Time for another update on my non-HDIB? comings and goings, of which there are few, but for good reasons — lots of moving and shaking in my life right now, which has slowed down the usual sabermetric drivel down to a slow drip.
As of late, I have started a small series over at Beyond the Box Score on the best and worst pitches of 2013. You can find my thoughts on four seam fastballs and sliders over at BtB. I used a very sophisticated method as well as an old Cray supercomputer to come up with my algorithm.
OK, fine. I used z-scores based on a handful of PITCHf/x derived values and went from there. On a Mac. Not as lustrous as working on a Cray, but there you go. Once things have calmed down to a dull roar, I hope to pick up the series where I left off by looking at best/worst change ups, cutters, curves, and maybe even split finger fastballs.
Dull roars are a nice segue into my latest article over at Camden Depot — this one is on concussions. Get it? Dull roars, headaches…yeah. Gauche attempts at humour aside, I discuss what a concussion is, how it is treated, and also dig into the little bit of data we have for players who suffered a concussion in 2013 and see how they did pre/post concussion, performance-wise. I hope to grab more data — Jon over at Camden Depot was kind enough to grab a good chunk of the 2013 and 2012 data for me already — and see how things parse out across several years of data. Fun? YOU BET. Analyzing the data, not concussions.
Now, different words, these being of the audio kind.
If you go to Tunes of i, you will find The Shift podcast, courtesy of Beyond the Box Score, with myself included. As we are all busy folks over at BtB, timing and scheduling matters are still being worked out, but we are well on track to be doing a ‘cast weekly, with us hoping for twice weekly sessions. While we are shooting to have guests on periodically, it will primarily be Bryan Grosnick, Andrew Ball, and myself bringing you sabermetric goodness over the airwaves, via a series of tubes. It will mostly be the other two guys, with me being awkward and rambling here and there, so listen to them, and feel free to laugh at my awkwardness. It’s a lot like this:
The fun doesn’t end there! Not only have I been doing the above, but I am also
painfully gainfully employed for the first time in nearly a year. While I wish I could say that it was for a baseball team or baseball related organization, I sadly cannot; hopefully, that day will come, but for now, consulting work in the bioinformatics field will have to help pay the bills for now.
That being said, I am proud to say that I will be working for a baseball organization in the form of a behind the scenes as a contributor -slash- intern for Baseball Prospectus, doing more stats and technology driven endeavours for those powers that be. I am very stoked to be helping out such a great organization and some great folks.
And if that wasn’t enough on my plate to keep me from posting as frequently as I once did at HDIB?, I will also be heading down to Orlando for this year’s Winter Meetings, doing what, I don’t know. Hobnobbing of some sort. If you’re down there, come say hello! I’ll be the pudgy white guy in a shirt and tie — you can’t miss me!
As always, thanks for reading and for your patronage.
With the retirement of New York Yankees closer Mariano Rivera, some things have or are becoming less and less a part of the game, the number 42 notwithstanding. In particular, Rivera was a rare breed of reliever that is slowly going the way of the buffalo, a remnant of how games were closed out before the 2000’s — the multi-inning closer.
Much like the reliever who relied upon the split finger fastball, which was en vogue for most of the 1970’s and 1980’s, only to be practiced by a scant few in the current day game, the closer who comes in to get the 4+ out save has apparently all but vanished. With the current landscape of the bullpen filled with one out specialists and closers who can only come in to a game in a clean inning, but only if you ask nicely, the reliever who can be relied upon to get more than one inning’s worth of outs stands out amongst the ultra-specialization of the 21st century bullpen.
Or is it? Is this Last of the Mohicans perception of the multi-inning save guy just a misguided narrative, or is there some merit to the notion that closers like Rivera or his counterpart in Boston — Koji Uehara — for example, are few and far between?
Let’s take a look at the last 25 years of saves, which is a reasonable swath of data to look at and figure out if the 4+ out save is truly dying a slow death. 25 years also covers the evolution of the bullpen and the role of a closer within the 9th inning, and the development of the setup man and the stat that accompanies said 8th inning guy, the hold.
