FG - Brandon Morrow's LOB Blues
An interesting article about Morrow's issues with men on base.
4 months ago
Frag
38 comments
0 recs |
Email
PrintComments
Here’s an article a few days earlier from Fangraphs basically looking at the same thing.
"We are all agreed that your theory is crazy. The question that divides us is whether it is crazy enough to have a chance of being correct."
- Niels Bohr
Sorry, unauthorized hotlinking of copyrighted material not permitted.
I think I've figured out his problem for 2011 (maybe)
He did have a much more “predictable” repertoire in 2011, with ~95% of his pitches being FB and SL. Also, I think he started pitching more up-in-the-zone with men on base than with bases empty, as his FB% jumps nearly 10% as men get on base. More HRs with men on base = lower LOB%, intuitively.
The problem with this theory is his 2010 numbers. Much more diverse repertoire, and FB% remained relatively constant as men got on base, but his BABIP w/men on base was ridiculously high.
Therefore, I have no clue what’s wrong with him.
His 2011 wRC+ is 26
something else that's strange
his FB% jumped while his LD% + GB% decreased with men on base in 2011, yet his BABIP still went up by .060. Shouldn’t that be the opposite?
His 2011 wRC+ is 26
GB is positively correlated with BABIP
not by a tonne, but more positively correlated than FB. also, there’s a lot of noise in there
I know
so shouldn’t the decrease of GB% usually go along with a decrease of BABIP?
His 2011 wRC+ is 26
Hey
Got a question for you. Morrow was pretty bad (unlucky?) BABIP-wise in 2010. It continued in 2011, but to a much lesser degree. I think it was ~.400 in 2010 and ~.350 in 2011. However, he induced much more GB in 2010 w/MOB than in 2011. Does the difference in batted ball profile sort of account for the difference in BABIP, and could it mean that he was equally bad in 2010 and 2011 w/MOB, just in a different fashion?
His 2011 wRC+ is 26
So, I get the following #s from Fangrpahs with MOB:
2010: 18.0% LD%, 38.3% GB%, 43.7% FB%, 9.6 IFFB% and 9.6 HB/FB% – .380 BABIP
2011: 19.4 LD%, 33.2 GB%, 47.4 FB%, 6.5 IFFB% and 9.7 HR/FB% – .335 BABIP.
So basically, as you say, the biggest difference is 5% less GB and basically a 4% in FB% and 1% increase in LD% from 2010 to 2011. Of those FB, Morrow basically gave up the proprtion of HR, but a lower % of IFFB, which basically have a BABIP of 0.
I ran there numbers through my basic BABIP calculator, where I use the following BABIPs for each batted ball type: LD 0.720, GB .235, IFFB .000, oFB (outfield fly ball) ~0.160. These are approximate, but work okay as a simple estimate, which is all I ever want anyway. So, just applying this at a high level for Morrow, on that 1% extra LD%, we’d expect an extra 0.500 over the GB, or roughly 0.5% overall. On the 4% FB, we’d expect a decrease of something like 0.100, or about 0.4% total. Which would basically cancel out.
And that’s what we see. I get an expected BABIP of .286 for 2010 (.094 difference) and .296 for 2011 (0.039 difference). So either the luck is terrible, or he’s giving up really, really sharp hits. As a point of comparison, I ran his overall numbers through my calculator and got an expected BABIP of .283 for 2010 (.342 actual) and .312 for 2011 (.299 actual). So overall in 2010, there was a huge gap, in 2011 it was close to the xBABIP. We see this to a lesser extent in the MOB, but both numbers are higher than expected. For the entire league last year, my calculators gives a BABIP for .299 vs. .295 actual, just to give you a idea that it’s pretty close.
Yeah
I recall reading that, and there was some good discussion there. What I find quite frustrating is that it feels like for all the digital ink spilled, we’re no close to an explanation. Fatigue, reliance on two pitches, problems out of the stretch, poor outfield defence – all of these feel plausible, there’s just not enough data to disentangle a lot of these effects. Actually, I’m relatively comfortable saying it’s not the outfield defence, as his BABIP, given the batted ball profile, was right in line if not slightly lower than what would be expected. Though it probably didn’t help much either given the guys who were ru out there for sigificant portions of last year.
