Pitches 76-100 -- 2007 -- Transition to the bullpen

[Fatter than Joey]

For anyone interested --

 

Curiosity got the best of me last night and I spent $5, to look at what happened to the Brewers in the 76-100 pitch range in 2007, I ate some nachos and then wrote an article.

 

The article which can be found here, could be considered misoyostic, however, on the other hand, there is data in tabular form as well.

 

Just some interesting numbers I found, nothing more, nothing less.

[Brawndo the Thirst Mutilat]

Interesting stuff, but a couple points of clarification on BABIP. I've seen some widespread misunderstanding on these boards lately.

 

1) BABIP against for pitchers: A pitcher's BABIP is often rightly cited as a helpful indicator of good or bad luck distorting the view provided by the more familiar stats (ERA, WHIP, even W-L). So, for instance, we can say that Chris Capuano actually (from a true talent perspective) was pitching about as well last year as he has in the past, but he'd had some outstanding BABIP luck previously that swung over to the opposite end last year. The current thinking on pitchers' BABIP against is that it IS indicative of a genuine spread in skill, but that the spread is quite small. As a result, it takes YEARS of data for an individual pitcher's skill or lack of skill to show up. Most pitchers are up and down in this stat (like Capuano), indicating that the best approach to assessing their current skill (and projecting their future performance) is to regress BABIP all the way to league average. If you're looking at someone who has pitched 10+ MLB seasons and has posted BABIPs consistently above or below league average, you might not want to do this, because it is becoming more and more likely with each additional data point that this performance represents a genuine skill (or lack thereof) rather than luck. There aren't many pitchers that fit this description though, so DIPS (FIP, xFIP, etc.) is still an excellent quick and dirty way to get a rough assessment of a pitcher's genuine skill level for almost every season and for almost every pitcher.

 

How this affects your study: Looking at this data, I'd be inclined to argue that Brewers pitchers, as a group, were massively unlucky from pitches 76-100 (and probably somewhat lucky on pitches 1-76), relative to the rest of the league (which you've provided as an excellent baseline). Yost can't really be blamed for bad luck concentrating in one particular chunk of pitches. If anything, the BABIP performance you identify is a mitigating factor that ought to instruct us to temper any criticism we throw at Yost for the lousy OPS-against performance his starters had from pitch 76-100. To wit: "Maybe Yost left them in too long, on average, as evidenced by the OPS-against data, but then again, maybe that OPS stuff was largely the result of lousy luck, as evidenced by the fact that most of the dropoff in results (OPS-against, ERA, WHIP, whatever) can be explained by the drastic rise in BABIP-against, which we think was probably just bad luck. If the starters were still striking out hitters, walking them, and allowing HR at basically the same rate from pitch 76-100, then Yost was probably right that they still had something left in the tank, whatever the results." What I'd really like to see is whether there was a larger-than-normal FIP dropoff for pitches 76-100 - if there was, then I think we'd be justified in blaming Yost for having too slow a hook.

 

2) BABIP for Hitters: Unlike pitchers, hitters demonstrate a wide and fairly stable spread in this statistic year over year. It is much less advisable to look at an unusually high or unusually low BABIP posted by a hitter as a sign that his actual AVG was a product of good or bad luck. It's not as stable as OBP or SLG or ISO, but it's at least as good an indicator of a hitter's true talent as AVG , and probably better. So it's not quite correct to say, for instance, that we ought to expect a significant dropoff in Ryan Braun's BABIP this year because his BABIP was SO high last year. It is correct to regress his BABIP to the mean somewhat more than we would for a vet, along with all the other measures (OBP, SLG, ISO, etc.) but only because we have less than a full season of MLB data for Braun, NOT because an unusually high BABIP is unsustainable. It is incorrect to view his huge BABIP as lucky, it is more correct to say: "We cannot be confident he's actually THAT good at BABIP because we have too little data, so we're hedging our bets and expecting some dropoff, while acknowledging that it's entirely within the realm of possibility that Ryan Braun is a BABIP monster the likes of which we've never seen before." The difference between those two views is fairly subtle, but I think it's important and a reason for optimism.

 

How this affects your study: Not at all really, I've just been looking for an opportunity to bring it up.

 

To the other unrepentant stat-geeks that populate these boards: if you feel I've mischaracterized anything about BABIP here, please do feel free to chime in, as I recognize I'm calling into question something that is at this point largely regarded as received wisdom on these boards (i.e. that hitters with high BABIP were lucky in the same way and to the same extent that pitchers with high BABIP against were unlucky).

[TerraceReserved]

What I'd really like to see is whether there was a larger-than-normal FIP dropoff for pitches 76-100 - if there was, then I think we'd be justified in blaming Yost for having too slow a hook.

