Archive for Pitchf/x

At the beginning of last season, I contextualized the swinging strike rate (SwStr%) (and refreshed those numbers after the season concluded). I had seen other analysts call certain pitches “above-average,” “below-average,” “elite,” etc. using the league-average whiff rate as a baseline. This is neither a criticism nor a judgment, as I absolutely did this before I had my statistically-driven epiphany. But understanding the average four-seamer’s or slider’s or cutter’s whiff rate lends additional context to any assertion one might make about the “elite-ness” of a pitch.

More recently, I wanted to convert discrete outcomes by pitch type into fielding independent pitching (FIP) statistics — namely, FIP and xFIP (expected FIP, which substitutes a pitcher’s rate of home runs per fly ball for the league-average rate). Let me warn you now: the results are very imperfect. It took some brute force on my part to get there, but I got there. I would wager that the the extreme (lowest and highest) values are probably a bit exaggerated. Regardless, it’s an interesting table to ingest:

Read the rest of this entry »

It seems as though each year, fantasy baseball analysts, “professional” and amateur alike, hone in on a new — or, if not new, then relatively untouched — metric or data set for their endlessly eager consumption. In 2015, FanGraphs introduced batted ball data to its leaderboards. In 2016, Statcast data was unveiled, although it arguably didn’t become popular until 2017, and before the 2017 season FanGraphs changed the game with its splits leaderboard. Baseball Prospectus has introduced myriad new metrics, too — DRA in 2015, DRC+ last year, etc. — and we began to lean into pitch-specific performance analysis last year. (The latter-most topic is relevant to what follows here.)

I recently joined Christopher Welsh and Scott Bogman of In This League on their podcast. I thought one of the evening’s questions was particularly topical and prescient (and I paraphrase): What will 2019’s it metric be? The question was asked with pitch values, something I’ve seen garner increasing attention on Twitter, in mind.

You can acquaint yourself with pitch values directly from the man who created them:

Read the rest of this entry »

Rob Silver, the 2016 National Fantasy Baseball Championship (NFBC) Main Event winner and high-stakes fantasy baseball extraordinaire, messaged me on Twitter a few days ago to ask a question: Which source of pitching statistics are most accurate? I’m paraphrasing. Also, I could paraphrase the question any number of ways: Which source should we be using? Which most reliably correlates with pitcher performance?

It was a question for which I had no answer. Admittedly, I use a variety of sources, none of which align with one another — something I have noticed before but about which I can do nothing but shrug and accept it as a quirk of being a sabermetrician who bears the struggle of dealing with publicly available data.

The sources cryptically mentioned above include the following:

Read the rest of this entry »

Some good pitchers, despite being good pitchers, throw bad pitches. And there are bad pitchers, too, who throw good pitches. Both are true, and one could make an argument a Venn Diagram of the two groups may overlap significantly, and that overlapping area is the group of pitchers toeing the line between breaking out and being unusable for fantasy purposes.

It stands to reason, then, that good and bad pitchers could benefit from easing off or completely abandoning their bad pitches. It’s one thing to evaluate a pitch based on its underlying metrics — its swinging strike rate (SwStr%), its ground ball rate (GB%), its velocity, and so on. It’s another thing to evaluate the pitch objectively by looking at its weighted on-base average (wOBA) allowed, which, I hope, in an adequately large sample, can indicate a pitch’s quality regardless of its peripherals. In theory, the larger the sample size, the greater the probability a pitch’s outcomes will converge with its inputs, such that the caveat “regardless of its peripherals” doesn’t actually mean anything. Given enough pitches thrown, the aforementioned underlying metrics will adequately inform the wOBA allowed.

Using PITCHf/x data from the last two years, I looked for (1) good pitchers who throws pitches that allow (2a) extremely bad wOBAs with (2b) unusually low BABIPs. Incurring high wOBAs on low BABIPs is less than ideal; if BABIP is subject to high variance and generally converges on the league average, then a bad pitch being “lucky” by BABIP suggests things will only get worse.

