Projection Accuracy: Early March Pitcher Counting Stats

Now that the analysis hitter projection comparisons (part 1, 2, 3, 4, and 5) are done, it is time for the pitchers to take center stage. For the first article, I’ll measure the accuracy of counting stats from early March. Razzball had a near clean sweep as it only missed on Saves.

First, here are the projections analyzed.

• Steamer (FanGraphs)
• ZIPS
• DepthCharts (FanGraphs)
• The Bat
• Davenport
• ATC (FanGraphs)
• Pod (Mike Podhorzer)
• Masterball (Todd Zola)
• PECOTA (Baseball Prospectus)
• RotoWire
• Razzball (Steamer)
• Paywall #1
• Average of the above projections

To create a list of players to compare for accuracy, I took the TGFBI ADP (players in demand at that time) and selected all the pitchers in the top-450 drafted players (30-man roster, 15 teams in the Main Event) in ten or more leagues. Then I removed all the pitchers who never threw an inning.

To determine accuracy, I calculated the Root Mean Square Error (RMSE) for two different sets of values. RMSE is a “measure of how far from the regression line data points are” and the smaller a value the better. Additionally, I included the actual and league average rates for reference.

I ran into a bit of an issue as 16 pitchers* were missing from one or more of three different projections.

Projection: Players missing
Pods: 7
The Bat: 6
Rotowire: 5

For this reason, I created two comparisons. One with all the projection systems but fewer pitchers and one with all the pitchers but fewer projection systems.

As for predicting Saves, two projections (Zips and the Bat) just don’t project Saves so I removed them. Also, I limited the player pool down to any pitcher from the initial subset who accumulated one Save or was projected for one Save.

Finally, here are the results:

Thoughts

• The boys over at Razzball crushed the rankings by being close to the innings thrown. And it wasn’t by just projecting fewer innings thrown. Zips and PECOTA were projected for fewer innings, but their per player accuracy was not as good as Razzball. Like with hitters, nailing the initial playing time projections led to good standings for the other counting stats.

• Two of the aggrigator projections (ATC and the average) consistently placed in the top half.

• The Saves rankings are in a comletely different order than any other set of rankings. If there was any one value I’d take an average of, it would be Saves. It’s the stat where doing more leg work can pay off and fantasy managers might as well leverage the Wisdom of the Crowds for the leg work.

• In early deep drafts, I might back away from the three projections that were missing players. The projections aren’t taking into account players being drafted in the top-450, yet alone in 15-team, 50-round draft-and-holds (750 players).

So that is all I have on pitcher counting stat projection accuracty for the early offseason. Getting close to the playing time as possible was key and in the next installment, the rate stats get compared.

* Brusdar Graterol, Kyle Wright, Chad Green, Garrett Crochet, Mike Foltynewicz, Hunter Harvey, Nick Wittgren, Logan Gilbert, Bryan Garcia, Trevor May, Sean Doolittle, Stefan Crichton, Rich Hill, Gregory Soto, Emmanuel Clase, Daniel Ponce de Leon