I think one thing is clear though – when the ball was put in play this year, he gave up much harder contact. In particular, the LD% increased from around league average of 17.8% to 22.4%, meaning it increased about 20%. And though I wouldn’t ead too much into it, the HR/FB rate increased significantly too. This could just be rergession towards average, but it could also be indictative of better contact (we expect high K guys to give up less sharp contact). This seems consistent with what Bruce Walton said to Tom:
His first year we went through some walk issues, base on balls, command issues. We pretty much ironed it out. We got to the point where now we left a lot of balls in the middle of the plate.Though xFIP really doesn’t show a difference, it completely ignores BABIP and HR/FB differences, and FIP the latter. But in this case, there’s a reason for a higher BABIP. As hugo noted his piece, the tRA went up accordingly – it’s a stat I don’t think gets enough appreciation, though the understanding of batted balls and averages on them is in its infancy. So I think it’s safe to say, that was 2010 was driven by an unsustainable BABIP, 2011 was much less so.
I also find the pitch mix explanation compelling, and to take a step forward I think it’s critical for him to get that third average pitch that he consistently relies on to keep batters honest.
I also find the pitch mix explanation compelling, and to take a step forward I think it’s critical for him to get that third average pitch that he consistently relies on to keep batters honest.
Look at his 2010 though; he used a lot of different pitches, still had bad BABIP & LOB%.
His 2011 wRC+ is 26
In 2010
He had a much higher BABIP (.342) than xBABIP (.283 – using my calculator). This suggests a lot of bad luck.
In 2011, his BABIP (.299) was close to in-line, actually slightly lower than his xBABIP (.312). This suggests the higher BABIP was deserved, and there wasn’t much luck driving it.
LOB% is strongly correlated with BABIP
Also, as per the article
He had even less of a repertoire when runners were on base, only throwing his fastball and slider, with a much higher % of sliders than without runners on base.
by Playoffs!!!!1 on Jan 28, 2012 10:05 PM EST up reply actions
but look at his BABIPs with men on base
it was much higher in 2010 than it was in 2011, yet his LOB was higher in 2011.
His 2011 wRC+ is 26
I looked into it
I’m wary of trusting the LOB splits for bases empty/men on base/scoring position, because if you look at Morrow for 2011, he has scoring position LOB of -15.1%, which is illogical and for 2010 it’s 1.5%, which intuitively feels inconceivably low. So I’m not sure the way those numbers are being separated out is good enough to rely on those numbers.
But maybe those problems are exclusive to scoring position splits only – they are a subset of men on base, whereas men on base and bases empty are mutally exclusive sets which sum to 100% (Morrow faced 777 batters in 2011, 449 with bases empty and 328 with men on base.) So as a second test, I took a weighted average of the bases empty LOB (100%) and men on base LOB (21.1%, side note, this should be called MOB LOB which sounds a lot cooler) and got 66.1%, which is close to the reported overall 65.5%. So maybe it’s alright. In which case, I’d argue it’s probably just a small sample size thing. We’re talking about relatively small numbers of batters faced. The other thing is, HRs with men on base would hurt the LOB rate but not affect BABIP. Using B-R’s HR Log, in 2010, Morrow gave up 11 HR, with a total of 9 men on base. In 2011, he gave up 21 HR with a total of 12 men on base. So it first instance this doesn’t seem to the case, but I’d need to think it through a little more before making a conclusion.
homeruns shouldn't affect strand-rate
if you look at how the statistic is calculated, it takes into account that homeruns are worth more than 1 run scored (I think it uses 1.4 runs / hr). As a sidenote, for a while last year, Tyler Clippard’s strand-rate was higher than 100%
"Look at me! I'm Tomokazu Ohka of the Montreal Expos!"