This post was going to be about several pitchers, each with their own problematic pitches, but I became too passionate about this single case. This is about Dylan Bundy, his abhorrently bad four-seamer, his fantastic slider, and how much his pitch selection is suffocating his potential. Ultimately, it’s about adding by subtracting.

Read the rest of this entry »

If you’re a nostalgic fantasy baseballer, you’ll remember that Patrick Corbin generated 3.5 wins above replacement (WAR) in his first full season of baseball before suffering the dreaded curse of Tommy John. (If you’re even more nostalgic, or more likely an Angels fan, you’ll remember he was traded alongside Tyler Skaggs for Dan Haren.) Corbin returned to baseball in 2015, and he shoved, seemingly indicating he suffered no ill effects of his surgery.

Yet 2016 was an unmitigated disaster, culminating in a midseason move to the bullpen and a full-season 5.15 ERA. A low strand rate (LOB%) is the blame — virtually no one suffers a 64.8% strand rate for a full season without some bad luck — but poor control and a home run problem complicated things. It appears to me Corbin ran afoul in two distinct ways in 2016.

It also appears to me he may have recalibrated himself. In his last three starts, he has struck out 23 and walked only four across 19.1 innings, good for a 1.86 ERA / 2.53 xFIP / 2.59 FIP.

Read the rest of this entry »

Yesterday, I borrowed PITCHf/x data from Baseball Savant to investigate how changes in batted ball velocity affected batting average on balls in play (BABIP) to a hitter’s pull side. If you’re too lazy to click, the short of it is: more velocity coincides with a better batting average. However! Lefties consistently fare worse than righties on ground balls to the pull side at all batted ball velocities.

This phenomenon can perhaps be attributed to the defensive shift. Or to the ease with which second and first basemen can convert singular outs at first base compared to their shortstop and third base counterparts due to the distance (and, thus, difficulty) of the throw. Or, most likely, to both.

But that’s not why I’m here. I’m not in the business to speculate — not today, at least. I’m just here to provide the facts in the form of some numbers I crunched in Microsoft Excel that, if you read yesterday’s post, you will probably find interesting. It has a nifty graph, if words aren’t your thing.

Read the rest of this entry »

Baseball Savant, a website maintained by Daren Willman, is a thing of beauty. Aside from some great leaderboards and applications, Willman hosts a database of PITCHf/x data. Using it without a game plan is like entering the Amazon without a machete — it can be unwieldy and overwhelming. Navigating just right bears ample fruit, however. I would like to share some of my fruit with you.

Because in 2015, PITCHf/x data began including batted ball velocity for most balls in (and out!) of play. Batting average on balls in play (BABIP) is a critical component to player success, and while there has been plenty of focus on it in the last decade — more so than, say, pitch framing, which is a popular but still-raw area of research — the baseball community would still benefit from a better understanding of BABIP, especially in light of more frequent employment of defensive shifts.

Intuition tells us that a harder-hit ball in play will have a greater probability of resulting in a hit. (Indeed, my expected BABIP equation from last year that helps corroborate such a claim.) Specifically, in regard to ground balls and defensive shifts, a hard-hit grounder will have a much greater chance of clearing a crowded first-base line than would a softly hit grounder.

Enter Baseball Savant and its very granular PITCHf/x data. Read the rest of this entry »

There are a sundry of middle/set-up relievers that can succeed in the closer role if given the opportunity. A few obvious: Wade Davis if something happens to Greg Holland… Wil Myers for James Shields and who?!; Ken Giles if the Phillies can somehow find a trade partner for the grundle-grabber; and Brad Boxberger if not Jake McGee. I assume David Robertson signs elsewhere and Dellin Betances steps in.

