May 9, 2013 at 8:52 pm by Franklin Rabon under Atlanta Braves
A lot of hands have been wrung throughout Braves fandom recently over the surprising struggles of once dominant closer Craig Kimbrel. He’s gone from being not just automatic, but completely demoralizing to the other team, to Durbin-esque in recent weeks, giving up key home runs repeatedly.
This has lead many Braves fans to speculate on what’s wrong with Kimbrel. I’ve seen repeated mentions of his velocity being down (possibly due to an undiagnosed injury, because Twitter doctors can smell injuries from 200 miles away), over-reliance on his fastball, different release point, or just plain old lack of mental toughness.
Let’s first address the physical claims, because most of them are easier to dismiss, because they’re just not factually true.
Here are Kimbrel’s careeer velocity, release point and movement numbers, courtesy of Brooksbaseball.net:
and here are his numbers for 2013:
We see that Kimbrel has ‘lost’ 0.1 MPH on his fastball. Given the sample size in 2013, this is even inside the possibility this is even just due to inaccuracy in radar guns. 0.1 MPH isn’t even remotely close to statistically significant, let alone being practically different from a ‘difficulty of hitting the baseball’ standpoint. Attributing Kimbrel’s struggles to loss in velocity is simply factually incorrect, and the product of lazy analysis by those who are simply unwilling to actually research their opinions before spewing them. I was recently engaged on twitter by a fan who said “it’s not up for debate that he’s lost velocity, I watch every game, it’s clearly 2-3 MPH less.” While he was right, that it’s not up for debate, it was because he was so factually incorrect that it’s not up for debate that he was wrong; this clearly shows why the common rebuttal to most analysis of “do you even watch the games?” is a non-starter. When we watch the games, as humans, we’re extremely prone to all sorts of known cognitive biases. I do believe this follower wasn’t being disingenuous, and actually did believe Kimbrel’s velocity was down, but due to confirmation bias, he was simply wrong about that factual matter. I watch every game, either from my seat on the 10th row along the first base side of the infield, where I have a pretty good view of what’s going on, or from my TV, where I have access to slow motion replay, and often times I do BOTH, thanks to MLB.TV replay. Even as such, I still don’t trust my eyes and memories when it comes to matters that can easily enough be looked up given the vast resources we have at our disposal today.
Looking at usage charts, we see that Kimbrel’s fastball usage is up by 8% (and since he’s a two pitch pitcher, his curve usage is down 8%). One line of argumentation goes that Kimbrel has over-relied on his fastball, allowing hitters to ‘tee off on it’ because they’re seeing it so much more. Simply put, an additional 8% of the time over such a small sample simply isn’t enough of a difference for that to be the case. Consider an at bat, this comes out to less than one additional fastball per at bat, which simply isn’t enough of a difference for a hitter to alter the way he approaches facing Kimbrel. Further, since Kimbrel will only face a hitter once per game, it’s not like seeing those additional fastballs allows hitters to better track the pitch, like might be the case with a starting pitcher. Even at that, given the nastiness of Kimbrel’s stuff, it isn’t clear that it would matter anyway. Steve Carlton was a starting pitcher with essentially the same two pitches, and it didn’t really matter. Carlton even always threw his pitches to the same two locations (fastballs up and in and sliders low and away).
Further, his pitches aren’t moving in a statistically different way. While his point of release has changed a bit (he’s releasing the ball from a slightly lower point), that should show up in contact rates (which haven’t really changed) more-so than HR/FB rates.
Another claim is that Kimbrel isn’t throwing the high fastball as much, and that the low fastball is easier to hit out. First, this defies all common baseball knowledge. Common baseball knowledge says that low fastballs are harder to hit out, while high fastballs are easier to hit out, but harder to make contact with. And while we can’t always trust ‘common baseball knowledge’ we’d at least need some sort of statistical evidence before we went around challenging it. And while there is some evidence that high fastballs are harder to make contact with, there isn’t any evidence that it would change HR/FB ratio without changing contact rate first. Kimbrel isn’t having trouble missing bats, his trouble is coming from what happens when contact is made. Finally, the location change is barely even noticeable anyway:
Fastball Location for his Career:
Some difference, but not really a substantial one. Certainly not enough to explain a more than doubling of his HR/FB rate.