So with that, let’s look at some data, all courtesy of Baseball-Reference.com and their invaluable Play Index tool. Using said tool, we will look at all saves of 1.1 innings or more between 1988 and 2013.
This first chart is simply a count of all of the multi-inning saves (let’s shorten it to MIS moving forward), broken down by year. The various coloured tiles denote a pitcher — as we can see, there were quite a few pitchers who notched a MIS back in the day, with quite a few of them notching multiple MISs in a season. The number at the end of each row is the number of MISs for a season; we can already see a drastic change in the role of a MIS, with 2005 being a particular watershed year for the ‘death’ of the MIS.
Rewinding a bit, let’s talk about the kings of the MIS:
Of the relievers with 10 or more MISs between our years of interest, we find one — ONE — who is still currently active moving into the 2013 offseason: Philadelphia Phillies closer Jonathan Papelbon. Adjusting the criteria to include pitchers who have made an appearance in the last three years and we grab one more pitcher, the aforementioned Rivera. All in all, this list is dominated by some of the bigger closer names of the 1980’s and 1990’s. Right now, there is a strong chance that the dying breed of the MIS closer is more truth than happenstance.
Just to compare/contrast some of the players who dominate the first five seasons of our 25 year span to those of the last five years, I have included the next two pie charts for comparison:
With this, we see a trend — the leaders of the given five year eras vary greatly in how many saves they have garnered to put them atop the MIS leaders list — no one in the last five years has notched double digit MISs, while in the early era, ten saves wouldn’t even be a drop in the bucket.
Just eyeballing simple counting stats isn’t enough, we need to look at more data to really see if things have truly changed. Let’s do that now and take a look at the average MIS across the years, with regards to the number of innings pitched and the average walk and strikeout rates per nine innings (BB/9 and K/9) notched:
While the walk (in orange) and inning (in red) rates are a little tough to see given how close the data run together, we do see a couple of trends — strikeout rates are up and average innings per MIS are slightly down; while there is a bit of ocean wave-like variability, it looks as though walks haven’t really changed much over the 25 years of interest.
So far, we have one solid and one potential change in the MIS over the last 25 years — more strikeouts per outing as of late, with the more recent MIS outings being shorter in duration as compared to those of yesteryear. Cool? I guess.
Let’s grab a couple more stats that can help evaluate the quality of a reliever’s outing — average leverage index (aLI), run average by 24 based out situations (RE24), and win probability added (WPA). I will leave it to the reader to peruse this reference to get a better idea of the finer grain details of each of these stats. In a broad sense, looking at these stats, we can get a feel of how valuable and crucial these MISs were to the success of the team.
Again, we see some interesting trends, with an additionally interesting drop in aLI from 2008 to the past season; WPA doesn’t appear to have much change across the quarter century, with RE24 also showing a little drop off in the last couple of years, but making a return to pre-2010 values.
Doing a Pearson’s correlation on all of these stats of interest across year, we get the following results:
Created with the HTML Table Generator
We find four of the six stats have a statistically significant correlation with year (P-value less than .05); in particular, a significant negative correlation between IP and RE24 and year and a significant positive correlation between K/9 and aLI and year is found. One caveat — RE24 is an additive stat, so the fact it trends significantly with innings pitched isn’t a huge deal. However, the fact that we find small (all Pearson’s R’s are very small, below .30) but significant trends in innings, strikeouts, and aLI do portend to the MIS evolving over the last 25 years.
In spite of a number of stats showing us the MIS of years past are not the same as the few that we do see in our current day game, we haven’t seen the death of this quirky save just yet. In fact, an encouraging spike in MISs in 2013 — a jump to 44 after only 17 being notched in 2012 — shows us that with some help from the likes of Uehara as well as Los Angeles Angels of Anaheim’s Ernesto Frieri and veteran Carlos Marmol doing their bit to keep the MIS alive, perhaps we shouldn’t be so quick to pen the eulogy of the multi-inning save.