Lets’ look at potential closers using reliever outcomes. Here are the average contact and balls in play-related outcomes for all relievers that qualified and specifically relievers with 10+ saves:

Our Filters: 

• Contact-related outcomes: For K-BB%, Ct% and SwStr%, I filtered simply by the general relief pitcher averages. Everyone below average in these 3 categories was filtered out.
• Command-related outcomes: For zone% and first-pitch-strike%, I used 1 standard deviation below average and 1.25 SD’s below average in BABIP as filters. 1.25 SD allowed me to omit only the relievers that had career BABIP’s higher than we would like to see for a closer or in general. I didn’t want to screen out Jenrry Mejia (1.24 SD below the mean) or Tim Stauffer (1.19 SD below), because there’s a possibility for BABIP regression.
• Balls In Play-related outcomes: I was lax on the balls in play outcomes. I went with a 40% Grounder rate and 45% Flyball rate as my filters versus the averages that you see above because below average fly-rates don’t mean much in places like Tampa (where Boxberger is elite but below average in fly-rates); and above average grounder-rates don’t mean as much with atrocious defense behind you (hence Corey Kluber’s unlucky BABIP, which should have been closer to .299 per end-of-season xBABIP/Inside Edge data), but I digress.

Using these filters, we’re left with a robust list of above-average relievers beyond just closers (scroll down for the noted filters):

What happens if we use the command (K-BB%)/contact (Ct% and SwStr%) related averages for relievers with more than 10 saves this year? 

…We’re left with some elite closers and then a few interesting names.

Last year, Danny Farquhar (just missed the list this year) had a top 20 swinging-strike rate – about a percent better than Fernando Rodney, but it was masked by his BABIP and left-on-base rate that killed his surface stats (4.20 ERA vs. 2.40 xFIP). This year, he actually outperformed his xFIP with a 2.66 ERA. After an early season MASH Report on Rodney’s velocity, I eyed Farquhar. At least keep him in mind next year if anything does happen to Rodney.

Josh Edgin (Mets for those of you that don’t know) has a top 65 contact-rate sandwiched between Mark Melancon and Jake McGee and even induced grounders 50% of the time. He has a pretty extensive repertoire as well. In order of usage: Fourseamer, Slider, Curve, Change, Cutter and Sinker. According to his Brooks Player Card, he has great swing and miss rates on his Curve (>56%), Cutter (50%), Slider (>42%) and Sinker (33%). Even his Change approaches 30%. This isn’t the case on his Fastball, but at 93+ MPH, it induces a decent amount of grounders (1.8 GB/FB). Keep in mind he had late-season elbow issues which effected his velocity by a MPH or so, but he could be called upon to get Mejia out of a jam. I like him better than an unhealthy Bobby Parnell and Jeurys Familia from a command perspective for another year.

Zach Duke did his best Craig Kimbrel impression prior to the R2M monster hitting him in August. Prior to 8/1, Duke had a 34.9% K-rate and 27.3 K-BB%. Kimbrel ended the year with a 38.9% K-rate and 28.3 K-BB%. I think August and September brought him back to his realistic value (~2.50 ERA, 1.15 WHIP). It will be interesting to see who closes for the Brewers if they let Francisco Rodriguez go. Both Duke and Will Smith have above average (even for RP w/ 10+ saves) swing-and-miss. Will Smith should have additional command next year, but Zack Duke induces grounders better which I like in my closers/in Milwaukee. They also have Jonathan Broxton. The hierarchy seemed to be K-rod-Broxton-Smith late last season. If that’s the case, they should use Duke more (former starter) and in higher leverage situations. He was equally solid against both lefties (.258 wOBA) and righties (.262 wOBA).

Oliver Perez everybody! I thought I could filter him out by his splits being a lefty, but like 2012, he was more effective vs. righties (and faced 44 more of them). The D-backs have Addison Reed, up-and-comer Evan Marshall as well as Daniel Hudson caught touching 97 MPH so if not by outcomes or splits, we can filter Perez out by opportunity.

Darren O’Day and Andrew Miller is part of one dominating Baltimore bullpen – one that gets referenced by anyone who thinks the Orioles can beat the Tigers in the ALDS. Notice that Zach Britton didn’t make either of the above lists! He was filtered out by his below average K-BB% (13.70%). It’s his 75+% grounder rate (hence the 81+% left-on-base rate and .215 BABIP) that keeps him elite in Baltimore. The only concern you can have with O’Day is a fastball velocity almost 2 standard deviations below the mean for relievers, but his arm angle combined with that slider still induces a 30+% whiff-rate on both pitches. Miller’s slider though is a world apart from O’Day’s: only Pedro Strop, Will Smith, Jake Diekman, Greg Holland and Oliver Perez induces more whiffs than Miller’s 55% according to Baseball Prospectus’ Pitchf/x Leaderboards.  I doubt we’ll see a closer-transition next year in Baltimore unless Britton’s GB/FB ratio takes a drastic dive because his HR/FB ratio, which approached 18%, could be an issue.

