top of page
Russell

Beating the Projections With Playing Time Adjustments

Updated: Oct 27, 2021

As I discuss in this note on projections, the people who make the more complex algorithmic projections are far smarter than I am. Projection systems like Steamer and THE BAT are widely used and trusted because they do an excellent job projecting player performance. They take far more data into account than the average fantasy player could possibly incorporate into a projection of their own (I make my own projections and I admit this inferiority). But there is one area in which a human touch can beat the projections: Playing time.


Every November, when the popular algorithmic projections go live, players download them, plug them into their own spreadsheets, and start ranking without giving any thought to playing time. That gives players willing to put in the time to make those adjustments a considerable advantage. Consider the following 2021 preseason hitter projections from Steamer:

  • Injury History Example - Mike Trout

2021 Steamer Projection: 150 Games | 671 Plate Appearances


Nobody could have seen Mike Trout's 2021 calf injury coming (Only 36 games played), but no human would have projected 150 games. Mike Trout has not played in 150 games since 2016. His games played in full-length seasons since that year are 114 (2017), 140 (2018), and 134 (2019). Even his 53 of 60 games played in the short 2020 season prorate out to 142. I would look at that history and project 140 games at the high end. That may seem like a small thing, but here is what it does to Trout's Steamer projection:


Steamer: 150 Games, 671 PA, 529 AB, 116 R, 41 HR, 106 RBI, 10 SB, .279 BA.

Adjusted: 140 Games, 621 PA, 490 AB, 107 R, 38 HR, 98 RBI, 9 SB, .279 BA.


That's roughly 8% shaved off of Trout's projection with basic common sense, and maybe that drops Trout out of the top 5 picks for you, and you would have avoided rostering him in 2021 as a result. These things matter.

  • Role Example - Jared Walsh

2021 Steamer Projection: 94 Games | 391 Plate Appearances


Walsh is the opposite of Mike Trout from a computer's standpoint. Heading into 2021, Jared Walsh had only appeared in two MLB seasons, and never in more than 34 games. But lost on the algorithmic projections is that Walsh finished 2020 with one of the hottest months in baseball. His 9 home runs and .337 batting average that month all but assured him a regular role entering 2021. And if you believed he could hold that role, then Steamer's projection of 94 games was wildly inaccurate, which turned out to be the case. Walsh played in 144 games, missing a small chunk of time with an intercostal strain in late July. Had you adjusted his projection to that of a full time player, you'd have beat the field.

  • Opportunity Example: Shane McClanahan

Steamer Projection: 71 Innings


After pitching 4.1 MLB innings in the 2020 playoffs, Shane McClanahan was on fantasy radars heading into 2021, but Steamer had him pegged as a reliever and projected him for 71 innings. The Rays developed McClanahan as a starter. Since his professional debut in Rookie ball in 2018, he had only appeared in games in a relief role on two occasions. Of course, the Rays do interesting things with pitchers, but an algorithm projecting innings does not know how to account for the nuances of Tampa pitcher usage. As a human, you can say to yourself: "Well, they've always used McClanahan as a starter, that's what they want him to be, and even if he's an opener or a bulk reliever, they want him to carry a load, so surely he can pitch more than the traditional 60-70 inning RP." And if you went through that process and adjusted McClanahan's innings projection for 90-115, you would have come much closer to his actual innings total (123.1) than the algorithm. This is a very common scenario every season. See Luis Garcia (Hou, SP) for another example.

  • Adjustments Based on News

If it is not clear by now, the best way to utilize projections is to copy them into your own spreadsheet where you can adjust and save as necessary. But it isn't all about adjusting what you see as inaccurate playing time projections. You need to stay on top of the news. I use three sources for this: NBC Sports Edge, Rotowire Player News, and Jeff Zimmerman's Mining the News articles. Take, for example, Zimmerman's most recent Mining the News from 10.13.21, which is filled with information that should prompt you to adjust playing time projections. It notes that the Angels plan to have a 6-man rotation in 2020, and Jose Suarez and Patrick Sandoval will be in it. It quotes Seattle brass saying that uber-prospect Julio Rodriguez will be on the MLB roster next season, possibly even from the jump. It says that the Rangers plan to have Dane Dunning throw more innings in 2022. That is only a small sample of the valuable information regularly found in these articles, all of which should lead you to tweak and adjust the playing time in your projections.


How To Adjust Playing Time


A tool for adjusting projections based on plate appearances or innings pitched is simple enough to make and invaluable to any fantasy player that uses projections for ranking players, making tiers, or any other kind of draft preparation. I use it almost every day. Linked here is a publicly facing version of what I've integrated into a number of my spreadsheet tools. You may use that tool, or download an excel version here.


--


Armchair Roto is an independent site. If you enjoyed this article, please consider using the links below to share it on social media. Thanks for reading!


291 views0 comments

Recent Posts

See All

Opmerkingen


bottom of page