Right
So it sounds like there’s some llinear weights (I’ve never looked at the mechanics of calclated LOB%)…which is fine, but then I would hesitate to draw conclusions about the effect BABIP would have in small samples and splits
Yes
strand-rate makes no sense in splits — zero runners on means that you would have no runners to strand (or allow to score).
Giving up a lot of solo homeruns makes a pitcher look like he’s stranded more runners because the calculation credits more than just the one run to the homer (and, thus, fewer runs to scoring from the basepaths).
The whole point of strand-rate is to isolate the effects of sequencing from BABIP. For a long time, people were just looking at BABIP and thinking “this pitcher wasn’t unlucky, he had a BABIP of .300, he just sucked” — by looking at strand-rate, we can determine, “Sure, his BABIP was normal but his sequencing was out of whack.”
Personally, I generally subscribe to the notion that the only real control that pitchers should have over strand-rate is related to their strikeout-rate, though I think there may be a seemingly-strange correlation with walk-rate.
I’d like to look into whether pitchers with higher walk-rates have higher strand-rates — I think they should because walks give a batter only one base and aren’t generally run-scoring plays. Thus, walks are both somewhat less likely to result in scoring baserunners (hits can be doubles or triples) and much less likely to drive in runners already on base. It should be a relatively simple model but I just haven’t run it yet.
"Look at me! I'm Tomokazu Ohka of the Montreal Expos!"
The whole point of strand-rate is to isolate the effects of sequencing from BABIP.
But, is it really doing this? That is, is strand rate independent of BABIP. I believe when you say strand rate, you mean LOB% (Fangraphs uses them both in their glossary entry for LOB%). From a theoretical perspective, it would seem to me that BABIP does matter (looking at the formula), though I haven’t fully though it out.
I pulled the data from Fangraphs for all qualified pitcher in 2011 (n=94). I ran a regression with BABIP as the sole explanatory variable, and got a highly significant t-stat of -5.97, and BABIP explained 23% of the variation in LOB%. So it would seem they’re not actually independent, and LOB% is not just telling us about sequencing.
Right, babip and strand are not independent of one another
The point I’m making is that looking at a pitcher’s strand-rate and BABIP together allows us to see if an inflated ERA is the result of more hittable pitches (when low strand-rates co-occur with a high BABIP) or poor sequencing (when low strand-rates do not co-occur with a high BABIP). By the way, I ran the numbers on strand vs. BABIP several months ago and found the same thing you just did (http://www.bluebirdbanter.com/2011/9/12/2418213/im-waiting-for-my-man-what-factors-into-a-pitchers-strand-rate).
What I want to do soon is use a better statistical program, multiple regressions, and relative importance estimators to really get at the nitty-gritty.
"Look at me! I'm Tomokazu Ohka of the Montreal Expos!"
Ok, gotcha
The link didn;t work, but I found ti from the archives. Don’t recall reading it back then, but I wasn’t a member then and much less frequent reader.
Right, in Morrow’s case we have a normal BABIP but low strand rate. Probably is poor sequencing, which is basically just bad luck, which would be good in terms of projecting good forward. Similar to Cliff Lee in 2010, though of course it made his season look merely great instead of reflecting his ethereal peripherals.
Just returning to your initial post, it would make sense that BB could affect the stand rate for the reason you stated, to me at least.
It all depends on how you look at the data
I do think Morrow’s problems in 2011 were also likely driven by a BABIP higher than it should have been
Bases empty: .273
Runners on: 335
Runners in scoring position: .355
Yet his K- and bb- rates were almost exactly the same with runners on. And, counter to the theory that what we’re seeing is real and should be a cause for a concern, Brandon’s LD-rate was actually lower with runners on (19.4%) vs. the bases empty (24.5%).
This is the opposite of what we should be seeing. In fact, there is probably a lot more evidence that Brandon’s elevated ERA (as compared with xFIP) was legitimate — he had a very high BABIP against over 600 batters. The high BABIP which led to Brandon’s inflated ERA in 2011 occurred over half as many at-bats, which, to me, suggests random variation.