An xBABIP review

On the last day of the season, @jeffwzimmerman provided me with Pitch xBABIP based on inside edge data. Let’s look at some of the bigger xBABIP differentials to keep in mind:

The last column depicts the z-score for BABIP differential. Francisco Rodriguez was expected to have a BABIP about 120 points above his actual BABIP. I highlighted (red/bad; green/good) the xBABIP z-scores as well so that you know whether or not to actually be concerned meaning sure Aaron Sanchez has the 5th biggest BABIP differential (over 2 standard deviations from the mean), but a .239 xBABIP is still utterly elite (3.34 SD’s from the mean). On the other side of the equation, it’s nice to see Evan Marshall, Carlos Martinez and Adam Ottavino (albeit in Colorado) with large BABIP differentials. Marshall and Martinez even have xBABIP’s over .5SD from the mean.

The last bit of fun

It was a very fun year to be doing bullpen reports for RotoGraphs. Aroldis Chapman broke the single-season strikeout rate of 2012 Craig Kimbrel (50.2%). He struck out 52.5% of the hitters he faced. Andrew Miller (42.6%) and Brad Boxberger (42.1%) also made the top 10 seasons ever. Dellin Betances (39.6%), Wade Davis (39.1%) and Craig Kimbrel (38.9%) made the top 20. Chapman’s swinging-strike% of 20% beat ’12 Kimbrel by .8%, but he couldn’t pass ’04 Lidge, ’03 Gagne, ’04 Gagne, ’02 Gagne or ’05 Lidge. Chapman, Miller, Doolittle, Boxberger, Betances (Wade Davis and Kenley Jansen close behind) all had historical, top 20 K-BB rates. Relievers dominate this list: only ’99 Pedro Martinez (#12), ’00 Pedro Martinez (#21), ’01 Randy Johnson (#27) and ’01 Pedro Martinez (#28) make it into the top 30, but it’s clear that we have a growing list of elite relievers.

From a fantasy perspective, thanks to 45+ saves totals out of Holland and Kimbrel, we had two relievers ranked in the top 20 pitchers. If Chapman didn’t miss time and Betances and Davis consumed the closer role, we would have had 3 others. Last year, Craig Kimbrel and his 4 wins, 50 saves, 98 SO’s, 1.21 ERA and .88 WHIP campaign made him the 3rd most valuable pitcher. This year with Kershaw, Cueto, Felix and Kluber, it would have taken even more.

If we combined the 3 more dominating performances exclusive of saves this year: Aroldis Chapman’s K-rate (52.5%) and saves total (36), Dellin Betances IP (90) – who was dominating in his own right, Wade Davis’ ERA (1.00) and Wins total (9) and Sean Doolittle’s WHIP (.73) – let’s call this guy Aroldellin Dooldavis, we would wind up with a 10.95 z-sum…just above Corey Kluber (10.72), but under Clayton Kershaw (13.74), Johnny Cueto (12.95) and Felix Hernandez (12.50). Even 50 saves wouldn’t have done the trick (12.43 z-sum):

Last second I changed the title from “Having a Presence” to “Commanding Attention”. If I were to talk about the team as a whole, I would have had to go with “Having a Presence,” but I’m focusing on their rotation, and well…they command our attention…

It’s rough to be a Mets fan living in Minnesota. It’s rough being a Mets fan anywhere, but watching the Twins and their approach to 100 losses doesn’t help. Despite being only 6.5 games back from the 2nd Wild Card to date, I don’t think the Mets will have enough offense to sustain next year, but that doesn’t mean they won’t be fun to watch. There will be quite the trio that commands attention – or even a quartet with the assumption that Noah Syndergaard makes an impact. And with a quartet…who knows?!

Read the rest of this entry »