subscriber fees alone. That's just TV. No ads, no magazine, website, anything. BILLION. Their ratings and the ratings of MLBN are huge.
BP's pretty successful for a niche site and we're growing, but anyone who thinks this is mainstream isn't getting it. OPS was trotted out last year by ESPN as cutting edge and by and large, people rejected it as too complex. The challenge of the next decade is to take all the knowledge and find out what's usable to a larger population.
... but you do get guys at ESPN like Neyer, Law, Olney and formerly Gammons who have been writing about OPS and other metrics for years. Heck, I found out about BP from Rob Neyer some 3-4 years ago. At least for their online content/website, ESPN seems "aware" of sabremetrics.
Verducci gives Hamels as an example, but I've read here (and other places) that despite the rise in ERA and the perception that he struggled, Hamels 2009 was not much different than his 2008.
I think that this is going to be the hard part of a rigorous study of the Effect - how do you define "underperform in the next year"? Hamels is a great example of how, if you used ERA to show underperformance, you'd probably be wrong. Do you look at peripherals? How the guy does versus what SIERA thinks he should have done?
Or do you just look at the likelihood of injury/lost IP in the next year, as Will suggested with the Rule of 30? I think that's what I'd like to see rigorously studied first. I think that's the more complicated study, though. You'd have to make sure the drop in IP was due to an actual pitching injury, and not a hitting injury or a freak Bobby-Ojeda-hedge-clippers-type injury. Those things shouldn't count, right? You'd also have to leave out IP lost due to being sent back down for ineffectiveness (unless an injury popped up later).
In any case, I hope Matt or someone else gets to do this.
What may clarify things more is if Verducci said something like "Some pitchers who are likely to get injured are more likely to do so after large increases in workload expose their propensity to injury". That is probably closer to the truth.
Did Anibel Sanchez have an underlying propensity for injury? Probably. Would that injury have surfaced so early in his career had he thrown fewer innings? Maybe, and my gut says yes.
Hamels is a bad example. The only thing that changed is he threw his curve and change less often, and his fastball out of the zone got hit more, and his bad BABIP inflated his ERA.
I found it very comforting that two pitchers that didn't fall under that effect are Derek Holland and Tommy Hunter, each of whom did not increase their innings load significantly in 2009.
Tom mentioned that only four of the 34 pitches he had red-flagged didn't break down and had lower ERA's the next season. I would be curious to know which four.
Last year alone, Kershaw, Jurrjens and Lincecum all had lower ERAs and didn't break down. Further, as previously mentioned, Hamels and Pelfrey's performance were both about the same despite elevated ERAs. John Lester's ERA went up but his xFIP went from 4.08 to 3.13 and you'd be hard pressed to prove he was worse last season than the year before. In fact he was better in every way except for a .20 rise in ERA that can probably be explained by BABIP. John Danks was a tad worse last year but not something out of the ordinary for any young pitcher.
I don't care all that much for the Rule of 30 or the YAE or whatever you want to call it. It's not a study. It's an observation that young pitchers should be treated with some care which I don't think is a ground breaking observation. There is no comparison between those who break the rule and those young pitchers that don't. Aren't ALL young starters prone to up and down years? If you make a list of young pitchers who had a good year and observed that they had an increase in ERA the next year, regardless of innings increase, would it be much different than a Verducci list?
That's what we really need to know and no one has done that. Matt, we're counting on you.
Does an observation have to be groundbreaking to be useful? Does having additional precision in that observation help? For example, we know that pitchers who throw a lot of pitches in a game have a tendency to get tired or hurt. Until things like PAP came around and the idea that 100+ pitch starts constitute risk/abuse, we didn't really know what that threshhold was. Similarly, how many "more innings" is "too many innings"? The Verducci Effect helps as an observational rule of thumb.
I don't mean that it's useless. But let's not make more of it than it is. I think this rule is being treated as ground breaking and really informative and I just don't think it is. To be honest, what is the difference between a "rule of 30" and saying you should be careful with young arms? It's Ok as a rule of thumb. Rules of thumb aren't a bad thing. But doing a study that compares rule of 30 breakers vs. non rule of 30 breakers isn't that hard to do for someone with time.
Actually, I have a lot of problems with PAP and the 100 pitch limit. Most of the studies that have looked at both have concluded that it hasn't helped to limit pitchers to 100 pitches as if 100 pitches is some magic mark. Today, you see very few 120 + pitch games. That's a good thing. But there's no reason to believe that breaking the 100 pitch mark is a problem.
Being careful with young arms -- and even older arms is a good thing. But I don't think the rule of 30 tells us that much. And I'd like to see it studied a bit and fankly I'm surprised it hasn't been.
And when the idea that young pitchers performance is naturally variable and that they're injury-prone leads to hard caps on innings jumps, AS IF THAT WILL CONTROL THE PROBLEM, is a sign that people may be mistaking correlation for causation.
Thanks for the update Will. I'm sure there are studies regarding the YAE/Verducci effect. But when I google it I get 22,000 plus hits and almost all of them seem to be articles describing the effect and saying how good it is and listing the next season's "victims." I did find this one by our very own Eric Seidman (I refer to him as Our because he's a fellow Phillies fan):
Anyways, here's something I agree with 100 percent:
"If it’s there or not there at a deeper statistical level — and I’ll be honest, I don’t have a firm grasp on Jeremy’s methodology — it’s great that we’re looking at it. I think on a more superficial level, it’s still a nice rule of thumb. I know there are others looking at it, but I do want to point out that I think the stats are much less important than getting to the root of the problem: how do teams effectively manage the health and workload of starting pitchers?"
A bunch of people have mentioned Hamels, and I would like to point out that while many of his peripherals didn't change, he suffered a bunch of minor freak injuries early in the season and couldn't break 90 on the radar gun until mid-May. He also threw 50+ less innings total this year.
Once again, I think people are trying to make the Verducci effect too specific. It's a rule of thumb, not a hard and fast rule. Think of it as nothing more than a red light on one of the Team Health Reports.
Has anyone studied warm-up pitches or practice pitches or whatever?
A pitcher might only pitch 50 innings in a season, but if he throws 20,000 practice pitches on his off-days, then over the course of the season his arm would have been worked harder than a guy who pitched 200 innings in games but didn't throw very much at all on his off-days.
Yes, they're meaningless. Frank Jobe did a study years ago that showed the ligaments didn't take stress until over 75% effort. There have been other confirming studies since then. Warm up throws, long toss, etc, only slightly add to ligament stress, but add to muscular fatigue.
62 comments have been left for this post.
Will/Kevin, can BP do an annual article on VE pitchers? The at-risks could be examined for the coming season, and those that were red-flagged the year before could be revisited.
Of the 10 Verducci highlighted, I'm particularly curious about Porcello, as he the Tigers seemed to measure him as much as possible. He also seemed to be asked to be a GB pitcher instead of power pitcher, which may be his more natural setting.
I kind of put it in the THRs.