I agree that Brandon needs a third pitch but not so much because his strand-rate last year was low as much as because I think batters will continue to adjust to him unless (until?) he develops one. I think, more often than not, if Brandon could repeat last season, he’d come out looking fine. However, I think that, unless he uses another pitch, he won’t be able to repeat last season.
Also, a third pitch could enable him to induce grounders, helping him shorten innings. He won’t be a true frontline starter if it takes him 20 pitches per inning.
"Look at me! I'm Tomokazu Ohka of the Montreal Expos!"
To be honest, I don't think it's so much how one looks at the data, as the limits of the data itself
Specifically, I think there’s a lot of theories at play, and we’re using split season samples looking at stats like BABIP, which is a volatile stat even in a full season sample, much less these smaller split level samples. To illustrate this, I’m going to reproduce the data you put above, with my calculated xBABIP based on batted balls in brackets to try and seperate “luck” driven BABIP differentials (relative to average) from “earned” BABIP differentials driven by batted balls (and wherever else I put data in brackets assume it’s my calculated xBABIP unless otherwise specified):
2011 Bases empty: .273 (.324) – 449 Total Batters Faced
2011 Men on base: .335 (.296) – 328 TBF
2011 Scoring Position: .355 (.296) – 186 TBF
2010 Bases empty: .312 (.280) – 348 TBF
2010 Men on base: .380 (.286) – 281 TBF
2010 Scoring Position: .411 (.273) – 158 TBF
So in 2010, we see that Morrow’s BABIP-xBABIP differential was realtively modest with no one on, and quite large with runners on (very large for scoring position, but the sample there is a subset of men on base, and so smaller, so I won’t comment further. Suffice to say, though, that if we backed out runners on base but not in scoring position – so basically, just runner on first – we’d get results that look more like bases empty than men on). In 2011, Morrow outperformed with bases empty, but underperformed with runners on similar to 2010, with smaller absolute magnitudes as in 2011, though similar magnitudes of underperformance relative to with no runners on.
Let’s combine those two years to try and build a more robust sample to filter out some noise:
2010-11 Bases empty: .290 (.304) – 797 TBF
2010-11 Men on base: .356 (.291) – 609 TBF
2010-11 Scoring Position: .381 (.285) – 344 TBF
So what we see is that with no one on, the BABIP is basically normal, relative to both league average and xBABIP based on batted balls. However, with men on base we see a big differential between xBABIP and BABIP. What is this? Bad luck? Some misclassification of batted ball (implicit would be miscalssification unique to batted balls with men on base, so I;m very skeptical)? Or is it possible that there’s something at work here, whereby the balls put in play, whatever their type, are just getting smashed, and the result is a much higher BABIP than normal. I think this may be the case, and I think it may come back to the lack of a third pitch. But I want to look into more data before expounding on this, as it’s just a hunch. Not to mention, as you note, the K% and BB% are similar, which means we need an explanation that explains the batted ball difference, but also why the K% and BB% are similar
One last thing to come back to. I alluded to the fact that the scoring position splits were even worse than the men on base splits, of which they are a subset. Backing out the former from the latter, we can see the numbers for when there are just runners on first (2010-11 numbers to get a more meanignful sample):
2010-11 Men on base: .356 (.291) – 609 TBF
2010-11 Scoring Position: .381 (.285) – 344 TBF
2010-11 Mon on first: .323 (.299) – 265 TBF
So we see it’s worse than no one on, but much better than with runners in scoring position.
Yep, I'm more convinced it has to do with pitch mix
I’ve looked into more of the data, and I’m becoming evermore convinced a signficiant part of Morrow’s problems is not having a third good pitch, or at least one he is confident in throwing. However, I’m no closer to explaining some of those weird base empty/man on splits above.