A few days back, I had a look at the statistic ‘pace’ and briefly discussed how pace affected the Washington Nationals pitching staff. Overall, we found that the Nats staff is fairly quick compared to the rest of MLB, with relievers being more apt to dawdle. We also discussed very briefly what, outside of a pitcher’s internal clock, could factor in to the pace stat — things like home run rate, holding runners on, and what have you.
Let’s return to this topic of pace and more specifically to the topic and its effects on the Nats’ big three starters — Gio Gonzalez, Stephen Strasburg, and Jordan Zimmermann. We have already seen some of their cursory pace data, but let’s now look at these three with a different lens; let’s now look at the role and effect of the catcher on pace as well as some of their other pitching stats.
First, some brief materials and methods discussion. Again using FanGraphs data, I grabbed game log data for all three for 2013, including pace, plate discipline data (think contact and swing rates), as well as some other standard data (things like pitch counts, home runs, walks, FIP, and xFIP stats). From there, I matched the games to each of the big three catchers for the Nats last season — Wilson Ramos, Kurt Suzuki, and Jhonatan Solano, with the idea of doing some statistical voodoo and breaking down pitcher stats (pace, xFIP, et cetera) by catcher to see if there were any significant differences in how quickly or productively each of the Big Three pitched, depending on who was behind the dish.
One more caveat to our methods here; for each pairing, I used the catcher who started as the catcher for each pairing. While there are a few games where the starter was relieved mid-game and a pitcher’s pace could possibly be affected by this change, I made the leap that the potential for this is negligible. Also, for the most part, catcher swapping was done later in the game, by the time our Big Three were pulled from the game; therefore, pitcher data should be consistent across catcher. While I could do my due diligence and break this all down by inning, the amount of manual manipulation of the data to do that is too much of a time suck, so here we are.
Good with that? A reasonable leap of faith taken? Moving along…
So. Our catchers. We have three, and here are their vital stats for this data set:
Ramos: 44 games caught
Solano: 4 games caught
Suzuki: 46 games caught
One quirk here as well — Solano only caught Strasburg, while Suzuki and Ramos both caught all of the Big Three. More on this later.
Now, some data, in the form of pretty charts!
First pace and pitches per inning for the Big Three; for the moment, I am filtering out Solano data.
Not too much here; overall, pace is pace for our Big Three, regardless of the catcher, with Suzuki getting pitchers to work a hair faster and hair more economically. I will save you the statistical gymnastics and tables, but there were no statistically significant differences in any of the pitching stats I grabbed across catcher.
The stats I looked at were:
xFIP – FIP
So with that aside out of the way, let’s keep looking data; here we look at the difference between xFIP and FIP (xFIP minus FIP) between catchers as well as strike rate:
With the difference between xFIP and FIP, it was my thought it could possibly portend to some sort of measure for catcher-pitcher dynamic, given that in general, both should trend together tightly for a pitcher, since the only huge difference between the two is how they handle home runs in their calculation. With that in mind, more positive numbers are desired, since it means that the pitcher-catcher combo did better than expected, at least with FIP as our yardstick. Does it really mean anything? Probably not; however, it is interesting to see that all three starters did worse than expected, on average, with Suzuki catching and that Strasburg’s Ramos outings were a full run worse. Again, there might be something here, but I am doubtful. Moving on to strike rate, we again see no real grand deviations with differences in catcher considered. Not shocking.
Now, let’s take a look at the Solano data; let’s also remember that this is based on FOUR GAMES, so we really can’t say much about it, but we can at least admire the differences seen with Solano’s signal calling:
Kind of quirky. With Solano catching, Strasburg improves across the board with the pitching stats of interest; again, it’s only four games, and there are a ton of variables and effects that we are not taking into account with the data as presented, but it is an interesting trend we see. Two things that are beaten into a pitcher’s head — work fast and throw strikes — that are purported to be the secret to success are seen with Solano, for Strasburg, in spades. Is it an effect of Solano and possibly some intangible rapport the pair has? Maybe. Could it be simple coincidence? Yep, could be that too. But it will be an interesting trend to keep an eye on in 2014.
