03-24-2024 MLB Preview - Flipbook - Page 12
MLB ’24
Sunday, March 24, 2024 12
SEASON PREVIEW
ORIOLES
The art of bird watching
How projection systems
explain Orioles’ 2023
breakout, and how they
can beat them again
By Matt Weyrich
T
he Orioles caught the
baseball world by surprise
in 2023, vaulting themselves out of their rebuild
in emphatic fashion to win 101 games
and clinch the American League East.
Behind breakout seasons by Gunnar
Henderson and Kyle Bradish, Baltimore exceeded even the most optimistic projection systems by a wide
margin. Those computer projection
systems are back to set the bar again in
2024, and while the two most reliable
sources might only be a couple of wins
apart, they see the Orioles on opposite
ends of the playoff picture. FanGraphs’
ZiPS pegged their projected season
record at 90-72 ahead of spring training, good enough for a repeat AL East
title. PECOTA, released annually by
Baseball Prospectus, sees Baltimore as
most likely to finish 87-75, fourth in the
division.
Crucial to understanding how each
system landed on these numbers is that
projections are largely dependent on playing time estimations. ZiPS, which stands
for sZymborski Projection System named
by creator Dan Szymborski, and PECOTA,
short for Player Empirical Comparison and
Optimization Test Algorithm, both develop
a baseline for each player based on previous seasons, age and historical comparisons and project how they will perform if
they play a full season.
Of course, not every player is going to suit
up for 162 games or make 30 starts. Few, if
Orioles top prospect Jackson Holliday, right, talks with first baseman Ryan O’Hearn during a pitching change in a Grapefruit League game at
Ed Smith Stadium in Sarasota, Florida. KENNETH K. LAM/STAFF
any, do in today’s modern era. ZiPS and
PECOTA each have their own methods to
estimatehowmanygamesaplayerwillseein
agivenseason,butnosimulationcanpredict
the randomness with which injuries occur
and how teams will use all their players.
“The question that people want
answered is they want to know how many
wins they’re going to have on Oct. 1, and
that’s not the question we’re really answering,” Baseball Prospectus’ head of quantitative modeling Jonathan Judge said in a
phone interview. “We are basically telling
you, if everything stays the same as it does
now and it plays out, then this is how it
would play out. But obviously, that’s never
how it’s going to work.”
The win totals published on FanGraphs
and Baseball Prospectus each year aren’t
hard numbers that are supposed to reflect
one single prediction. Rather, the number
is just the tip of a bell curve representing
the most likely outcome while showing
how likely it is that a team falls short or
exceeds their projection.
Turn to Projections, Page 13