At a high level, we know he relies on fastball/slider. We also generally know that sliders tend to play up to same-handed hitters, and they are easier for opposite handed hitters to hit (I believe this is because they break in on the hands of same handed hitters). So whereas against righties the slider gives him a huge weapon to pair with the fastball, essentially two plus/plus plus pitches; against lefties, he’s left more exposed and really only has one plus/plus plus pitch, so batters will either be able to manage the slider, or sit fastball (more on this relationship and how I think it works below).
The first hint was the composition of the line-ups Morrow faced. In 2010, he faced 60% LHB and in 2011 56% LHB, whereas the RHP as in the 2010 faced about 51% LHB in both thouse years. So line-ups were being stacked with lefties everymore so against Morrow than the typical lefty. Presumably, opposing team think lefties can particularly take advatage of Morrow.
At first glance, looking into the stats, we don’t see a huge
2010 LHB: 27.5 K%, 13.1 BB%, .336 BABIP, 3.40FIP, 3.87xFIP
2010 RHB: 29.5 K%, 6.7 BB%, .351 BABIP, 2.81FIP, 2.91xFIP
2011 LHB: 29.1 K%, 8.7%, .296 BABIP, 2.89FIP, 3.06xFIP
2011 RHB: 22.4%, 9.1 BB%, .302 BABIP, 4.64FIP, 4.16xFIP
If we look at this, it appears Morrow has significant peripheral splits last year, and this year somewhat reverse splits. I imagine part of this is pitching more in the zone in 2011, as Bruce Walton implies in Tom;s interview, but I don’t know why the peripherals against righties declined from 2010-2011. In any event, I think this is separate, as I want to look more into the batted balls. There is not a big difference in the BABIPs, but I noticed the LD% is drastically difference. if we break BABIP down by batter type and year, again against my xBABIP, we get:
2010 LHB: .336 (.291) – 437 TBF
2010 RHB: .351 (.269) – 254 TBF
2011 LHB: .296 (.328) – 375 TBF
2011 RHB: .302 (.293) – 340 TBF
2010-11 LHB: .314 (.311) – 812 TBF
2010-11 RHB: .323 (.283) – 594 TBF
There’s a lot of noise in the differentials, but if we combine them, we find that for LHB, Morrow has a high BABIP, and it closely relfects the an xBABIP based on his batted balls – he gives up a lot of line drives. Against righties (a somewhat smaller sample), we has a higher BABIP, but we would expect a lower BABIP. Keep in mind that as a fly ball pitcher, we’d expect a lower than league average BABIP, so the .283 epxectation against righties seems in line, whereas the xBABIP against lefties is quite high, which relfects the aforementioned line drives.
Next, I went to the PITCHf/x data from Joe Lefkowitz. For 2011, Morrow used his fastball 59% of the time against RHP, and 61% against lefties. His slider, however, is 34% against RHB, and only 24% against LHP. However, the swing and miss is similar – 36% against RHB, 32.2 against LHB on sliders, and roughly 20% against both on dastball, which explains why the peripherals are roughly in line.
In other words, in my view Morrow faces a trade-off against LHB. If he uses his slider the same amount, batters will catch on, and it will lose effectiveness. So he cuts back its use, but that means he throws the fastball more, since he doesn’t trust his other pitches as average or better. So LHB sit on the fastball. It’s still really good, so he gets swing and misses, but when he misses with one, it gets smashed. And, if we look at the LD% on fastballs, against LHB it’s 29%, against RHB its 17%. And where does that come from? The ground ball rate falls from 35% to 22.4%, almost exactly the difference.
What about 2010? Well, he used sliders a lot less to both hands. In 2010, he apparently used a change-up about 16% of the time against LHB, which got a lot of ground balls and few line drives, but few swings and misses. He also used a curve ball about 13% of the time. My guess is, he had trouble locating these, and this is why he dumped it in 2011, and the BB% went down. The curveball appeared very effective against righties, so the only reason to dump it would be a command problem. Again in 2010, he has a big split in the LD% on his fastball – 17% vs. 25%, but similar swing and miss%. So LHB would sit fastball, take some lumps and punish mistake fastballs.