Could all of this be luck?
It could be that too; these BABIP rates by catcher are interesting just as an aside, not so much as a predictor of success (BABIP stinks for that), but how different the rates are across catcher and it speaking to there possibly being an effect of pitch selection amongst each of the pairings. Another study for another day, I suppose.
While there are many aspects of the fulfilling the duties of a catcher that haven’t been considered here, by the looks of it, the Nats are fortunate to have a pair of catchers in place for 2014 — Ramos and Solano — who won’t be poisonous to the overall productivity of their Big Three starters and might even be positive influences on their staff.
One of the many tangibly intangible things that a pitcher can do to endear himself to umpires, his defense, and their adoring TV viewer public is to work fast. Pitchers also benefit themselves by working fast, as one of the basic tenets of successful pitching is to interrupt a hitters timing — by controlling the pace of an at bat, the greater chance a hitters innate cadence and timing for each pitch can be oh so slightly interrupted.
Even with some of these tiny advantages to working fast, some pitchers just…
One of the many side stories of this year’s World Series that exemplifies this was Clay Buchholz‘ pace to pitch. Glacier-like was the overarching sentiment as to how quickly a Buchholz outing went; given the age-old complaint that any Boston Red Sox game takes longer than most, this wasn’t a new revelation.
So was it all in our heads? And how do Washington Nationals pitchers do in this arena of pitching?
FanGraphs can help us out with this — with the pace statistic.
Pace tells us, on average, how many seconds a pitcher take in between pitches of an at bat. For 2013, the average time between pitches was 22.9 seconds, when you include all players who made an appearance as a pitcher (*waves at Skip Schumaker and John McDonald*).
Again looking at all pitchers, starter or reliever, but using an innings pitched cutoff of 10 innings and we find that the fastest pitcher was New York Yankee Vidal Nuno, at a 17.2 second pace; on the other side of the coin, Tampa Bay Rays reliever Joel Peralta was the slowest, at 31.9 seconds.
However, this doesn’t tell the whole story; while time between pitches is obviously a hindrance to a quick game, so are the number of pitches thrown. More pitches thrown, the more time in between pitches, the longer the inning is — simple math. So with that in mind, let’s pull out our abacuses, TI-82 calculators, and perhaps some scratch paper, and figure out who are the yin and yang of pitching pace.
So, we have pace, innings pitched, and total pitches, courtesy of FanGraphs. From there, we simply need to calculate pitches per inning, then multiply by pace, then divide by 60 to get a rough estimate of how long an inning is for a given pitcher, in minutes.
Doing all of this voodoo brings us to this:
Cool? Kinda, I guess; to circle back to the Buchholz reference and close the book on that, he comes in at a 24.2 second pace, 15.04 pitches per inning, which gives him an average inning pitched in 6.09 minutes. This ranks him 54th among 146 pitchers with at least 100 innings pitched and fourth out of the five Red Sox starters with at least 100 IP. Yes, Clay is a quick worker, compared to his carmine brethren.
Enough of that, let’s talk WASHINGTON NATIONALS BASEBALL. Remember that? Barely, I know. Sarcasm aside, let’s have a look at how the Nats staff pans out with this whole pace thing. A priori, we are led to believe that Jordan Zimmermann is a very fast worker and Stephen Strasburg kind of plods along.
Any truth to this?
Well, yes and no; ZNN appears to be a quick worker, for sure — not too far behind the pace setter (pun intended) Nuno at 4.79 minutes per inning, but Strasburg isn’t too far behind him. So we have a wee bit of a misconception of Strasburg’s pace, even when pitch count is factored into a pitcher’s overall pace. We also find Nathan Karns conspicuously absent from our list; for whatever reason, Karns and two other players’ paces were not listed by FanGraphs.
