Reviewing 2017 Pod vs Steamer Projections — Home Run Upside

Before the 2017 season, I decided to switch up my typical “Steamer and I” series posts where I discuss one player in which my Pod Projections differ from Steamer. Instead, I compared my projections in specific fantasy categories to identify upside and downside guys my forecasts hinted at versus Steamer, representing the crowd. Let’s begin our recaps of the new series with the home run upside group, which are those hitters I projected for significantly more home runs than Steamer. Note that I extrapolated the actual Steamer home run projections to match the same number of at-bats I projected, so playing time differences wasn’t a factor.

Pod HR > Steamer Extrapolated HR
Name Pod Projected HR Steamer Projected Extrapolated HR Actual HRs Winner
Ryan Schimpf 28 21 14 Steamer
Nolan Arenado 40 35 37 Steamer
Daniel Murphy 20 15 23 Pod
Khris Davis 37 32 43 Pod
Jackie Bradley Jr. 23 18 17 Steamer
Stephen Piscotty 22 17 9 Steamer
Matt Carpenter 22 18 23 Pod
Trevor Story 29 25 24 Steamer
Jake Lamb 26 22 30 Pod
Ryon Healy 22 18 25 Pod

Ughhh, I hate tying. Each of us “won” five players, but perhaps a tie is still an impressive result for the Pod Projections considering they are all done by hand, whereas Steamer is calculated by a computer through complex models and algorithms. Of course, the sample size here is tiny.

Yeah yeah, we all knew I was a Ryan Schimpf fan heading into the season, and I probably would have cleaned up with this projection had he not BABIP’d .145! I didn’t foresee a minor league demotion coming, so that killed my chances. But I was right about his power in the time we did remain with the Padres.

Ha, Steamer gets a win for Nolan Arenado, but one more homer would have given me the victory. He finally BABIP’d over .300, which I knew was coming at some point, but surprisingly his HR/FB rate slipped for a second straight season, which ultimately led to my loss here. It’s surprising given the leaguewide power spike, but everything is different at Coors, so perhaps the offense didn’t benefit as much in that park as everywhere else. Or, Arenado just had a slightly down home run year.

Daniel Murphy. Exhibit A for the value a human projector adds over a computer projector. I knew the 2016 power spike had an explanation behind it and therefore believe in its sustainability. A computer, on the other hand, was completely unaware and couldn’t possibly project anything but significant regression.

If you used Statcast’s Brls/BBE and my xHR/FB rate equation, you would have gotten the validation needed to believe in Khris Davis‘ mammoth power. Both of us figured some regression from his lofty heights of 2016, but Steamer figured far more than I did, likely without the important data I had at my fingertips.

Perhaps it was too soon to toot my own horn, because in the case of Jackie Bradley Jr., his Brls/BBE made me overly confident in his power skills heading into 2017. His HR/FB rate fell from his 2015 and 2016 levels, which cut into his homer total, but the lowest fly ball rate since his 2013 debut also contributed to the drop.

We were both way off on Stephen Piscotty’s home run total, partially because hamstring and groin injuries limited him to just 341 at-bats. I was off on him again though because he continues to hint at great power potential with his xHR/FB rate components, but fails to make good on such promise.

Despite playing through injuries and seeing his HR/FB rate decline for a second straight season, Matt Carpenter hit one more homer than I projected, which was already four more than Steamer. That was accomplished thanks to a sky high fly ball rate of nearly 51%, easily a new career high. I wouldn’t count on that again, though his offensive output has been almost identical the last three seasons, so whether he sells out for homers or balances power and batting average, it all combines for a similar package.

Trevor Story was one of the biggest stories of 2016 and given his mediocre minor league record, we wondered how much regression to expect in 2017. Complicating things was a huge Brls/BBE and xHR/FB rate that validated his power. I actually wasn’t that far off on his HR/FB rate, having forecasted an 18% mark versus a 16.2% mark that he ultimately posted. But, just 503 at-bats and an increased strikeout rate meant my projection looked more off than it should have.

Jake Lamb is another in the Murphy mold with a story behind his 2016 power spike. That made it easier for me to believe in a repeat, whereas Steamer only had his mediocre historical power numbers to lean on.

Last, but not least, is a third member of the “swing change goes well” group, Ryon Healy. I fully believed in Healy’s power, but was more concerned about his BABIP, which did decline, but not by nearly as much as I feared it could.





Mike Podhorzer is the 2015 Fantasy Sports Writers Association Baseball Writer of the Year and three-time Tout Wars champion. He is the author of the eBook Projecting X 2.0: How to Forecast Baseball Player Performance, which teaches you how to project players yourself. Follow Mike on X@MikePodhorzer and contact him via email.

4 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
motherMember since 2016
7 years ago

Pods, I love your work and am a past purchaser of Pods Projections. As you allude to in your commentary, HRs in a season are a product of two variables: HR/PA and PAs. Equalizing the two projections’ PAs is good, as I imagine it’s really hard to nail each player’s PAs (as Jumbo Schimpf and Mike Trout showed us this year, for different reasons), so would comparing your and Steamer’s projected HR/PAs to actual HR/PA provide even greater projection accuracy insight?

Keep up the good work!

motherMember since 2016
7 years ago
Reply to  Mike Podhorzer

Right. But my point was in comparison to the actuals. Since PAs are arbitrary/random/difficult to predict, wouldn’t comparing projected HR/PA to actual HR/PA be a more insightful metric to compare between projection systems than just HRs? That way you are taking the projected vs. actual PAs out of the equation.

If HR/PA seems difficult for readers to intuitively grasp, you could always display the metric as HR per 600 PA, or something like that to make it more relatable.