Major League Baseball’s Statcast is one of the most important recent advances in measuring baseball player performance. The camera and radar-based systems now exist in all 30 MLB ballparks, as well as some Spring Training facilities. In essence, Statcast captures the motion of everything on the field: the players, the ball, probably even umpires and base coaches if someone wanted to know. Statcast is a kind of centralized, standardized collection of scouting data, which has previously been collected by a disaggregated network of individual people, all with diverse methods and biases. For the quantitative communities within the league and without, Statcast is perhaps the largest single new source of data since PITCHf/x was adopted in 2007. If used appropriately, the publicly available Statcast data can give fantasy owners a critical edge as well.

Prepare for your fantasy baseball draft with our Mock Draft Simulator

Caveats

For fantasy baseball purposes, we care most about the quantification of batted-ball exit velocity and launch angle, pitch velocity and location, and perhaps running speed. Since Statcast systems were installed in all MLB parks in 2015, sharp-eyed analysts have noticed that data is missing from the public records. In 2015, Tony Blengino reported on the missing data and determined that about 25% of all batted-balls had no recorded velocity, with weak grounders and popups failing to register most often. Rob Arthur reached a similar conclusion, showed some improvement from 2015 to 2016 and demonstrated a method to fill in the missing data. David Kagan has given a physicist’s perspective. Jeff Zimmerman has also calculated and compiled “corrected” exit velocity readings using an imputation technique.

Furthermore, work at Baseball Prospectus has shown systematic bias between the Statcast systems in different ballparks and described a method to correct for it. Andrew Perpetua has stepped back and noted that batted-balls with the same exit velocity and launch angle can lead to different results depending on weather. While a good amount of Statcast data is publicly available through Baseball Savant (more on this below), the inner workings are still pretty opaque. In a recent edition of the Statcast Podcast, Mike Petriello and Tom Tango indicated that 100% of batted-ball exit velocities and launch angles would be recorded in 2017. The fact remains that, except for small snippets like this, we do not know what improvements are being made over time to the measurement systems, calibration procedures, and post-processing techniques. We may expect the data to improve over time, but without checking, it will be hard to know when equilibrium is reached. As such, any models built upon the available Statcast data should be considered experimental, treated with appropriate skepticism, and updated frequently.

Resources

Baseball Savant is the portal for public Statcast data. Developed and maintained by Darren Willman, the site has an excellent leaderboard, including the new “Barrels” metric which represents a player’s ability to hit the best kinds of batted balls: those with exit speed and launch angle that have historically produced a batting average over .500 and a slugging percentage over 1.500. Players can be examined individually, and the underlying data can be queried and downloaded using the Statcast Search function.

Originally rolled out with a series of posts on RotoGraphs and summarized recently, Andrew Perpetua maintains xstats.org, which incorporates many of the corrections discussed above, as well as incorporating horizontal spray angle from Gameday fielding locations, and classifies every batted-ball against the historical results of similar batted-balls. Using a player’s actual strikeout and walk rates, he builds a complete “expected” batting line, that is, the line a player earned from his granular batted-ball profile rather than the actual outcomes. Conceptually, this is a big evolutionary step forward from simply regressing a player’s home run per flyball rate (HR/FB%) or batting average on balls in play (BABIP) toward his career average or the league average. We can now tell much more confidently whether a player “deserved” a high BABIP or HR/FB, or was simply lucky. Ryan Brock showed that, at least for the halves of 2016, xStats used predictively could beat Steamer for young players with limited MLB track record.

I personally have developed a more simplistic expected isolated slugging metric (xISO), using data available from Baseball Savant and FanGraphs leaderboards. I maintained a daily updated spreadsheet during the 2016 season, which calculates xISO using two different models, both of which have been updated this pre-season. I do not pre-process the Savant data to correct for ballpark, temperature, or missing batted-balls, but I use only the more reliable 2016 data to determine model coefficients. The latest xISO model achieves very good correlation with actual ISO for qualified 2016 hitters, and earlier form of the model helped me correctly predict second-half power declines from four out of four hitters. I only got one of four power increase predictions correct, but injuries played a role in three of the cases.