Yunesky Maya? Yeah. Moving on…
The bottom part of the list — let’s lovingly dub them DAWDLERS — are primarily relievers. While we know that Tyler Clippard enjoys long walks around the mound, blowing into his pitching hand, licking his hand, and doing a number of sundry things on his pre-pitch checklist before dealing, we also see another quirk about the pace stat and an underlying component of it, courtesy of hitters, with the dawdlers — home runs.
Yes, the more home runs you give up, the more time you have to wait for the hitter to get around the bases before you can throw your next pitch. While I will save the rigorous statistical analyses for the effect of home run rate on pace for another day — and as we can see in the above table, it isn’t something that appears to be highly correlated — it is an unfortunate aspect of pitcher pace.
Keeping with eyeballing trends, we also see with the Nats that relievers are a tick slower than starters, in general. While homers do play a role, the fact that relievers are in-game during typically higher leverage situations — a fancy way to say there are men on base with the game on the line — the need to change your timing to the plate and even throw over to first base on occasion to keep a runner close takes precedence over keeping your pace number low. Strategery at its finest.
So again doing some quick eyeballing, the Nats average a 6.03 minute inning; countering
Cubans outliers and using the median, Nats pitchers come in at about 5.7 minutes per inning. Comparing that to the MLB overall (and again using a 10 IP cutoff) and an average 6.3 minute inning or 6.2 minute median inning, and we see that Nats hurlers are a tad quicker than the average team.
Here’s how the entire MLB pans out; note that this chart does not include data from players who swapped teams during the season and are thus noted as playing for ‘—‘ by FanGraphs.
So there you have it; the Angels are DAWDLERS on average, while the Braves and Cardinals, bastions of all that is unwritten and unheralded, are the quickest. Snark aside, both of those teams more than likely have an organizational mantra that predisposes guys to work fast, and if you look at their team pitching stats, it’s definitely not hurting their stock.
Pace; it’s not just salsa.
It’s been awhile since I’ve last posted here and while I would love to tell you I have spent the time away from HDIB? analyzing the L.O.M.B.O. data and results in preparation for a manuscript that will be submitted to the International Journal of Sport Grit and Want and Desire and Other Things No One Can Truly Measure But Dammit We Try — the IJSGWDOTNOCTMBDWT for short — alas, I have not.
I have however, been busy writing about the Shutdown. Put away the pitchforks and take a look at what Stephen Strasburg and Jordan Zimmermann have done before and after Tommy John surgery and having their innings limited post operatively across a number of stats and categories:
For the Orioles fan in your life, I have lots of prose dedicated to Manny Machado and his MPFL injury and his prospects for a healthy return and what that means for the Orioles in 2014:
Looking to 2014, part of a series over at Camden Depot breaking down 2013 and 2014 by position.
I will be helping out Camden Depot with a couple more of these year in review write ups, focusing more on the O’s bullpen; you can also find more of my nerd math writings over at Beyond the Box Score, if you enjoyed the Strasburg/Zimmermann/TJ bits discussed.
Anyone know what IJSGWDOTNOCTMBDWT’s impact factor is? Might have to shop L.O.M.B.O. around…
I like to think of myself as something of a scientist. An out of work scientist, but that’s a whole ‘nuther ball of wax. As such, I look at articles and research such as this with a curious eye. For those click averse, it is an article discussing the virtues of ‘grit’ over IQ as a predictor of success in life.
It’s fascinating research that lends credence to the notion that perseverance and volition will get you further in life and will allow you to accomplish long-term goals more so than just being a Smarty McSmartypants. While the premise of the work is still something I wrestle with in terms of its application to everyday life, the research and its results did strike a chord in my baseball mind.
‘You gotta have heart’.
‘Gritty, gutty, and scrappy will conquer all’.
