First Pitch: Jeff Locke and the Defense of FIP
I’m a big Louis CK fan. I’ve mentioned this at times on the site, or on Twitter, and especially on Facebook. Although I don’t really make my Facebook page public, so you’ll just have to trust me on that one. I think I’ve grown to like Louis CK even more in the last year because of his comments on divorce. His comments about pretty much any subject are extremely hilarious, but also extremely accurate. It was no different for divorce. So last year when I got divorced, I found his jokes to be hilarious, and totally true.
Before you “aww like a puppy died”, as Louis CK would say, know that one of his most accurate jokes is how divorce is always good news. In a “Tim Williams” quote, rather than using a Louis CK quote, I can tell you there are things I no longer have to do. For example, I never have to step foot in a Pottery Barn again. If Michael’s is having a 50 percent off sale on frames, I don’t have to walk around the store finding something interesting, or try to run out the clock looking at cats at PetSmart next door. If I see a hot girl in a bar, I can talk to her, fail to pick her up, then go home and masturbate thinking about her without any guilt. OK, that last one was Louis CK inspired.
There’s one Louis CK divorce quote that I thought about in a different way tonight. I’ll post the quote below, then try to tie it into baseball so this isn’t just an excuse to talk about Louis CK. Plus we’re like three paragraphs in, and hardly any baseball talk.
“It’s hard to start again after a marriage. It’s hard to really look at someone and go ‘Maybe something nice will happen’. I know too much about life to have any optimism. I know even if it’s nice, it’s going to lead to shit. I know if you smile at somebody and they smile back, you’ve just decided something shitty is going to happen. You might have a nice couple of dates but then she’ll stop calling you back and that’ll feel shitty. Or you’ll date for a long time and then she’ll have sex with one of your friends or you will with one of hers and that will be shitty. Or you’ll get married and it won’t work out and you’ll get divorced and split your friends and money and that’s horrible. Or you’ll meet the perfect person who you love infinitely and you even argue well and you grow together and you have children and then you get old together and then SHE’S GONNA DIE.
That’s the BEST CASE SCENARIO, is that you’re gonna lose your best friend and then just walk home from D’agostino’s with heavy bags every day and wait for your turn to be nothing also.”
That “I know too much to have any optimism” part is kind of the way I feel about advanced statistics. In the past, back before I knew about AB/HR, strand rates, or BABIP, I would look at a starter like Jeff Locke and think “that guy could be an ace”. Now I know too much. I don’t even look at the ERA. I look at the FIP numbers. And because of that, I know things aren’t going to be as good for Locke in the future.
James Santelli wrote about that subject this afternoon. I’ve written about pitchers regression using FIP a lot in the past. There was one notable article back in 2011 when the Pirates were contending, but almost every single pitcher was performing above their heads. If you look at the comments in both articles, you’ll see a lot of the same arguments.
1. Let’s just enjoy it.
I don’t want to pick on Lee Young, because I like Lee. But I did laugh when I saw the same response from Lee in two articles about the same topic separated by about two years. And you know what? Lee is right. You should totally enjoy it.
I can’t tell you how much I enjoyed watching Locke’s last start where he went seven shutout innings, giving up just three hits. Sure, it was against the Astros, but it was still nice to see. It was the same way against a better team when he threw seven shutout innings against the Cardinals. I’ve covered Jeff since he first entered the organization. I’ve always liked talking to him, and I’ve always pulled for him to succeed.
But there’s a line you have to draw when you’re covering a team. You need compartmentalization skills. Locke is a good guy who I always enjoy talking to, and who I wish to succeed. But none of that can enter into the evaluation of him on the site. And neither can anyone else who is on the site. It’s our job to point out things like “Jeff Locke has a 4.48 FIP”, and explain that he’s probably going to regress. We’re not saying “don’t enjoy this”. We’re kind of saying “enjoy it while it lasts”.
There was also a “hater” comment on Twitter, which I found funny. It was only two months ago where this site was saying Locke should be in the rotation while so many were saying he should be a reliever or in Triple-A. Now we’re saying that he’s not a top of the rotation guy, but a strong number four starter. Again, it’s just part of the job to point these things out. It’s not “hating”.
You should totally enjoy what Locke is doing. And if you hope that Locke defies the odds and becomes the next Jeremy Hellickson, then go for it. I hope that happens too. I’m not sure any Pirates fan would be any different. But that would be the exception, not the rule, and that’s not the way to lean when presenting analysis.
2. Why Can’t Locke Be Different?
You could almost make a “Five Stages of FIP Grief” article out of this. The most common thing I see is a sort of bargaining/denial approach where you hope that Locke (or anyone else) is the exception to the FIP rule. Maybe there is something magical about him that leads to a lower FIP? (Statistically, there probably isn’t.) Maybe it’s the defense, manager, or some other team factor is helping him? (Then why do other pitchers have a FIP that is much closer to their ERA?) Maybe the stats aren’t reliable? (Which is more likely: a group of stats trusted by a lot of very smart people are unreliable because they’re predicting a regression about one player, or that one player is going to regress and the stats are correct?)
