FanPost

Expected Runs Versus Actual Earned Runs or Why Jon Lester Hasn't Turned The Corner.




I'm not really a math guy. Let's face it, math is hard and often not fun. Most people don't like getting too bogged down in the numbers and that is no less true in baseball, even though this is known as a stats driven sport. It's easy to stick to simple numbers. Batting average, earned run average, home runs, runs batted in, steals, saves, fielding percentage and errors are all easy to calculate and easy to explain to someone who doesn't follow the sport. They're approachable and generally cover the results of a game pretty well.

That said, most of them do a pretty poor job of predicting future success or failure. I get it. Following sports is fun because, well, it's fun. If you need to take a stats course at your local community college to keep up, it's no longer fun and then what's the point? Well, I'm here to tell you that if I can keep up, so can you. I never got past algebra 2 in high school, and never further than intro to stats in college. I only took that because my liberal arts degree in anthropology required it.

The way I started coming around on some of the more complex methods of measuring what happens in games was through fangraphs.com. They present a ton of data in a form that's easier for me to digest. I'm a visual learner and consumer. I like movies and TV shows more than books and when I'm feeling particularly cultured (or I'm trying to impress a girl) I prefer walking through an art museum over listening to an orchestra. Fangraphs is one of the most popular baseball sites out there these days, and there are others starting to put together similar repositories like brooksbaseball.net which has a phenomenal amount of data to sift through, and charts for nearly everything you can imagine.

So how does this all relate to Jon Lester and what we can expect going forward? Well, let's look at him from a few different perspectives. Let's begin with his last four starts in which he's posted a 1.24 ERA and averaged 7.25 innings per start. He looks like he's turned the corner, right? Even if we stretch back to the All Star break he has a 2.31 ERA with opponents hitting .236/.286/.360 off of him. Well, in the last four starts he has 16 strikeouts and 10 walks. That's a 1.60 k/bb ratio, which isn't very good. His k/bb since the All Star break? 3.17

Lester's season long k/bb is 2.47 which is identical to his career rate. So that 1.60 is an abnormality and one that wouldn't normally correlate with a run of good results (a low ERA). How do we express that disconnect? Stats like FIP and xFIP are a good start and are readily available at fangraphs.com. (If you can't tell, I like the folks over there.)

I'm about to give a quick explanation on how FIP and xFIP work. I'll make it as quick and painless as I can, I promise. So bear with me a moment.

FIP, or fielding independent pitching, is a stat designed to look like ERA while removing the aspects of ERA that are reliant on defense and luck. It is a measure of the results in plate appearances where the pitcher was primarily responsible for the outcome. Home runs, strikeouts, hit by pitches and walks. xFIP is expected fielding independent pitching, which is a regressed version of FIP. Put simply, they replace the actual home run rate with the league average home run rate since home run rates are very fluky. This normalizes the data a bit and correlates better with future results than either FIP or ERA.

Why am I giving a mini-lecture on FIP and xFIP? It's because it's important to understand the basic idea behind the two stats before I make my argument about Lester. In short, FIP is the results with defense and luck taken out as best we can manage. xFIP is the same thing, but with luck on home run rates also removed.

Now, let's get back to Lester. His ERA looks fantastic over his last four starts, and even his last 7. His FIPs and xFIPs for the individual games do not, however. Here's a quick graph that shows each stat for each start throughout the year.

Lestererafipandxfip_zps85c5962a_medium

You can see that most of his starts have a separation between the three stats, but there is a consistent gap in his last four, especially between ERA and xFIP. And this is a good way for these stats to be used. Some people will rely on only on ERA or only on xFIP ignoring their limitations. Comparing ERA and xFIP gives is a good idea of just how good a particular start was or where a pitcher has been on the performance spectrum and whether we can expect recent results to continue.

In this case, I would be surprised if Lester continued to keep runs off the board and go deep into games without some changes in the underlying peripheral statistics like strikes outs and walks. In other words, he's been a bit lucky and that luck can't be counted on to last.

Lester is an interesting pitcher in that his decline over the last two seasons hasn't been linked to an obvious physical decline. He's throwing a little softer than he used to, but is still one of the hardest throwing lefties in the majors. He's lost between 1 and 1.5 mph on his fastball since his peak back in 2009.

Ed00d55f89a8e8c9d5d0a60449e797f2_medium

via www.brooksbaseball.net


Digging through the data at brooksbaseball.net there is something that jumps out at me. Lester's curveball is being thrown less often and has been less effective, specifically less consistent, than it used to be.

A good way to try and measure this is to look at line drive rates on each of his pitches over his career. Line drives are almost exclusively a result of good contact. Fly ball data includes a mix of good and bad contact, as do ground ball rates. In order to see which pitches are giving up consistently good contact or are generating consistently poor contact, we need to look at a clean data set and so I'm going to use line drive rates.

549cd22d309b72a10b4a2ac8fb88f224_medium

via www.brooksbaseball.net


This is why I like charts so much. I could stare at the raw data all day and not pick up on just how inconsistent the curveball has been, but one look at this and I immediately notice that the last two seasons have seen an incredible amount of peaks and valleys with that pitch. In fact, it's the only pitch that correlates with his overall results quite well.

Lester was dominant from 2008 through 2011. His line drive rates were all fairly consistent on all of his pitches. In 2012 we see a brief spike with the cutter and in 2013 we see one with the change up. Over both years there are five spikes with the curveball.

What is driving that inconsistency? We can't be sure. It could be a mechanical issue related to a changing arm slot. Throwing overhand is not natural and wears down the shoulder over time. Some pitchers have their release point move from overhand down toward sidearm as they get older. Not a lot, mind you, but a subtle shift from 12 o'clock to 11 or even just to 11:30 could cause significant problems in retaining mechanical consistency. So maybe as he's getting older and his shoulder is slowly, naturally, wearing down, he's brought his release point out away from his head a little to compensate.

It could also just be the minor drop in velocity. If his fastball has gone from a pitch that hitters had to remain twitchy in the box to keep up with to a pitch they can flick their wrists out at and at least foul off when they get one unexpectedly, it would allow them to sit on that curveball more often.

Or, maybe he's tinkered with his mechanics a bit throughout his career, as most pitchers do, and at some point between 2011 and 2012 he lost his feel for the curve and just never got it back.

As a Sox fan, I hope it's the third possibility as that is correctable. If it's one of the first two, then Jon's days as an elite starter and the anchor of a staff are probably over. That doesn't mean he can't still be a very good starter and a boon to a rotation, but being an ace requires consistency and consistency is something Lester hasn't had in the last two seasons.