MASTER NOTES: Foolin' With A New Metric

As the old saying goes, “fool me once, shame on you, fool me twice, I might add you to my fantasy baseball pitching rotation.”

I should explain.

Within the next few days, will post an article I wrote for the “Facts & Flukes Spotlight,” a regular in-depth look at players who are showing unusual production.

I was asked to turn my spotlight onto BAL starter Chris Tillman, who is, somewhat surprisingly, off to a $17 start this season, with an ERA under three and seven wins in his first 11 starts. The question to be answered is always, “Is this guy for real?”

Part of the rationale for the Spotlight series, and a big part of the fun of it, is that we writers are encouraged to use the newer data and analysis tools on sites like FanGraphs and Brooks, which provide a lot of tools for using PITCHf/x data, and Baseball Savant, which offers a way into the Statcast data from

It sometimes takes a little while to load all the data into a spreadsheet, because not all the sites use exactly the same protocols for identifying players—first name, space, last name; last name, comma, first name, with or without a space; and so on. Even the space between the names can be problematic, because some sources use the space that you get from the spacebar on your keyboard, while others use a space that is different enough to prevent matching the names from one data source to another. There are some numbering systems as well, but nothing has been comprehensively standardized.

But once all the data are in the spreadsheet, and all the names are the same, that’s when it gets fun—slicing and dicing the data, combining metrics, breaking them apart, calculating new ones.

That’s what happened when I was looking into Chris Tillman. Using the PITCHf/x “Plate Discipline” data from FanGraphs and manipulating with some more basic pitch-count data, I found myself looking at some data columns with pitch counts of about 100 starting pitchers that included how many pitches each pitcher had thrown that were in the strike zone or out, and were swung at or not.

And that’s when I thought of Fool Rate.

The idea stemmed from the concept that a pitcher is often trying to fool the hitter or deceive the hitter or otherwise discombobulate the hitter into doing something he doesn’t want to. As far as these pitch counts were concerned, I thought, a pitcher could be said to have fooled a batter if he got the batter to swing at a pitch outside the zone, or to not swing at a pitch inside the zone. So if I added those two pitch types together and compared with the total of all pitches, I’d have the percentage of pitches on which the pitcher had fooled the batter.

Okay, it’s not like discovering the law of displacement of fluids, which is just as well because nobody wants me running named down the street shouting “Eureka.” But it still seemed like a promising idea.

After calculating each pitcher’s Fool Rate, I wanted to test the list to see if the Fool Rate was a workable, helpful metric connected in some real way to fantasy stats like ERA and strikeouts, and to important underlying stats like hard-contact rate. But that had to wait.

The first test I always run in these circumstances is an old-fashioned eyeball. I sort the list from highest score to lowest, and see if the cream has risen to the top of the statistical pail and the thin skim milk has dropped to the bottom.

The top of the Fool Rate list was 38%, which I thought was astonishing, by Noah Syndergaard of the Mets and Aaron Nola of the Phillies. The rest of the top-10 were Kyle Hendricks, Zack Greinke, Collin McHugh, Gio Gonzalez, Alex Wood, Adam Wainwright, Josh Tomlin, and … Bartolo Colon.

I thought this was a decidedly mixed bag, and frankly a little disappointing, with an aggregate 3.79 ERA and 1.20 WHIP. I should have expected a wily vet like Colon might sneak in, but McHugh and Wood just didn’t suit the Fool Rate hypothesis.

But the pitchers just outside the top-10, some within less than one full point of the top-10, included Clayton Kershaw, Cole Hamels, Tanner Roark, Jose Fernandez, Rich Hill, Corey Kluber, Jake Arrieta and Madison Bumgarner.

At the bottom of the list, 13 pitchers were under 29%, with James Shields at just 26%. Shields is really struggling this year. so this was a good sign—if not for his fantasy owners or the Padres, at least for the Fool Rate concept.

The other names on the bottom were Yordano Ventura, the aforementioned Chris Tillman, Danny Salazar, Chase Anderson, Phil Hughes, Anibal Sanchez, Jeff Locke, R.A. Dickey and Wily Peralta. Aggregate ERA of 4.70, WHIP of 1.42.

That’s more like it!

It wasn’t like the Fool Rate had aced the test. Maybe a C-plus or B-minus.

The next test was a correlation test, comparing Fool Rate to various other metrics to see how tightly connected they are. As a quick reminder, correlation values run from plus-one to minus-one. At those extremes, the correlation is perfect—any change in Fool Rate would cause a proportional change in the other metric. A correlation of zero would be evidence of a perfect non-connection, saying in effect that the two variables were moving without affecting each other at all.

I was hoping for correlations in the 0.50 range or better, and I didn’t get them. All the correlations were in the 0.30-0.38 range—there’s a connection, but not a really strong one.

Still, Fool Rate could be useful to identify possible situations where a pitcher is outperforming or underperforming his skills, which is the first step in deciding how to manage that pitcher for fantasy purposes. And down the road, I’ll try to dig into this a little more deeply, ’cause I have questions: Is Fool Rate consistent over time? Does change in Fool Rate indicate change in performance?

And I have to think more about the weaknesses in the metric. One big one that came to mind while I was preparing the Spotlight piece was that hitters often take an in-zone pitch not because they’ve been fooled but for legitimate reasons, like a take sign from the manager or waiting on a particular pitch and/or location on a 3-2 or 2-0 pitch and not getting it.

So for now, it’s not a philosopher’s stone. But it something to file away and use cautiously. The last thing we want is a Fool Rate metric that we use to fool ourselves.

Click here to subscribe

  For more information about the terms used in this article, see our Glossary Primer.