You’ve heard plenty of these paroxysms, so I will stop there.
But is there something to having 25 gritty, gutty sons of bitches on a baseball team that might cure all ills, first round draft picks be damned?
Inspired, I devised my own survey for baseball players — the Longitudinal Obstinance and Moxie Barometer for Organizations, or L.O.M.B.O.
With it, I hope to determine whether there is something to a gritty personality. It’s just recently devised, so I need some help acquiring data. In fact, consider this an invite to take the survey yourself — you can find it here.
Give it a go and see how gritty of a player YOU are. Here’s a quick and dirty breakdown of the scoring:
0-2: Lazy. No heart. You don’t run out ground balls and pimp home runs, while also possibly peeing in pools in the process.
3-5: Occasionally inspired, when the mood strikes. You are Adrian Beltre in a contract year.
6-8: You’re full of mettle, but you don’t bring your lunch pail to work every day. You’re not just having fun out there. You lack the will to win.
9-10: You’re a member of the Eckstein family. Holy shit, you’re covered in dirt even after showering and eat New Hampshire granite for breakfast, you’re so fucking gritty.
*EDIT* Since I am nicht so gut with Survey Monkey and it doesn’t look like they have an automated scoring system, here is the scoring rubric:
|1||Yes = 1||No = 0|
|2||Only lefties = 1||Often As I Can = 2||Stroke = 0|
|3||One = 0||Two = 1|
|4||Laser = 0||On a Hop = 1|
|5||Weight = 1||Batting Avg. = 0||Bunts = 2|
|6||Yes = 3||No = 0|
While last night’s Washington Nationals game against the St. Louis Cardinals will be remembered for the almost no-hitter for rookie starter Michael Wacha (and rightfully so), there was an interesting side story in the 7th inning, courtesy of Bryce Harper.
Jimmy, roll the tape:
(Click to unleash the gif)
Did you see it? Did you notice the last second hop/skip/jump in the box towards Wacha before he swung at a 96 MPH fastball?
Here’s a different angle of the swing, which really gets the point across:
(Click to unleash the gif)
First of all — wow. In a game determined by milliseconds, Harper takes something already tough to do and laughs in the face of it, making an adjustment as Wacha is throwing a pitch. A 96 MPH FASTBALL.
Second — why? My thought is that Harper was guessing changeup — Wacha’s biggest secondary pitch, and one that was especially good last night in his 8.2 innings of one-hit ball. His 32 changeups last night came in at a 62.5% strike rate and a 15.6% whiff rate. When you also consider that Wacha threw just three curveballs and one cutter, you get a better appreciation of Harper’s mindset in this at bat. Already down in the count 1-2 and knowing he wasn’t going to get anything breaking, he assumed he wouldn’t get something hard, so he set up in his usual fashion in the batter’s box, then scooted up to get to that assumed changeup before it darted out of the zone.
Here’s a plot of the pitches he saw in the at-bat; as you can see, Harper was one step ahead of Wacha as far as getting that changeup, eventually striking out on the pitch:
Pretty darn impressive. To not only be able to move the feet, keep your swing mechanics intact while doing so, look for and react to a changeup, but then get a high-90’s fastball, and still be able to catch up to it enough to just foul it off is all sorts of amazing and impressive. But you know what? It’s been done before.
Remember Hal Morris?
While his church league softball batting stance wasn’t as egregious as Harper’s, Morris also employed a foot shuffle before getting the bat through the zone. With a .304 batting average, relatively low 12.3% strikeout rate, and 14.6 fWAR over a 13 year career, Morris was surprisingly effective with the approach. Also to note is while Harper all but leaped towards Wacha with his swing, Morris’ was more of a shuffle up towards the plate, with a small step towards the pitcher; small difference, but an important one.
While the end result wasn’t terribly desired — Harper ended up striking out — it nonetheless was an interesting look at Harper’s approach and how he is able to not only make adjustments, but make them in real-time; a very rare