Fantasy Freestyle: Using Statcast Analytics for Fantasy Decisions

As a fantasy owner, I like to look at what the progressive front offices inside the game are doing in their player evaluation. What are the brightest minds in baseball doing to pick players? I often find concepts and ideas that I carry over to my player evaluation in my leagues to help me construct my fantasy rosters.

Here are some Statcast-derived metrics I've observed in recent years that forward thinking front offices are using that I think are useful for fantasy owners.

Part one: Contact quality: exit velocity plus angle

Contact quality logged by Statcast is one of my favorite peripheral statistics to use. What type of contact is a hitter making? What type of contact is a pitcher giving up? If a hitter has a low BABIP, is he making lousy contact that is easy for fielders to convert into outs or is he running into some bad luck on good contact hit right at fielders? If a pitcher is underperforming some of his defense independent run estimators, is he getting knocked around with well struck contact or is he being snake bit by factors outside of his control? Contact quality can tell us a lot about this.

Contact quality is specifically referring to a combination of exit velocity and batted-ball angles. These angles include vertical launch angles and horizontal locations. The angle of the batted ball is important to include; exit velocity alone only tells part of the story, and can often be diminished with poor launch angles. The sample sizes aren't huge yet, as Statcast has only been public for two years, but we have a good idea of what batted ball velocity and angle combos are the best.

Launch angles between 10 and 25 degrees are line drives. Launch angles help eliminate some of the subjectivity in classifying line drives, which have been somewhat of an issue for batted ball classifications in the past. Generally, anything with an exit velocity above 70 mph between these angles will go for a hit at a high rate, with higher exit speeds producing a higher extra base hit rate. The trajectory of a line drive is well constructed to knife through a defense and find grass to land in.

Below is a table of batting average and isolated power by exit velocity range for batted balls with launch angles between 10-25 degrees (line drives), as tracked by Statcast.

There is a steady increase in isolated power as line-drive exit velocity climbs, with power exploding after 100 mph. The 70-80 mph line drive range mostly contains flares that fall in front of outfielders for singles. Batting average drops to around .550 between 85-99 mph, as these line drives are more likely to be hit at a depth where outfielders are playing. Once line drives cross 100 mph in exit velocity, they are more likely to be hit over outfielders’ heads, and batting average climbs back up to over .750.

Groundballs have launch angles under 10 degrees, and the higher the ground-ball launch angle, the more a grounder carries its velocity through the infield. Not all groundballs are created equal, and not all groundballs are good for the pitcher. A groundball can have a widely different expected outcome based on the contact quality. For example, groundballs with launch angles between 0 and 5 degrees have resulted in a batting average of about .440 league wide. The outcomes tell us that these types of ground balls are poor quality for pitchers and good quality for hitters. On the opposite end, the lower the launch angle on the ground ball, the less the ball carries its exit speed through the infield. The downward impact into the ground in front of home plate slows the ball's velocity down and makes it easy for fielders to convert into outs. Batted balls with launch angles between -20 and -25 degrees have resulted in a batting average of about .100 league wide. The difference between a high-launch-angle groundball and low-launch-angle groundball is night and day.

Below is a table of batting average by ground ball launch angle. Batting average progressively dips the lower the launch angle of the groundball.

Late, nasty, downward movement can drop the pitch below the barrel of the bat and cause these low launch angles. Late sink can make the pitch deceptive, which makes it difficult for the batter to hit the middle of the ball. Zach Britton's ridiculous sinker is a great example of that. Jeurys Familia’s sinker is another good example.

Here’s an example of a low angle ground ball generated by Zach Britton. It has a launch angle of -27 degrees with an exit velocity of 85 mph.

Looking at launch angles alone on grounders can show a lot, because even high exit velocities that leave the bat at low angles can appear as weak contact to the eye because of how the downward impact into the ground slows the speed of the ball down as it makes its way through the infield. Groundballs with an exit velocity of 100 mph have resulted in a batting average of about .450 on above average (greater than -10 degree) ground-ball launch angles. The same 100 mph exit speed on below average (less than -10 degree) ground-ball launch angles has dropped batting average down to about .210. It’s the same exit velocity off the bat, but the vertical angle the ground ball leaves the bat at makes a large difference. It’s a good example for why exit velocity alone is only part of the story.

Launch angles above 25 degrees are flyballs, with popups above 50 degrees. Flyballs with exit velocities under 95 mph rarely go for home runs, with a home run rate less than 1 percent. The horizontal angle is especially important for flyballs because the outfield wall is deeper in CF than it is down the lines. A flyball hit to CF needs more velocity than a fly ball hit to LF or RF to clear the fence for a home run.

Below is a table of fly ball home run rate by exit velocity range, with fly balls defined as batted balls with launch angles above 25 degrees.

The table shows why the horizontal location is important for estimating production on fly balls. An expected outcome formula using only vertical launch angle and exit velocity isn’t ideal for flyballs because the center field home run rate is lower than left field and right field.

The higher launch angle on a flyball, the more difficult it is to generate a high exit velocity. Only 4 percent of flyballs with a launch angle above 40 degrees had an exit velocity over 100 mph, with it dropping to 1 percent above 50 degrees. This is compared to 13 percent of flyballs below 40 degrees that had an exit velocity over 100 mph.