This site allows one to download the entire data, so at some point in the future I’m going to do that and try and nail more of this down – particularly seeing if there’s some interaction between the handedness and the bases empty/men on base splits, because this doesn’t tell us a lot. There’s some data on the pitch selection with bases empty/men on/RISP, and he uses the fastball a little bit less and the slider more while cutting back on other pitches, but there’s not a very strong pattern, and it’s not entirely consistent from 2010 to 2011, at least prima facie. I’m wondering if what he does is different to lefties and righties.
Anyway, to sum up: I definitely think he needs a third pitch, especially to lefties. It sounds like the cutter is what they’re going to try, but failing that, I think to take the next step he needs to try to develop the curve or change-up. It doesn’t have to be great, but he needs to be able to reasonably command it to keep hitters off balance.
If anyone reads this
Please excuse the number of mistakes. I forgot to preview and edit before hitting post.
The first hint was the composition of the line-ups Morrow faced. In 2010, he faced 60% LHB and in 2011 56% LHB, whereas the average AL RHPas in the 2010faced about 51% LHB in both those years.
Against righties (a somewhat smaller sample),wehe has a higher BABIP
I'm not sure how you've reached this conclusion
I definitely think he needs a third pitch, especially to lefties
from these data:
2011 LHB: 29.1 K%, 8.7%, .296 BABIP, 2.89FIP, 3.06xFIP
2011 LHB: .296 (.328) – 375 TBF
Morrow was pretty great against lefthanded batters last season, possibly because (as hugo’s article said) he mixed in the change and curve slightly more against them.
As long as the hypothesis that unless hitters swing and miss, they crush the ball is concerned, it’s basically conjecture without hit f/x data. We’re talking about a few hundred plate appearances in 2011 and different issues in 2010. I really don’t think we have enough to go on here. As I said before, I do think Brandon needs to be able to rely on a third pitch, not because he hasn’t been successful in the past (theoretical success, not results-based success) but because I don’t think he can maintain that kind of success in the future with just the fastball and slider.
"Look at me! I'm Tomokazu Ohka of the Montreal Expos!"
Apologies for the belated nature of this reply
Morrow was indeed a great three-true-outcomes pitcher in 2011 against lefties (and against righties, and in 2010 against righties – not so much against lefties). But, those three outcomes amount to just under 40% of his PAs, and the whole reason we’re discussing this in the first place is because those other 60% are causing his results to diverge from what we would expect based on the peripherals.
As you note, the 2011 BABIP against lefties was not out of the ordinary, at least at first glance. But given a 26%+ line drive rate, the BABIP isn’t going to stay there, which is why I was focussing on xBABIP to try and seperate out, as you call it, theoretical success (or lack thereof) and results-based success. And I agree that ultimately, we would need the hitf/x tool to make determinative conclusions, but we don’t have that and from what I’ve heard won’t be getting it anytime soon, so we’re stuck with the data points we have, and to make of them what we can.
So, from what do I draw the inteference (I think conclusion is too strong given the limits of the data, and maybe I got a little carried away with the strength of my conviction)? I see the following points:
1) A 2011 LD rate of 26.2% against lefties, vs. 17.8% against righties. In 2010, also a split, 19.4% against 15.5%.
2) Morrow has big platoon splits in the LD rate off his fastball in both years. Yet, he used the fastball 60% of the time in 2011. His slider did not have big platoon splits at all – not in the batted ball profile, not the swinging strike rate, not in the swing rate. Just the usage. It would seem to me lefties are sitting fastball, though I want to dig into this more..
3) In 2010, Morrow had lesser peripherals against lefties, when he was using a curveball and change-up more (though, I read that what pitchf/x identifies a sinking fastball might actually be the change-up, because he throws it quickly, so the change in change-up use might not have actually been that big). In 2011, the peripherals improved, yet the batted ball got worse, once we focus on xBABIP rather than BABIP. This is correlation, not causation, but when we fit some of these pieces together, it seems to me a picture emerges.