A point we’ve alluded to that we’ll now make explicit is what has changed with Kimbrel’s results? Strikeout, walk and other rates, again, aren’t statistically significantly different. What has changed is Kimbrel’s home run to fly ball ratio.
The HR/FB ratio jumps off the screen, while other rates are roughly in line with his career.
Simply put, HR/FB ratio is by far the most unstable rate for a pitcher of any statistic there is. Even for starting pitchers, it usually takes 2 or more full seasons to have any idea what a pitcher’s real HR/FB ratio really is. In Pizza Cutter’s landmark rate stabilization study, HR/FB ratio didn’t even come remotely close to stabilizing for starting pitchers over an entire season. The point is especially poignant for relievers. As Pizza Cutter said:
Now consider the sample and stat that we’re fretting so much about with Kimbrel, the sample is 52 PA. It typically takes close to 2000 PA for a pitcher’s HR/FB to begin to stabilize.
What we have to consider here is the “signal to noise” ratio. That is, how much of the results we’re seeing are due to a real change in skill, and how much are random variation. Given the sample size and variance of the statistic that’s giving Kimbrel so much trouble, we’re talking about 98.75% noise and 1.25% signal. That means that for the 52 plate appearances Kimbrel has faced hitters this season, HR/FB ratio is literally 99% random chance. Not a change in skill, not something the pitcher is doing differently, but purely random variation. We’re essentially trying to come up with explanations for a coin flipping contest when we try to explain why Kimbrel’s HR/FB ratio has jumped so much. While there might be some change in skill involved, it’s several hundred times more likely that it’s just pure dumb luck, that it’s not even really worth considering “what’s wrong with Kimbrel” at this point. It’s dramatically more likely he’s the same pitcher he’s always been, just running into small sample weirdness. Fans hate “small sample size weirdness” as an explanation, but this is largely why average fans aren’t professional MLB general managers.
Craig Kimbrel had an incredible season last year, but the problem is, even over a full year, a reliever’s performance is over a relatively small sample, and is thus prone, as pizza cutter stated, to an absurd amount of random variation. This is why it’s an absurd idea to pay any reliever a lot of money, even coming off a year like Kimbrel had last year. Because even if they do perform at an obscenely great level, you simply cannot depend on that type of performance from year to year.
Why I’m saying that the good news with Kimbrel is also the bad news is because while nothing is wrong with Kimbrel, that’s the point, nothing is wrong with him; what we’re seeing is simply what we should expect from him from time to time. You can never expect a reliever to put up a season like Kimbrel did last year, and we should expect down years where Kimbrel blows 7-10 saves more often than a year like last year, even if Kimbrel is one of the greatest relievers in baseball history. Further, with Kimbrel, even though he’s approaching 100 saves, he simply hasn’t pitched enough for us to even know what his actual rates should and will be once they start to stabilize. While we fully accept that a starter can have an anomolously good or bad year, we often fail to realize that over his entire career, Craig Kimbrel has pitched less than an ‘ace’ starter’s full season workload. It’s entirely possible that when Kimbrel’s HR/FB ratio does stabilize, it will be higher than the ~8% rate he’s seen over the first couple of years of his career. The bottom line is that we don’t really know if Kimbrel is an ‘all-time’ level talent or merely very good. It’s not only possible that his career won’t be as good as Billy Wagner’s, it’s actually extremely likely.
