THE BOOK--Playing The Percentages In Baseball

You wrote:

“The historical precedence is that it would be very difficult for a pitcher to repeat his situational pitching with regards to batting average on balls in play (BABIP). Not only did Happ have a low BABIP overall (.270, which was among the league leaders), but in high-leverage situations, it was an unfathomably low .”

I think a study I did once goes along with this. It is “Do Pitchers Give Up their Expected Number of Runs Based on OPS?” I acually broke it down by OBP & SLG. But I took some other factors into account, like handedness and strikeout rates. Pitchers generally got the ERA they deserve. Here is the link

http://cyrilmorong.com/Pitrunnexp.htm

I don’t think pitchers, in the long-run, can pitch to the situation.

Fantastic stuff Tango (and David).

When talking about a disconnect between ERA and FIP or any other differences between traditional stats and saber ones with a “layperson”, I don’t like to say, “So-and-so got lucky or unlucky...” One, they will not understand what you mean. Two, there is really no clear definition of “lucky” in this regard, nor is it even accurate in some cases, so couching it that way is practically meaningless. 

If a pitcher has a .270 BABIP and his FIP is much higher than his ERA, some of us might say that said pitcher got “lucky” and that he really was not as skillful (or whatever word you want to substitute) as his ERA might suggest.  That is a ridiculous statement.  Maybe his team played spectacular defense behind him.  Maybe he allowed very weakly hit BIP.  Maybe the BIP behind him just happened to be hit right at his fielders.  To blindly say that he was “lucky” because his BABIP was lower than average and we think that most pitchers have little actual control over their BABIP, is patently absurd, especially to almost any non-sabermetrician.

The only think that should be said, and the only thing that is patently correct is that the pitcher’s ERA in the future is more likely to be closer to his FIP than to his ERA, after accounting for defense and parks and any other contextual adjustments.  You can add that specifically his BABIP is more likely in the future to be closer to league average adjusted for the defense of his team and his home park.

If you really want to talk about true talent, you can say something like, “In addition, his actual talent as a pitcher is probably less (or more) than his ERA last year (or whenever) suggests, for the following reasons.” And then you might articulate some of those reasons, such as, “Research has shown that a pitcher’s BABIP, which greatly affects ERA, can fluctuate up or down a lot over a season or even several seasons, much more so than a pitcher’s K, BB, and HR rates, such that a bad pitcher can actually have a good ERA for a whole season by virtue of having a good BABIP, and vice versa for a good pitcher.”

But to start out the conversation by telling a non-sabermetrician that so-and-so, despite appearing to have had a great season, like Happ, actually had an average season and just got lucky, or is actually not so great a pitcher and just got lucky, is not a good way to explain what you are trying to say. 

And, as I said, it really makes no sense at all.  It would be like trying to tell someone, “Oh, no, David Wright didn’t have a poor season in the HR department (he didn’t?).  He just got unlucky (no, he only hit 10 HR).”

Do you expect anyone to take that sentence seriously?

That is very similar to some of the sentences we hear from sabermetricians with regard to pitchers and pitcher seasons like Happ.  And you wonder why people like Mike Silva don’t take these Happ and ERA and FIP_ discussions seriously?

MGL: are you referring to my article, one of the readers here, or just making a general statement?

In my case, I never mentioned luck, just persistence of performance.

Good points MGL.  The biggest problem I see is that many of us are trying to write blog posts in the 500-1000 word range while also trying to add a touch of novelty to our posts and explain a new concept. It’s a careful balance between providing information and keeping the audience.  The more I ramble about research, the fewer people make it through my post and enjoy it. 

I’ll be writing for a Phillies blog when it opens its doors in a few days.  I was prepping an article about FIP, xFIP, and tERA (is bbFIP available anywhere?) yesterday.  I’m trying my best to make my stats lessons interesting.  The concept behind this particular post is to show the reader how to use those 3 stats to say Roy Halladay is better than Johan Santana.  My goal isn’t to tell the reader about the components of those stats, my goal is to get the reader to see merit in them and follow the links to primers at the end of the article.

So I used these stats which are typically used to draw conclusions on fangraphs but are nevertheless not close to the full picture.  Then I did what any blogger is supposed to do, I drew conclusions.  I wanted to say that I thought Johan’s apparent abilities in stranding baserunners and suppressing BABIP should bring expectations for the two closer together, but I couldn’t do that without first explaining BABIP, LOB%, and how research shows that these two factors are typically outside of the pitcher’s control.  I didn’t have room for that.

So I then wrote another post which I titled “The Greatest Danger”.  It basically says that while my post is not intentionally misleading, it is misleading because it only looks at the question (Doc or Johan?) from one angle.

As I grow more familiar with all the material out there, it’s becoming harder and harder for me to draw a conclusion.  Now that I’m getting my feet wet writing for a blog (and hoping to find better opportunities to blog or work in baseball in the future), this growing difficulty in drawing conclusions is worrisome.

Chris Young is a good comp. for J.A. Happ.  Young has consistently beaten his FIP and has had remarkably lower BABIPS than the norm. 

in High and Medium leverage situations his BABIP’s were in the .230’s. 

they’re both long guys with long arms that have decieving fastballs.