 

Wouldn't it be possible to discover this simply by looking at HR, K's and BB's from pitches 76-100? If those rates are worse than in pitches 1-76, that would mean a worse FIP, as FIP is just a compilation of those numbers in ERA's format. I don't have BR.com's PI, but I think it'd be interesting if that was looked into.

 

And Brawndo, excellent overview on BABIP.

[Fatter than Joey]

Brawndo Thanks for taking the time to read my crap, and put together what you did. -- The "study" was something that largely caught my eye, and I thought noteworthy.

 

BAbip --- I really didn't do a good job explaining my thought process for including those columns. I have always understood extreme high BAbips, to be a sign that a pitcher is pitching unsustainably bad. I expected the BAbip to be uniformly high, because I thought our pitchers were pitching in a compromised state. I suppose I didn't consider that someone would see 4 pitchers at ~.400 BAbip as an event of luck -- but that conclusion is as valid as any.

 

FIP -- Is really hard to breakdown for this, as we are dealing with a pitch range which could span innings, that is to say I don't know how to flesh out IP in pitch range data.

 

As per your request here are the HR BB K rates per Plate appearance for 2007

 

Cappy

1-75 .025 .079 .201

76+ .049 .089 .179

 

Bush

1-75 .030 .048 .185

76+ . 042 .075 .101

 

Suppan

1-75 .019 .069 .126

76+ .022 .087 .119

 

Vargas

1-75 .030 .086 .198

76+ .064 .100 .107

 

Hope that helps, and thanks again.

[logan82]

Although I disagree with blaming Yost for this, I don't think it is out of the realm of posibility that a tired pitcher would get hit harder and have a higher BABIP. Sustaining a high BABIP was touched on a while back when there was a discussion about Ryan Braun regressing. I believe there were a couple guys brought up that maintain a BABIP around .330-.340. Low K's and high BB's greatly influence BABIP. That's why guys like Pujols and Fielder have BABIP's around .290 even though they hit balls hard. Nothing relating to pitchers though.

[TooLiveBrew]

Low K's and high BB's greatly influence BABIP. That's why guys like Pujols and Fielder have BABIP's around .290 even though they hit balls hard.

Pujols's career BABIP is .320, though - and the funny part is that it's precisely why you describe... he's a ridiculous balance of high BB & low SO totals.

[logan82]

TooLiveBrew wrote:

Pujols's career BABIP is .320, though - and the funny part is that it's precisely why you describe... he's a ridiculous balance of high BB & low SO totals.

I see he is between .292 and .316 since 2004. I wouldn't think those are to far from league average to really cause much notice. I was just going from memory. I remembered that it was surprising, at the time, he would have such a low BABIP.

[Brawndo the Thirst Mutilat]

FTJ - thanks for the update. I'd need the league average dropoff from pitches 76-100 as a baseline to know whether or not those numbers are unusual, but it sure looks like those guys were pitching at a much lower true talent level than you'd expect when they were allowed to pitch deep in games. In other words, I'm mostly convinced that Yost left them in games too long.

 

On the other hand, this mostly quashes the widespread criticisms of bullpen overuse, since Yost didn't really have much of a choice but to burn his relievers up - 4 out of every 5 starts was being taken by someone that turned into a pumpkin at the stroke of pitch 76. I'd hesitate to forecast a repeat performance for this year, as we're no doubt talking about small samples here, but you don't need very large samples for K, BB and HR rates to become meaningful.

 

I suppose I didn't consider that someone would see 4 pitchers at ~.400 BAbip as an event of luck

 

The thing about BABIP is that (and I think this should be intuitive for most people) the outcome of an event where the batter has successfully put the ball in play is determined mostly by the batter's hitting skill, secondly by the fielding skill of the defenders, and only a distant thirdly by whatever innate BABIP suppression skill the pitcher might possess. At the point that you're looking at a BIP, the hitter has already succeeded at his basic objective. This is why BABIP rates need to be treated as meaningful skill-indicators for hitters, and random luck fluctuations for pitchers. They aren't really the products of "luck", it's just that there's WAY too much noise to pick the statistical signal out unless you've got a pitcher with years and years of data, most of which happens to point in the same direction (oddly enough, the only kind of pitcher for whom this tends to be the case is the knuckleballer, for whom BABIP rates are usually lower year in and year out).

 

So, for all practical purposes, it makes sense (statistically at least) to treat the pitcher as having nothing whatsoever to do with the outcome of a BIP. That's why I questioned whether those obscenely bad BABIP rates should be taken as evidence that there has been a dropoff in pitcher performance after pitch 75. If anything, I'd use those rates to explain the dropoff in pitcher performance we could measure any number of ways (including stats as basic as ERA) as something that is not terribly meaningful; at least not indicative of a true measurement of a decrease in pitcher skill for those situations.