That said, I probably overstated the case a little bit and was not careful with the language I used (it was fairly late at night). Since I wrote the above, I’ve thought of a few ways of testing the batter-by-batter data that might confirm or discredit the above thoughts about how those points fit into my hypothesis. Don’t know when I’ll get around to that, or what it yields, but if there’s anything interesting I’ll toss a FanPost together or soemthing. Also, I still feel like some of the anomolous things we see in Morrow’s nubers are unexplained (why the poorer results against righties in 2011), so I’d like to see if there’s anything there that can be explained.
The problem here is one of circular reasoning
My point is that I don’t think pitchers (including Morrow) have much control over their BABIP (aside from being able to keep it slightly lower by giving up more flyballs and pop-outs) strictly because I don’t think pitchers are able to suppress linedrives over the long-term.
Look at the leaders in LD-rate last season . . . I haven’t run the numbers but it looks like any correlation between being good and suppressing hits is weak at best (also see fangraphs: http://www.fangraphs.com/blogs/index.php/expected-babip-for-pitchers/ )
"Look at me! I'm Tomokazu Ohka of the Montreal Expos!"
Alright
I see where you’re coming from now, and given that, I understand your skepticism. I hadn’t read that Fangrpahs article, it was about two years earlier than when I began reading Fangraphs and don’t think I’ve seen it linked before. That said, the article doesn’t really provide any evidence, so much as th last two paragraphs posit as fact that LD% and thus BABIP regress towards league mean. I intend to take a look into this more (by that I mean, survey the material out there and then maybe look at some data).
Intuitively, I’m skeptical that pitchers have so little effect on their LD% to the point where we basically say there’s none. That said, I haven;t really looked at at. My xBABIP calculator was just a toy I built last year to look at hitters and dump in thier batted balls to get an idea if early season BABIPs were sustainable based on batted balls. Nonetheless, the way I was using it above was exactly what the FG article was suggesting – “The best way to use this statistic is to attempt to validate a pitcher’s current BABIP. For instance, a pitcher might have an high line drive percentage and a high BABIP” – since Morrow has some funny numbers in this regard, especially last year against lefties.
Morrow has now had two seasons in which he has significantly underperformed his peripherals, but interestingly, it appears for two different reasons – in 2010, an abnormally high BABIP (unsupported by xBABIP) and in 2011 sequencing, as we discussed above. THe best case scenario is that this was just really, really bad luck, and going forward we can expect results in line with peripherals. Which would be fantastic, because no way does he come as cheap as he did on his contract without that. Worst case, there’s something driving the underperformance yet to be fully identified, which may be more persistent. In this vein, I still think there might be something to the pitch mix hypothesis.
Anyway, thanks for the link, and the discussion.
yep, it's certainly possible that Morrow should underperform his peripherals
but I think that requires some belief that he was lucky in either sequencing in 2010 or overall BABIP in 2011.
It would be nice to see him mix his pitches better going forward because it’s going to be harder and harder for him maintain the even-peripherals-based success that he’s had the past two years and it would be nice to see him pitch a little more efficiently (which I think mixing in pitches and getting grounders earlier in counts would do).
The discussion has been fun, I’ve really enjoyed your work so far (even if I haven’t had the time to comment on it as much as I’d like) and please keep it up.
"Look at me! I'm Tomokazu Ohka of the Montreal Expos!"
Timely Addendum
So the other day there was an article at Fangraphs looking at Josh Johnson and how he added a curveball to his arsenal in 2011. I don’t know if you saw it, but here’s the link.
I found it interesting for another reason: Josh Johnson is a very successful RHP with a plus fastbal land slider. Unlike Morrow, he also has and uses a good changeup, which gives him the tradional three plus pitch arsenal needed for start success. And obviously, the results are good.