Finally, one point I’ll briefly address is the claim that the reason why Kimbrel’s ‘problem’ hasn’t shown up in any of the measures we’ve examined above is because it’s mental. There is a claim floating around out there that Kimbrel has simply got into a mental funk. I’m actually not one to immediately disqualify the mental aspect of the game. Players absolutely can go through mental funks that negatively impact their performance. However, for it to be real, it has to negatively impact their performance in a tangible way. What mental toughness does is makes a pitcher less likely to throw poor pitches. A lack of ‘closer mentality’ doesn’t however make good pitches magically become hittable. Kimbrel’s 96 MPH fastball with the exact same amount of movement didn’t magically become easier to hit out of the park because it contained less will to win. Rick Ankiel’s ‘intangibles’ were absolutely real, and caused an absolutely real inability to throw strikes. If you’re going to claim ‘intangibles’ you have to be able to show how those intangibles actually impact the game, you can’t use ‘intangible’ as if a magic wand that explains away everything that isn’t immediately apparent.
If Kimbrel had ‘a problem’ and getting back to his 2011-2012 type results were simply as easy as ‘fixing the problem’ it’d almost be less worrisome, the reality is that there’s nothing to fix. We should marvel at Kimbrel’s 2011-2012 results, but to expect we’ll see it year in-year out, or ever again at all, is simply folly. And most importantly of all, we absolutely shouldn’t pay Kimbrel down the road like we expect his 2011-2012 performance.
May 9, 2013 at 3:56 pm by Ben Duronio under Atlanta Braves
While most of the talk is about Craig Kimbrel’s struggles in the ninth inning over the past week — which I also opined on yesterday at RotoGraphs – there should also be some cause for concern with Eric O’Flaherty.
This is O’Flaherty’s final season before free agency, which most likely means it is his final season with the Braves. Much like Brian McCann, O’Flaherty’s position is not necessarily a position of need and the team will want to fill his spot with a cheaper and younger player. The sooner it is understood that O’Flaherty will be gone, the easier it will be for you to cope with if you are the kind of person who has difficulty watching players leave. It will happen, so let’s move on from that.
The problem with O’Flaherty hitting free agency, is that his strikeouts are dropping at a rapid rate and his velocity has dipped down to 90mph from his peak of right at 92mph. Is that a huge, fearful drop? Not quite, but I can assure teams have access to his velocity readings and the fact that it is continually dropping along with his strikeout rate (22.3% rate in 2011, 20% in 2012, 15.5% in 2013) will be a big caution sign for potential users of his services.
What it means for the Braves is that we probably should not expect O’Flaherty to bounce back with quite the same level of confidence that we have in Kimbrel. If I had to estimate, I think O’Flaherty’s ERA will only rise from its current level of 2.40 rather than drop — which is considerable when you look at his 0.98 and 1.73 marks from the past few years. He still gets himself a lot of ground balls which should lead to a better ERA than FIP and xFIP, but I doubt we see the O’Flaherty we saw for the past two seasons going forward. It would certainly be great to have that type of performance, but given his weakened velocity and lowered strikeout rate, I am not very confident in that happening.
May 9, 2013 at 9:00 am by Andrew Sisson under Atlanta Braves
The Braves continue on their 10 game road trip with a stop out west in San Francisco. Probables for the four game series are Teheran v. Vogelsong (10:15), Hudson v. Cain (10:15), Maholm v. Bumgarner (4:05) and Medlen v. Lincecum (4:05).
Even with the B squad lineup yesterday, the Braves were able to finish off the Reds, taking two of the three. Mike Minor had another great start going 7IP, 4H, 1ER, 3BB and 7K while throwing a career high 117 pitches. Dan Uggla, currently riding a six game hit streak, jacked out two homers giving him seven on the season. The struggling Uggla all of a sudden has a 110 wRC+, not bad.