As usual, I rambled off topic and left out my point.  Using “luck” is a cheap shortcut to get through a post without losing the audience.  Then when someone questions luck in the comments I can say ‘blah blah research blah read this link’ and cover that base.  It’s lazy and it’s not ideal, but when there’s a space constraint and a large message, then something has to get cut.  Most people seem pretty open to accepting that luck is a factor in the year to year variation of stats, what people seem to be wary of is when it’s suggested that us stats people can distinguish luck from skill.

A team, I like to show how Happ compares against other pitchers in other situations.  Looking at his OPS against (not perfect, I know) you can see that in 2009, with no one on base he pitched like Blanton (.753).  With men on base, he pitched like Roy Halladay (.650).  With men in scoring position, he pitched better than Pedro at his peak (.479).  With a man on 3rd, he pitched like… well, there is no description for how he pitched (.415 with --3, .310 with 1-3, .143 with -23, and .221 with bases drunk).

Then, you can ask the question:  If Happ really has the repeatable ability to limit opposing hitters when there are RISP, why doesn’t he pitch like that all the time?  He’d be the best pitcher in baseball history!

That generally gets accross the idea that there is indeed a little luck involved in baseball.

Tango, I was just speaking in general terms not about your blog post on Fangraphs.  Your post was excellent.

“The concept behind this particular post is to show the reader how to use those 3 stats to say Roy Halladay is better than Johan Santana.”

Those are dangerous waters to tread for the same reasons I articulated above.  If pitcher A is 16-8 with a 4.50 ERA and pitcher B is 8-16 with a 4.00 ERA (ERAs are adjusted for context), for me to say that pitcher B was the “better pitcher” is again presumptuous and will not carry much weight to a lay person. Who is to say that the pitcher with the lower ERA as opposed to the better w/l record is the “better” pitcher?

Again, you have to be really careful about throwing about statements and conclusions using words which have no clear-cut meanings…

Situational Wins will not help you to answer the question about whether Happ’s NIBBs were “smart walks” that didn’t hurt the team, as Happ’s pitching coach claimed.  The reason is that there is more to the situation than just the baseout state.  WPA/LI is similar to RE24 in that it gives a good approximation of the number of runs that can be expected to be added from the particular baseout states instead of an average baseout state.  But who is batting next, and in the future, is also a factor in the decision to risk a walk by pitching extra carefully to a batter.  To adjust for that you need to calulate the number of runs that ACTUALLY scored either on the walk or during the rest of the same inning.  No runs scored on Happ’s 54 NIBBs, and only 17 runs scored in the rest of those innings.  Since 23.3 runs were expected to score from the baseout states before the walk, each walk cost his team nothing.  This is only an approximation of course the proper methodology would adjust the prior baseout state for the actual run potential of the batters in the lineup and also adjust for his team’s defense, but it is close enough to show that Happ’s pitching close was correct.  Happ did chose his times to walk someone wisely. But he was also very, very lucky in the outcomes, as you point out about other aspects of his pitching.  And the likelihood of his performing that well in the future is minimal.

MGL (or anyone else here),

Would you be interested in reading the article and offering a critique?  The link below is to the google.doc.  If you do choose to take a look, keep in mind that I’m writing to a very specific audience that I’m trying to introduce to DIPS (perhaps against their will).  Things get more impartial after the first paragraph/hook...sorry Mets fans.  My email address is if you have comments.  Sorry for the thread hijack.

https://docs.google.com/Doc?docid=0AbCib-AD2n2VZGRrdHc3c21fOGM3OGM5c2Zj&hl=en

Peter, since Happ stranded a ton of runners on base, then I expect him to… strand a ton of runners on base.  So, I can’t look at the actual of what happened after he issued his walks.  It’s not as if he knew he was going to strand them.

You are totally right about the on-deck hitters.

Tango - I think you have the reasoning backwards.  I would contend that he stranded more because he wisely chose WHEN to pitch more carefully and risk the chance of putting a man on via a walk.  Pitching more carefully as a conscious strategy does not always end in a walk, of course.  It often ends in a strikeout or weakly hit ball.  He went to 3-2 counts on the batter in 19 of his 54 walks, so it is quite likely that he got more than average number of strikeouts in those situation as well.  It also probably contributed to his below average BIP.
Only time will tell how much of this was luck and how much was a repeatable skill, but what actually occurred last year does not support refuting Happ’s pitching coach’s claim.

As far as the value of the on deck and following hitters go they are more likely to raise the expected runs instead of lowering it.  He walked the number 3 hitter 6 times, the number 4 hitter 6 times, and the number 5 hitter 7 times.  Plus, even if you lowered the expected runs to 0 each walk would only be worth .056 runs, way below .30.

I was joking today that it would be ‘funny’ if Happ pitched to about the same ‘counting stats’ level next year - by improving his underlying statistics. Won’t that just be a headache to explain?

I mean, there is some non-zero chance of that, given Happ’s AAA statistics.

MGL:  Excellent comment.  I’ve been guilty of using the “lucky” word, but will be more careful now.

Harvey, thanks.  When it comes to serious communication, I am a real stickler for words and their meanings.  It is especially important to keep that in mind when one person is trying to inform or educate another, or in a serious debate among equally knowledgeable persons…