A Matter of Trust

How might we define a player that Statcast “trusts”? First of all, let’s stick with ISO as a single metric. While not usually a fantasy category, ISO captures the ability of a player to hit for power, which is almost always a good thing for HR, RBI, and R. Unlike batting average, almost all the factors that lead to high ISO, like hitting the ball hard and at the right angles, are captured by Statcast, which is why we can learn so much about ISO from the data. Next, for each player, we can compare the following data points: 2016 ISO, 2017 Zeile Consensus Projected ISO, and 2016 xISO. For xISO I will show both mine and one calculated from Perpetua’s xstats.org. A “trust” designation will be given to a player whose 2016 xISO closely matched his actual 2016 ISO, but whose projected 2017 ISO is significantly lower. By this method, I’ve identified ten players that may own more power than the projection systems are willing to forecast:

 
Daniel Murphy (2B/1B, WSH)
Murphy’s power outburst in the second half and playoffs of the 2015 season is well documented, as is the swing change from whence it allegedly came. Murphy carried his new approach forward into a huge 2016 campaign. With such a long major league track record, projection systems have been understandably slow to react. Statcast-based metrics, however, suggest that Murphy earned his 2016 power production. He already owns the second highest consensus AVG projection and is a good bet to beat his power projections, even entering his age 32 season.

Brandon Moss (1B/OF, KC)
At 33, Brandon Moss is unlikely to provide much profit for fantasy baseball owners. He will never hit for a high batting average, but the return of his power in 2016 seems trustworthy. In AL-only leagues and deep daily move mixed leagues, Moss could be a sneaky bench option to deploy against right-handed pitchers.

Trevor Story (SS – COL)
After showing reasonable power in the minor leagues, Trevor Story debuted in Colorado last year and started destroying baseballs at a historic rate. While a .296 ISO might not be realistic to expect again, a mark above .250 seems safe, unless Coors field is relocated closer to sea level.

Evan Longoria (3B – TB)
He may be getting less patient as he ages, but 2016 saw Evan Longoria hit a career high 46.8% fly balls. A healthy 94.6 mph exit velocity on line drives and fly balls earned him xISO estimates in line with his .248 actual ISO, his highest since 2011. If he can maintain his 2016 batted-ball mix and avoid excessive strikeouts, Longoria has a good chance to beat his power projections.

Jose Altuve (2B – HOU)
Altuve is probably the third best player in fantasy baseball. He might even end up being the best. Much of his value comes from batting average and stolen bases, but Altuve has taken large steps forward in power over the last few seasons, from an .080 ISO in 2013 to a sizzling .194 mark last year. The projection systems conservatively expect some regression in 2017, but Statcast indicators are more optimistic.

Jonathan Lucroy (C/1B – TEX)
More strikeouts, more flyballs, more power. Lucroy also hit near .300 and chipped in five steals in 2016. His .208 ISO was deserved, and fantasy owners should be confident that Lucroy’s lukewarm 2015 is a distant memory.

Matt Carpenter (1B/2B/3B, STL)
Carpenter is an established hitter who has earned his power output these last two years. As with many players on the wrong side of 30, his major question mark is health.

Khris Davis (OF – OAK)
Davis has officially adopted (and adapted) the Krush moniker from his Baltimore Davis counterpart. While his average will likely hover around .250, so will his ISO, and 30+ home runs should come easily.

Jason Kipnis (2B – CLE)
A rotator cuff injury is jeopardizing the start of Kipnis’ season. If healthy, though, The stout second sacker seems to have discovered his power stroke by hitting more fly balls, and especially more pulled fly balls. This is a consistent trend among this group of hitters and hints at a deliberate approach that seems likely to be sustainable.

Nick Castellanos (3B – DET)
At number 223 in the consensus ADP – 22nd among third baseman – Nick Castellanos does not cost a lot to acquire. The owner who does, however, is buying a 25-year-old who hit .285 with a .212 ISO as a 24-year-old and probably deserved an even higher ISO. A fractured hand derailed his 2016 season, and two prior years of subpar production have saddled him with some skepticism, but Statcast data hints that Nick Castellanos could be ready to deliver solid power numbers.

• 5 Bounce-Back Infielders to Target
• 5 Bounce-Back Outfielders to Target
• 5 Bounce-Back Pitchers To Target
• 5 Negative Regression Infielders To Avoid
• 5 Negative Regression Outfielders to Avoid
• 5 Negative Regression Pitchers to Avoid

Subscribe: iTunes | Stitcher | SoundCloud | Google Play | TuneIn | RSS

Andrew Dominijanni is a correspondent at FantasyPros. For more from Andrew, check out his archive and follow him @ADominijanni.