This is my third year of doing fantasy baseball projections for Draft Buddy. They are available on the website, and included in the Draft Buddy Excel download. Each season is getting a little less time consuming.
The roots of my process are the same outlined in Mike Podhorzer’s Projecting X 2.0 and Tanner Bell’s accompanying Excel spreadsheet. Instead of manually projecting a player’s raw stat, I’m projecting the underlying skill metric that affects that statistic.
Take for example, home runs. I don’t just forecast 35 HR for Mike Trout. Instead, I forecast the individual components that influence home runs – a player’s strikeout and fly ball percentages, and home run per fly ball rate.
If you want to know more, read Mike’s book! I highly recommend it.
[Editor’s Note: Draft Buddy receives a commission on purchases of Projecting X 2.0 at the above link through their affiliate program. Although we aren’t blowing smoke here, Chris bought the book years ago.]
My Secret Sauce
Using Mike’s process and Tanner’s spreadsheet as the base, I worked towards a more programmatic approach to determining each player’s underlying component. I’m referencing the past three seasons across not only the MLB level but also AAA and AA levels. To help alleviate small sample sizes due to injury, COVID-shortened season, etc., I fill in the missing playing time with MLB averages over the last three years for players of the same age.
Here is a basic example: Let’s say the 29 year old hitter I am forecasting only had 100 plate appearances last season. In those 100 PA he posted a K% of 26.0% and a BB% of 7.4%. I consider 400 PA the threshold, so I need to “make up” the other 300 PA with MLB average for 28 year olds (since this was last year when the hitter was 28).
Over the past three seasons, 28 year old MLB players have posted a 23.9% K% and a 9.0% BB%. Now, for this hitter I use a weighted average of (100/400*26.0%) + (300/400*23.9%) to find a K% of 24.4%. Repeat the same process for BB%, which in this player’s case works out to 8.6%.
These are the numbers I use for this player for last season (2021). I repeat the process for 2020 and 2019. Then another weighted formula finds the metric I use to calculate the player’s raw stats based on playing time I manually input.
Forecasting the playing time for anyone in the upcoming season is the bread and butter of any projection system. No one can predict injuries, so you have to go with what you know at the time. I prefer to reference other projections such as ATC and FanGraphs Depth Chart projections to get an initial idea of a player’s projected playing time.
Then, following free agent signings and spring training, additional playing time adjustments are made and pushed through the projections system for updates. Thankfully the lockout resolved and a 162 game regular season is ahead of us. I was pessimistic after my first set of 2022 projections, published prior to each side seemingly digging in their heels.