As James pointed out in his article today, the chances that Locke is the exception are extremely low. Like I mentioned in #1, we all hope that he’s the exception. But the odds are heavily leaning towards the rule, which means Locke is likely headed for regression.
3. What About (Insert player who doesn’t apply to the conversation, suggesting that the stats would have been wrong about him)?
In that 2011 article, someone brought up Jose Bautista, saying that he discredits the approach of “here is what he will do going forward, based on his past stats”. I’m not sure how Bautista applies to pitching, but I kind of see the general point. That’s also a good point to make. All we can do with stats is look at them, analyze them, and use them to try and predict what will happen in the future. But players will improve. It might not always be as drastic as Bautista’s improvement.
I picked on Lee earlier in this article, so I’ll give him credit for something he brought up today. Locke’s FIP since his first three starts is around 3.45. He’s walking fewer batters, striking out more, and that’s leading to better numbers. The question is, can you take out those first three starts? Do they represent something that Locke is still capable of? Or has he moved on, and the last six starts are a reflection of the pitcher he will be going forward? If it’s the former, then his season FIP holds up. If it’s the latter, then he’s going to be a better pitcher than his FIP indicates, and the regression won’t be as steep.
The key here is that Locke controls his future. ERA and FIP both tell the same story. They focus on the past and what a player has done. You use those to predict what a player can do in the future. Everyone does this. Even if you’re campaigning against FIP, you’re doing this. You’re just doing it with ERA. FIP just happens to be a better indicator going forward.
But that’s all based on the past. If Locke starts striking out more and walking less (like he has been in the last six starts), then he’s going to see a lower FIP. That’s because he will be a different pitcher than what the overall stats say. Just like ERA, a FIP can go up and down with more data. It looks at the entire season, and the more data you get, the more things can change. So Locke’s FIP can go down, although it’s not guaranteed, and it’s not the only possibility. There’s also reason to believe that Locke could go this route, since he did have strikeouts all throughout the minors, and had much better control than what we’ve seen so far.
In another “Player who doesn’t apply to the conversation” topic, there was discussion about Tom Glavine today. In his best years (1991-2000), Glavine had a 3.13 ERA and a 3.56 FIP. So he outperformed the numbers, but not to the extreme that Locke is out-performing his numbers this year. And that’s another case where you’ve got a player who has done that for a long time. Locke has out-performed his FIP over nine starts.
The Purpose of the Stats
A few weeks ago Colin Wyers wrote a great article at Baseball Prospectus about sabermetrics, and the proper way to use stats. The article talked about how stats are used incorrectly these days, aimed at ending discussions, rather than starting them. I felt the article today about Locke brought about some great conversation. It looked at what was most likely going to happen, but at the same time there were some great discussion points. Without the article, are we going to dig into Locke’s numbers and see that he has turned things around after his first three starts? Chances are we’ll now be watching him closer to see if he continues pitching like the last six starts, while avoiding looking like the guy from the first three starts.
However, we also need to consider what we know, and just like the Louis CK quote above, we know too much. We’ve been through this before. We went in with the denial and optimism that something nice would happen. And what was the result in 2011?
Pitcher – ERA at the time / xFIP at the time / ERA the remainder of the season
Paul Maholm – 3.08/4.15/5.03
Jeff Karstens – 2.55/3.81/4.66
Charlie Morton – 3.80/3.71/3.86
Kevin Correia – 3.74/4.15/7.68
James McDonald – 4.40/4.50*/3.93
Some of those pitchers regressed too far. Morton was accurate. I noted McDonald’s xFIP because I did a separate study looking at his season minus the first four starts. He got a late start in Spring Training, and probably started the year too soon. That led to some poor outings, and he was excellent after those first four starts. But the xFIP was in the 4.00 range. So his ERA the rest of the year was in line with the xFIP when you took out those first four starts. That’s a case where FIP correctly predicted improvement on the season numbers.
We know that when it comes to FIP predicting a future regression, chances are we’re not going to see “something nice happen”. Instead we’re going to see the expected regression. In those rare cases the pitcher will avoid a regression. But that’s probably because he changes the way he pitches, leading to things like increased strikeouts and decreased walks. In turn, his FIP will also go lower, meaning you’re not going to see a pitcher finishing with the extreme separation that Locke has.
Whether you believe any of this or not doesn’t matter. It’s not like Locke will regress because we point out his FIP, or that he’ll avoid that if enough people doubt the validity of FIP. The hope is that the focus on FIP can lead to three things: preparation, discussion, and enjoyment. Prepare for what is most likely going to happen. Discuss the chances of that happening, including what Locke could do to improve his FIP. Most importantly, enjoy what is currently happening, especially since the chances are strong that it won’t last.
Links and Notes
**The 2013 Prospect Guide and the 2013 Annual are both available on the products page of the site. If you order them together, you’ll save $5.
**Check out the new episode of the Pirates Prospects Podcast: P3 Episode 4: Are the Pirates For Real? Plus a Jameson Taillon Interview.