One of the best parts of yesterdays game was when Juan Francisco took J.J. Hoover deep with the bases loaded. Not only did it essentially put the game out of reach, but as you probably know, these two were traded for each other at the beginning for last season. So far as a Brave, Francisco has played in 116 games posting a 1.1 fWAR. In his 284 appearances, he has put together a .247/.287/.449 triple slash, .313 wOBA and 97 wRC+. As a Red, Hoover has thrown 47.1 innings, accumulating 0.1 fWAR as a reliever. His ERA-/FIP-/xFIP- currently sit at 67/104/116, with a 24.0% strikeout rate and 10.7% walk rate. Overall the trade has worked out great so far, already a one win advantage. At the time of the trade, the Braves were able to give up an expendable bullpen arm in the minors for a player, with potential, at a position of need. To many fans, yesterday’s grand slam probably closed the book on their opinion of this trade.
FanGraphs ran a piece yesterday about Kimbrel’s implosion from Tuesday night. This has been highly discussed on Twitter, but for those who haven’t been around for that I’ll give my quick thoughts. There is nothing wrong with Kimbrel. His velocity is right where it’s always been, his strikeout rate is 40.4% and his walk rate is the lowest it’s ever been. Just remember, he struck the previous two hitters out before the two knee-high fastballs were sent deep. We preach it a lot on here, but sometimes random variation and fluctuation of results happen, it’s just the way things are and can probably best explain the past few outings. If you’re expecting to see 2012 Craig Kimbrel again, well don’t, you may never see that type of season again.
Jason Heyward will begin his rehab assignment today in Gwinnett.
May 8, 2013 at 11:00 am by Mark Smith under Atlanta Braves
I’ll get better at remembering to put this up after the games, I promise.
MVP: DAAAAAAAAN Uggla .272 WPA
LVP: Gattitude -.160
12:35 PM on Sports South and MLB Network
Mike Leake isn’t a terribly good pitcher, but relative to other 5th starters, he’s pretty good. He won’t strike out many hitters (16%) or walk many hitters (6%), but he gives up home runs like it’s going out of style (14.5% HR/FB – average is 10.6%). Leake is a Kitchen Sink Pitcher that tries to confuse the hitter by throwing so many pitches although none of them are particularly good. He’s a pretty blah pitcher, especially for a former top-10 pick, but he adds some hidden value with the stick, hitting .275/.306/.354 for his career, and is not an easy out at the plate.
It’s Mike Minor Day again! Considering the ballpark and the fact that both pitchers are a bit homer-prone, today’s game has a fairly decent chance of getting out of hand. The Braves are throwing out a “Get Away Day” lineup, however, so the Reds would have the advantage if Minor wasn’t so much better than Leake. It would be awesome to grab a series win tonight before the 7-game trip out West to San Francisco and Arizona.
May 7, 2013 at 5:30 pm by Mark Smith under Atlanta Braves
MVP: Devin Mesoraco .490 WPA, McCann .151
LVP: Craig Kimbrel -.813
7:10 PM on Sports South and MLB Network
Home Bailey is a pretty good pitcher. though he’s never quite lived up to his top-prospect pedigree. For his career, he has struck out 18% of batters while walking 8%, but over the past three seasons, the strikeouts have perked up while the walks have declined. Essentially, he has evolved as he has implemented his splitter, whose usage continues to rise. Bailey throws hard, but his fastball is generally fairly straight, and he’s never fully decided on a breaking ball that he can gets swings-and-misses with. 2013 has seen a jump in strikeouts, and while it’s a small sample, he is utilizing his splitter more than in the past, and it may be the secondary pitch he’s always needed. But overall, one can expect a lot of fastballs, a few more sliders than curves, and plenty of splitters for lefties.
Kris Medlen takes the mound looking to keep the Braves in the win column. Juan Francisco also returns to the lineup tonight to help battle Bailey, and Bailey and his fastballs seem like a good night for Juan to come back, as opposed to last night with Arroyo and his myriad breaking balls. This lineup is starting to thicken out again, and hopefully, Jason Heyward will be back early next week to finally put it all into place. For now, the Braves have another solid match-up for tonight and will hope for a few more long balls.