 

If there was an abnormal dropoff in K, BB, and HR rates on those pitches, that's indicative of a genuine problem that last year's team had and a reason to worry a little bit about running into the same problem this year.

[Ennder]

Pitchers mostly control BABIP in two different ways. The type of results they get (GB vs FB) and the types of pitches they throw. Ground Ball pitchers tend to have a higher BABIP against and fastballs tend to also have a higher BABIP against.

 

Keep in mind as well that the NL in general had an .817 OPS against from pitches 76-100, by far the highest of any pitch range.

[Fatter than Joey]

FTJ - thanks for the update.

 

My pleasure, I'm just tickled someone is taking me up on this discussion.

 

On the other hand, this mostly quashes the widespread criticisms of bullpen overuse,

 

Yeah -- I will not pretend I am objective about Yost, but I guess I want to see if I can better understand what I am hating him for. http://forum.brewerfan.net/images/smilies/smile.gif

 

Yost didn't really have much of a choice but to burn his relievers up

 

Right, but Yost did have a choice to hook his starters sooner. I think there are cases when a manager needs an out, and he has 2 options -- starter or BP, if the starter fails, then he has to go to the BP anyway to get that out and the manager would have been better served to just put in the reliever initially.

 

I'd hesitate to forecast a repeat performance for this year

 

The thing I have a hard time wrapping my head around is how 4 pitchers could have such remarkably bad and similar numbers, in the same season, when none of them had ever a history of doing that in the past. I can think of 2 things that could consistently affect 4 pitchers.

 

1. The rigid nature of Yost's BP transition strategy

 

2. Defense -- If the defense improves, perhaps pitch 76 will happen later in the game, and Yost won't have to squeeze his starters so hard.

[logan82]

Fatter than Joey wrote:

1. The rigid nature of Yost's BP transition strategy

 

2. Defense -- If the defense improves, perhaps pitch 76 will happen later in the game, and Yost won't have to squeeze his starters so hard.

If it was strategy, why wouldn't we have seen the same things in previous years with Capuano and Bush. It comes down to either pitching your starter to long or going to Aquino/Spurling/ect. Sort of like deciding wether you want to be kicked in the balls or in the head.

[TerraceReserved]

I found some good information in The Book with pitchers HR, BB, and K rates as games progress. It uses times through the order, and although pitch counts would be better, these numbers will probably work fine as a baseline for pitcher performance.

Time HR% BB% K%
1 .028 .088 .168
2 .032 .082 .158
3 .033 .080 .140

For every time through the order there are about 30-35 pitches thrown, so roughly pitches 1-76 are the first 2 times through the order, and 76-100 is the third time through.

For our 4 pitchers, here are their average rates for 1-76, and below them 76-100:
HR% BB% K%
.026 .071 .178
.044 .088 .127

Compare that to the league average list above now altered, with the first two times through the order averaged, and the 3rd time through below it:
HR% BB% K%
.030 .085 .163
.033 .080 .140

Wow. What we're seeing here is that our 4 pitchers gave up a whole lot more HR than league average, walked more when the league average actually goes down, and strikeout significantly less batters than the league through pitches 76-100. What to attribute that to, I'm still not sure, but those are pretty significant numbers. My geuss (and hope) is that this was all just bad luck, but 4 pitchers on the same team, all going through this very noticeable decline late in the game? I really don't know....
.

[lcbj68c]

It comes down to either pitching your starter to long or going to Aquino/Spurling/ect. Sort of like deciding wether you want to be kicked in the balls or in the head.

 

I agree. I have to believe the strategy of bullpen management will change a little bit based upon the names that you have out there in the bullpen rather than a handful of AAAA or castoffs (aquino, spurling,etc). I believe there was a ton of "damned if you do, damned if you don't" scenarios with the bullpen management last year. This year, they seemingly have a crop of guys they could turn to in order to get that 1 batter out and not have to worry about the next pitching move. Like last year when it was suggested to bring in Cordero in the 7th and 8th with seemingly no quality option after him.

[Ennder]

Capuano and Suppan have a career long history of pitching poorly from pitches 76+. I'd expect then to keep having problems.

 

Bush hasn't had problems except in 2007 and Vargas is pretty normal in that range.

[Fatter than Joey]

Capuano and Suppan have a career long history of pitching poorly from pitches 76+.

 

No they don't -- at least to the degree that they did in 2007.

[Ennder]

Capuano and Suppan have a career long history of pitching poorly from pitches 76+.

 

No they don't -- at least to the degree that they did in 2007.

Well they certainly do have a history of pitching poorly from 76+, but no not as bad as last season. I still would expect them to continue to pitch poorly in that range even if it isn't as bad as they did last year.

[Fatter than Joey]

Fair enough