Making Simple Edits to Projections for R, RBI, and PA

In the RBI table, we find the average leadoff hitter for NL teams scoring 699 runs drive in 0.076 RBI per plate appearance.

And using this last table we can now multiply out an actual projection. There’s no difference between the AL and the NL in number of plate appearances, so no need for separate columns here.

For this example, let’s assume a 150-game season for Heyward. The table above tells us Heyward would be expected to see about 4.632 plate appearances per game. Multiplying our earlier numbers out we get:

• 4.632 PA/G * 150 G = 695 PA
• 0.130 R/PA * 695 PA = 90 R
• 0.076 RBI/PA * 695 PA = 53 RBI

Granted, this assumes Heyward would play every inning of those 150 games. To the extent you think he’d be substituted out in the middle of some games, you’d want to adjust accordingly. And selecting the Cubs for this example highlights an interesting oddity… They frequently bat a position player 9th. So perhaps Heyward’s RBI per PA should be ratcheted up some…

For reference, the Steamer projections for Heyward’s R and RBI are 91 and 71, respectively.

The Steamer projections also have Heyward for 146 G, 659 PA, 18 HR, and 18 SB. You could also adjust those counting stats about 5.5% based on our higher PA projection (695/659 = 1.055). That would bump the HR and SB both to 19 (1.055 * 18).

The Season-Long Tables

If multiplication isn’t your thing (you’re probably at the wrong website), here’s a series of full-season equivalents. They’re set to 162 games, so you’re still looking at some math to bump a player down to 155 games, or however many you think they’ll play.

For example, if we thought Heyward would play in 162 games, we would estimate 97 R, 57 RBI, and 750 plate appearances.

This last table is simply the sum of the 162-game R and RBI tables above. Because R and RBI are valued nearly identically in your standard rotisserie league, this can be a quick resource to gauge the effect of a change in batting lineup.

For example, if the Cubs re-signed Dexter Fowler to bat leadoff, bumping Heyward to fifth, Heyward might actually see his value rise as his combined R and RBI production goes from 155 for the season to 165.

 What’s the Benefit Here?

Again, my goal was to create an easy-to-use resource for those stats that are largely team-dependent. Surely we could use Steamer, Zips, or the Fans projections as our R or RBI estimate. But when news comes out that a player is being lifted from 8th in the lineup to 1st, who knows how quickly those projections will update. And they’re surely some weighted average calculation including probabilities of the player hitting in a variety of spots in the lineup. If we hear news and want to make an aggressive projection that they’ll stick in that spot, we need a resource to help us do that. Hopefully this can help in those situations.

Alternate Approaches

I’m far from the first to tackle this challenge. I was able to locate several other approaches you may want to consider. One significant weakness in the method described above is that it completely neglects batter skill. All else equal, those with higher OBPs will score more runs and those with better hitting skill and power will drive in more runs. Some of the approaches below attempt to incorporate both batter skill and team dependencies, so they likely lead to more accuracy but lose an element of simplicity.

Here are those alternate approaches I identified:

• Jeff Zimmerman, Hitter Evaluation (Part 2): Runs and RBIs (2011) – Jeff outlines an approach that incorporates skill, lineup, and team factors.
• Mike Podhorzer, Projecting X 2.0 (2016) – In the sequel to his first book, Mike takes on the challenge of R and RBI with his own formulas. He takes an approach very similar to what I outline above, but he’s calculated his own R and RBI ratios that also incorporate batter skill elements neglected above. Note, I’ve worked with Mike to develop an Excel file that accompanies the book which you can read about here.
• Rudy Gamble, Lineup Position Impact on Runs & RBIs (2009) – Similar in theory to what I outline above except Gamble uses indexes to adjust for each lineup spot. If you like his approach, with a little bit of time, I think you could adjust the tables above to create updated numbers (this was written in 2009).