May 7, 2013 at 9:00 am by Andrew Sisson under Atlanta Braves
It was a good start to the road trip Monday with the Braves holding off the Reds, 7-4. The probables for the rest of the series are Medlen v. Bailey (7:10) and Minor v. Leake (12:35).
Brian McCann went 0-4, BB in his first game back. Of course, it may take a couple of games for McCann to get back into the flow of things and adjust to big league pitching, but having him back behind the plate was a comforting feeling. Whether this means anything or not, he appeared noticeably thinner from previous seasons.
Andrelton Simmons was the big story last night, popping out two home runs and continuing to show off his defensive wizardry in the field. All of a sudden, his batting line don’t look so terrible, and his 93 wRC+ is actually above league average for his position.
Jason Heyward took BP yesterday for the first time since surgery. According to DOB, he will likely take two more days of BP before going on a rehab assignment. His return date is still up in the air, something the club will monitor day-to-day. So far, so good.
Mike Viso of Big Leagues Magazine, had a great interview with Braves prospect Matt Lipka (you’ll have to scroll through a couple of pages to get to the interview). Lipka, because of where he was drafted out of high school, feels like he has been around forever making little progress through the system. Because of this, it is easy to throw the “bust” tag on him, however we have to remind ourself that he’s still pretty young. When I went to see Lynchburg play last week, I again had one of these moments, realizing he just turned 21 and by far the youngest player on the Hillcats roster.
The Mets/Braves rainout from this weekend has been rescheduled for June 18 as a split doubleheader.
May 6, 2013 at 5:30 pm by Mark Smith under Atlanta Braves
7:10 PM on Sports South and ESPN
Great American Ballpark
Great American Ballpark has always been known as a hitter’s park, and it isn’t because of the altitude. As you can see, Turner Field dwarfs the one in Cincinnati, especially to right field. The advantage to left is somewhat negated by a higher fence, but long fly balls are more likely to get out there as well.
Reds ISO Maps
The Reds have a good offense, but it is one that can be pitched to. Shin-Soo Choo is an OBP machine, but he has a hole on the inner half of the plate. Zack Cozart isn’t very good at hitting but hits there because middle infielder/bat control, and he likes the ball up-and-in. Joey Votto, Brandon Phillips, Jay Bruce, and Todd Frazier are where things get dicey, but they each have their holes – not much, away, in (esp. against LHP), and away, respectively. Devin Mesoraco has a top prospect pedigree and is still trying to become an everyday player, but he isn’t there yet. And Derrick Robinson probably shouldn’t be on a major-league roster.
Do you have that one pitcher you always happen to see when go to a game played by a certain team? Well, Bronson Arroyo is my guy, and when considering which game of the series I would go to, I purposefully avoided this one. It’s not that Arroyo is a terrible pitcher. He, in fact, a pretty average pitcher who limits his walks (6%) while not striking out many (15%) and giving up his share of home runs. He doesn’t throw hard, so mixing his pitches and locations is key, but as you can see, he leaves the ball up in the zone, which causes him to be homer-happy at times. Looking at his pitch selection, he throws the kitchen sink at both types of hitters, but the real difference is when he gets ahead and goes heavy on the curve to RHH and goes with more sinkers to LHH. Again, expect a lot of junk pitches as only about 41% of his pitches are fastballs.
Brian McCann returns tonight to the aid of Paul Maholm both at the plate and behind it. His return shifts Even Gattis to LF and Justin Upton to RF until Jason Heyward gets back. Arroyo seems a poor match-up for Gattis, but there’s little doubt Gattis could put one out of GABP. Maholm will look to counter a tough lineup, but the best hitters on the team are left-handed, though Phillips and Frazier are nothing to sneeze at. The key, as always, is to limit the baserunners that get on before the middle of the order, but it’s especially necessary in a park like this. The Braves will need to patient on offense and wait back on Arroyo, and if Arroyo is just slightly off, this game could end up in a laser show.