Pitchers who skillfully generate medium depth flyballs that are easy for outfielders to catch–think vintage Tyler Clippard–will often outperform some of their defense independent ERA estimators. Clippard's change up can get a hitter lunging forward at the pitch, wrecking his base and timing, which cripples the power of the flyball.

Here’s Clippard’s change up generating a harmless flyball with an exit velocity of 88 mph and a launch angle of 33 degrees.

Where can fantasy owners find contact quality statistics for help in their leagues? The main Statcast database is located at MLB.com's Baseball Savant, and all public Statcast data originates from MLB.com. Andrew Perpetua has also done a lot of great work on Statcast. You can find his work at xStats.org. Andrew takes every batted ball velocity and angle combo available of every hitter's batted ball and looks at the prior outcomes of similar batted balls to create an expected value. Some of Andrew's contact quality stats include xOBA, xBABIP, xBA, and value hit percentage. Expected stat formulas like these are a useful tool for fantasy owners as a quick and easy way of checking hitter contact quality.

I also like looking at the percentage of batted balls or percentage of at bats that fit a certain hard contact range. For example, Kyle Hendricks limited batted balls with an exit velocity over 100 mph and a launch angle above 0 degrees to 12 percent of his batted balls, which is well above the major league average of 17.3 percent. He ranked fourth-best among starting pitchers in preventing hard contact by those standards.

While contact quality for hitters very important, it is not everything. Batters need to be able to use the whole field to avoid defenders overloading in areas that most of their batted balls go. Some hitters are also better at finding vacated areas and hitting the ball there, like up the middle. Because of this, estimated outcome formulas using only contact quality do not tell the entire story of a hitter's batted balls.

Part two: Spin rate changes possibly raising red flags

As someone who has watched a lot of Matt Harvey over the years, including first hand at spring training games early that March, it was clear something was wrong with him last April. His fastball and slider velocities were down from early in spring training. His breaking pitches had flattened out, and the overall precision was not there. His control was still pretty good; his walk rate was around 6 percent, which was down from his 2013-15 rate, but still above average. It was his command, the precision on his pitch locations, that really dropped off. Hitters were hitting tons of line drives on flat pitches located in bad spots. He just didn't look right.

I noticed the spin rate on his fastball was down from 2015. At first I thought this was because his velocity was down, because fastball spin strongly correlates with velocity. But the spin on his fastball had dropped even in the same velocity range. On pitches between 95 and 96 mph, Harvey's average spin rate was 2,317 RPM in 2015. In 2016, it fell to 2,235 RPM. On July 4, Harvey's last start of the season, his average fastball spin rate bottomed out at 2,168 RPM in that velocity range.

There isn't a lot of public information on this, but there is some indication that changes in spin rate can be an indicator of unhealthy fatigue or worse, an injury. Driveline Baseball, one of the leading innovators of pitching research, wrote this recently on their website:

"There is a growing belief that drastic changes in spin rate (usually sharp decreases) can be a sign of an injury or pending injury.

Each pitcher is going to have a small range around their average spin rate. Sharp changes from that normal range can be seen as an early sign of unhealthy fatigue. So teams can monitor their pitchers spin rate to see if there are any sustained drops of spin rate from their average. This would also work hand in hand with monitoring a pitcher’s velocity."

It turns out Harvey was pitching through Thoracic Outlet Syndrome, which wasn't diagnosed until July. His fingers were numb, which made it difficult to grip and control the ball. The nerves in his shoulder were bothering him, too. It explains a lot; having trouble feeling the baseball sounds like a difficult obstacle to overcome for a pitcher. Harvey said these symptoms became bothersome towards the end of spring training, which is around the time his struggles on the mound began. They became more pronounced in his final start of the year.

Of course, we didn't need a spin rate drop to tell us that something was wrong with Harvey–the command loss, the velocity drop and the flat pitches did that. But the drop in spin rate may be another tool to give us more evidence that something might be wrong.

At least one major league front office specifically shut down a pitcher last year after observing a drop in spin rate. The San Diego Padres saw that Ryan Buchter's average fastball spin rate had fallen to 2,109 RPM during an August 19 appearance, down from an average of 2,294 RPM in 2016 prior to that appearance. The Padres felt he was fatigued and wanted to be proactive in attempting to prevent an injury. Buchter's average fastball spin rate bounced back to 2,285 RPM after returning from the DL.

Another famous example of spin rate and injury is Wade Davis. Davis' average fastball spin rate fell from 2,390 RPM from April-June to 2,249 RPM in July. Davis was placed on the DL shortly after with a flexor strain in his right arm.

I am currently less confident about using spin rate changes as something that is definitively meaningful for fantasy owners, but there's some evidence that it's worth paying attention to. More research will be needed to make conclusions.

Statcast is not perfect yet

Statcast is mostly wonderful, but it isn't perfect yet. The system has a difficult time consistently recording velocities and angles of certain types of batted balls, mostly on the low angle end and high angle end. About 9% of batted balls were missing an exact exit velocity and launch angle in 2016.

Another issue with Statcast was brought to light here at Baseball Prospectus. Jonathan Judge found here that the Trackman systems installed in major league ballparks can influence exit velocity readings from ballpark to ballpark, ranging from +1.2 mph on the high end to -0.9 mph on the low end. This is why I try to avoid using average exit velocity whenever I can.

Despite its imperfection, I think there is useful information available to us as fantasy owners from Statcast, and since it's relatively new, there's a good chance the rest of your league isn't overly familiar with it yet.