The Orioles Are Still Here
Picked last in the AL East despite their inspiring '22 run, Baltimore is proving the predictions wrong... again.
The 2022 season was something of a coming-out party for the Baltimore Orioles, after years of irrelevance at best (and dreadfulness at worst). While the O’s found themselves sitting 11 games under .500 by mid-June, a 21-9 run over the next 30 games — including a 10-game winning streak — jolted the team right into the playoff conversation, where it would stay for the rest of the season. Even though Baltimore ended up falling a few games short of its first postseason bid since 2016, last year felt like the start of a brighter future for Charm City baseball.
Or at least, that was before the preseason forecasts came out for 2023. Despite Baltimore’s relatively successful 2022, FiveThirtyEight’s composite forecast pegged the Orioles to finish last in the AL East this season with a 75-87 record. None of the component systems that went into their composite picked Baltimore to win more than 78 games. And my own series of simulations using Out Of The Park Baseball 24 largely agreed, calling for the O’s to win 78.2 games and finish in the basement of the division.
And yet, the Orioles keep defying the odds.
As of Thursday morning, Baltimore is 20-10 — good for the third-best record in baseball — with MLB’s seventh-best runs-per-game differential (+0.77) to back it up. Compared with its preseason forecast at FiveThirtyEight, Baltimore has nearly doubled its playoff odds (from 16% to 30%) despite playing in the division with baseball’s best combined record (the AL East is 98-58). With the possible exceptions of the dominating Tampa Bay Rays and the surprising Pittsburgh Pirates, Baltimore is the feel-good story of the MLB season… for the second season in row!
History rarely repeats itself that way. Longtime sabermetric readers will know about the phenomenon Bill James called the “Plexiglas Principle”, which is essentially just a manifestation of regression to the mean. Teams that improve a lot in one season tend to fall back toward their previous records some (not all) of the way in the following season. Exceptions to this rule do exist, but they are few and far between in baseball history.
Before the Orioles improved their record by a whopping 31 wins from 2021 to 2022, there were 52 AL or NL teams since 1901 that played at least 100 games in back-to-back seasons and improved by at least 25 wins per 162 games. Those teams averaged an improvement of 28.8 games between seasons, rising from 63.1 wins per 162 to 91.9 on average. (Not too terribly different than Baltimore’s rise from 52 wins in 2021 to 83 in 2022.) Of those 52 teams, 14 (or about 27%) went on to improve again the following season — led by the early-1900s New York Giants, who went from 56.3 W/162 in 1902 to 97.9 in 1903 and then an incredible 113.0 W/162 in 1904. Good ol’ John McGraw sure knew how to build up a winner.
But that means 38 of our 52 teams (or 73%) fell back at least some in the following year — with 23 of 52 (44%) giving back 10 or more games, and 8 of 52 (15%) giving back 20 or more games. So on average, our group that had improved by 28.8 games on average between Year 1 and Year 2 ended up declining by an average of 7.3 games in Year 3, going from 91.9 wins per 162 back down to 84.6. That’s the Plexiglas Principle in action: what goes up usually must come back down.
But the current Orioles are bucking that trend in a huge way. After improving by 31 wins per 162 between Years 1 and 2 of their sample, they are up another 25 (!) wins per 162 so far between Years 2 and 3. If that were to hold up, it would obliterate what McGraw and those 1904 Giants did, with their mere improvement of “only” 15.1 wins per 162 from Year 2 to Year 3.
Obviously, it’s pretty optimistic to think the Orioles will keep playing at a 108-win pace going forward — no matter how great Cedric Mullins, Adley Rutschman, Jorge Mateo and Yennier Cano look so far. (This is the part where I also note that Baltimore has played MLB’s easiest schedule by average opponent Elo.) But even if Baltimore instead wins at a rate commensurate with its underlying BaseRuns-predicted winning percentage over the rest of the season, a 94-win campaign would represent an 11-win improvement over 2022, and the sixth-best follow-up campaign for anyone in our historical sample of rising teams.
(And in fact, after that 20-10 start, Baltimore simply needs to go better than 63-69 over the remainder of the season to ensure it joins the relatively exclusive historical club of teams that improved by 25+ W/162 in a season and then improved again, by any amount, the following season.)
Simply put, the Orioles are defying the projections because the projections count on the overriding rule of regression to outweigh the possibility than any one team is an exception — even one that was the fourth-youngest team with Baseball America’s fourth-best farm system last season, aka the type of team that might be predisposed to break out and then keep its momentum going forward. But for now at least, Baltimore is proving that the exceptions to our historical rules are actually what make baseball fun and unpredictable.
Filed under: Baseball