In the article, there’s a table showing Johnson’s pitch usage in different situations. Like Morrow, he throws a lot of fastballs and sliders. In fact, against RHP (in 2010), he’s essentially a two pitch pitcher: fastball and slider, as Morrow was in 2011 (in 2011 Johnson used no change-up, but mixed in the curve a little bit). But against lefties, against whom the slider is less effective, he mixes in the change-up to keep hitters honest. He still uses the slider against LHB frequently when ahead in the count, but rarely when behind in the count. So this got me thinking, what does it look like for Morrow?
I’ve replicated the chart for Morrow in 2010 and 2011 below, using pitchf/x data from Joe Lefkowitz’s site. I also added the splits for when there are 2 strikes, and with RISP. The chart is too wide to fit in this thread, so I’ve attached in a seperate reply to an above thread such that it appears below. One important note: Morrow throws a pitch that pitchf/x classifies as a split-fingered fastball. Because Morrow throws so fast, I’m pretty sure this is actually his change-up. For 2011, Morrow had no change-ups other than this, and regardless, didn’t throw many. In 2010, there were some pitches actually classified as change-ups (10%), and roughly an equal number of splitters for a total of about 20%. Some of those splitters may actually have been fastballs, so treat the number of change-ups in 2010 with a grain of salt.
Morrow's 201-2011 pitch usage
A couple notes, as the results are pretty clear, by some thing that jump out to me:
- In 2010, the fastball is pretty consistent between L/R regardless of the count. In 2011, it’s overall much higher, but he used it a ton when behind in the count and he used it more against lefties almost regardless of the situation.
- The slider usage is much smaller than 2011, and smaller than Josh Johnson used, but Morrow’s split is similar to Johnson’s. Morrow did not use his slider against lefties unless he was ahead or had 2 strikes, so basically just as a finishing pitch. In 2011, he ramped up the usage, but particularly so against RHB and LHB when ahead or to finish off a batter.
- The change-up usage went way down. In 2010, he used it a lot against lefties, and particularly when behind in the count. In 2011, hardly used at all.
- Ditto the curveball use, though this is primarily a RHB pitch, and used against lefties when he was ahead.
- Taking all this together, and the fact that his peripherals were better in 2011 against LHB and slightly worse against RHB, I make the inference that had trouble commanding the change-up and throwing it for strikes. When he took it out in 2011, his command improves, but he was essentially left with the fastball, which LHB sat on and waited for mistakes to hammer (hence the high LD%).
- He also shelved the curveball, maybe due to command problems as well, maybe because he felt he just didn’t need it.
I think this tends to support the proposition that Morrow needs a third pitch, whatever it be, if only to keep hitters off balance, particularly LHB. I should also note that the cutter Morrow introduced the last couple starts in 2011 show up as a pretty distinct cloud (when you look at horizontal break v. vertical break and speed) in those starts, but pitchf/x classified them as sliders because they didn;t see enough of them to classify them separately. I intend to look further into this, particularly how batters succeeded in some of these splits (when he threw 78% fastballs to LHBs, is that where they hammered him)?
Very nice work
your hypothesis that lefthanded batters would sit and wait on the fastball is intriguing and you could be onto something — the results in the table may support that hypothesis.
On the other hand, generating causal links (particularly for something that seems like it may be as random as babip or linedrive-rate) is tough. It’s not like there’s a big cluster of two- and three-pitch pitchers who gave up the most linedrives in baseball last season. To be honest, what concerns me with Morrow isn’t the linedrive-rate, it’s the flyball-rate. Hopefully the addition of the cutter (as you mentioned above) helps him bring that down.
Great work and thanks!
"Look at me! I'm Tomokazu Ohka of the Montreal Expos!"
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
Baseball Nation Recent Stories
In Association With
Big League Stew - MLB - Yahoo! Sports
- Giancarlo Stanton misses cutoff man, and Logan Morrison is thankful he did (Video)
- Snoop Dogg delivers memorable first pitch at U.S. Cellular Field
- The Juice: Mesoraco’s grand slam helps Reds complete four-game sweep of Braves
- Please hold: Indians’ Joe Smith jumps the gun on relief appearance
- Local radio host honors Cubs’ Tony Campana with parody song
