2016 xK% Upsiders vs 2016 Steamer Projections
So just as I had hoped when I published my 2017 starting pitcher strikeout rate upsiders using my xK% equation, followed by my recap of my 2016 upsiders, it generated some excellent discussion. I made mention in my recap article, that was also requested in the comments, that I should also compare how xK% did versus the 2016 Steamer projections. It’s not enough to share a list of 12 guys with upside, show that as a group they did improve, with 66% doing so, and call xK% successful. If every forecasting system projected improvement as well without my xK% equation, then perhaps the equation isn’t providing any incremental value. So let’s bring my recap table back, but this time with the 2016 Steamer projections included.
| Player | 2015 K% | 2015 xK% | 2016 K% | 2016 Steamer K% | 2016 K% – 2015 K% |
|---|---|---|---|---|---|
| Tim Lincecum | 18.0% | 21.6% | 16.0% | 16.8% | -2.0% |
| Danny Duffy | 17.4% | 20.2% | 25.7% | 21.3% | 8.3% |
| Steven Wright | 16.8% | 19.5% | 19.4% | 15.3% | 2.6% |
| Chi Chi Gonzalez | 10.7% | 13.1% | 11.3% | 13.7% | 0.6% |
| R.A. Dickey | 14.3% | 16.7% | 17.3% | 15.9% | 3.0% |
| Chris Young | 16.6% | 19.0% | 23.2% | 15.8% | 6.6% |
| Edinson Volquez | 18.2% | 20.5% | 16.3% | 17.3% | -1.9% |
| Matt Moore | 16.6% | 18.8% | 21.2% | 21.1% | 4.6% |
| Jered Weaver | 13.5% | 15.7% | 13.4% | 13.9% | -0.1% |
| Bud Norris | 18.8% | 21.0% | 20.6% | 20.9% | 1.8% |
| Marco Estrada | 18.1% | 20.2% | 22.8% | 17.4% | 4.7% |
| Nick Martinez | 13.8% | 15.8% | 8.9% | 14.4% | -4.9% |
| Averages | 16.1% | 18.5% | 18.0% | 17.0% | 1.9% |
As you should know by now, I’ve never been afraid to review my work, even if I was wrong. The results here are a mixed bag for the xK% equation. Of course, let’s be reminded that this is just 12 pitchers, which is a tiny, tiny sample size. Perhaps I’ll take on a bigger project, use my new, yet-to-be-published xK% equation (don’t get too excited, it’s just updated coefficients!), and do some backward looking xK% vs Steamer projection analyses.
Before checking each individual pitcher to determine whether 2015 xK% or the Steamer projection was closer, I got excited because for the group, xK% performed better. However, Steamer was actually closer on seven of the 12 names, which is strange!
One more caveat — xK% is not, and never was intended to be, a projection. It’s backwards looking. So Steamer has a built in advantage because it’s actually a 2016 projection, whereas xK% is really just telling us what the pitcher’s strikeout rate “should have been” in 2015. It doesn’t account for age, league/team changes, late seasons pitch mix or velocity changes, etc.
When I do my Pod Projections, I project the components of my xK% equation, which then spits out a strikeout rate projection. More than half the time I’d say I keep what the formula calculates, otherwise will manually adjust it slightly. So it could still technically be used to put together a forecast.
I think the Matt Moore projection was possibly the most impressive and surprising. He was returning from TJ surgery in 2015 and posted a weak strikeout rate, so it looked rather optimistic for Steamer to be forecasting his strikeout rate to jump back above 20%. Either the projection engine was brilliant or it just lucked out and was blissfully ignorant of Moore’s injury.
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.
Nice work. Looking forward to more articles (from you and other authors) that use the Steamer projections as a baseline when evaluating a new indicator.