Preseason MLB Regression Report: Identifying MLB Teams Ripe for Regression in 2024

Looking back at final 2023 numbers and setting the table for 2024

Feb 24, 2024; Sarasota, Florida, USA; Baltimore Orioles starting pitcher Corbin Burnes (39) throws a
Feb 24, 2024; Sarasota, Florida, USA; Baltimore Orioles starting pitcher Corbin Burnes (39) throws a / Kim Klement Neitzel-USA TODAY Sports

From traditional statistics such as batting average, to "+" statistics like wRC+ to WAR, you name it and someone has measured it for Major League Baseball.

While I love the numbers and the day-to-day grind of building a statistical profile of a season of baseball, I sometimes wonder if we've overcomplicated the analysis in some respects.

That's why I love the Pythagorean Theorem for Baseball. Give me two easily obtainable numbers, plug them into a formula and you have your answer, in this case, what a team's winning percentage should be.

What is the Pythagorean Theorem for MLB?

This formula is well-known, widely used, and typically comes within a few games of a team's actual record at the end of the season, which is remarkable when you think about it.

But I wondered if it was too simple and began tinkering with a calculation that combines the Pythagorean Theorem formula with actual wins and losses under the working theory that no matter how much you outscore your opponents you can never erase a loss.

Eventually, I settled on a formula that works well in identifying teams that are overperforming or underperforming given their run differential and actual won-loss record at any given time.

The comparison of these two measures for the 2023 season is below, specifically the two middle columns.

The number we're after for this analysis is on the far right and is simply actual wins minus the "MLB Regression Rpt xWins" and is an estimate of the number of games a team has overperformed (if positive) or underperformed (if negative) by at the time of the report.

The report identified the Marlins as a regression candidate fairly early last season, as they racked up one-run wins at an incredible rate.

It turns out the Marlins did eventually regress in August, but still finished the season 33-14 in one-run games, which impacted their Pythagorean expected win percentage greatly, resulting in an 8.79 game difference between actual wins and Pythagorean expectation.

Which MLB Teams Will Regress in 2024?

The MLB Regression Report wasn't perfect but was closer to the Marlins' actual number of wins because, as mentioned, it takes into account the actual wins and losses as part of the formula.

Candidates for negative regression (won more than they should have) are generally near the top of the report and teams for positive regression (won less than they should have) are generally at the bottom of the report.

Teams that have a very high or low one-run game-winning percentage (Marlins and Padres in 2023, for example) tend to be at the top or bottom of the report, respectively.

Over time we would expect those one-run games to even out.

But that's not the only way a team stands out on this report.

The Kansas City Royals were 21-20 in one-run games in 2023 and only won 56 games, but are widely viewed as a team that will be improved in 2024.

The Pythagorean formula placed their expected wins almost 8 games higher, while the MLB Regression Report saw some upside, but was a bit more conservative because of what happened in the win-loss columns.

It's also important to note that this report should be viewed as a starting point for such an analysis, not the answer in and of itself.

The Royals largely added pieces to a team that should have won more games last season and is primed for marked improvement in 2024.

The consensus is the Marlins are headed for a tough season, in part because of personnel changes, but also because they played such a high percentage of one-run games last season and the odds of winning 70% of one-run games two consecutive years is small.

Even in a good season, the Marlins margin is thin and therefore subject to more volatility, as shown by the team's 24-40 record in one-run games in 2022.

The Orioles 101 wins in 2023 also seem impossibly high to replicate even before injuries to Kyle Bradish and John Means, though the addition of Corbin Burnes and injury to New York Yankees ace Gerrit Cole are mitigating factors.

Much as there is nuance in handicapping in general, there's nuance in this type of analysis.

There are many nuances in the case of the San Diego Padres. While the data suggests they are headed for a big improvement, the amount of turnover in personnel leaves me uncomfortable in recommending San Diego as a team that will positively regress over 82 wins, though it is difficult to imagine the team not improving on a 9-23 record in one-run games and 2-12 in extra innings.

The roster could be much worse subjectively, yet with some luck the Padres could end up with a better record in 2024, I'm just not willing to bet on it at this point.

On the other hand, with the Padres win total placed at 83.5 and Fangraphs projecting 82 wins, it's certainly not unreasonable to believe the mid-80s in wins is possible.

The plan is to publish this report periodically to assist in identifying teams that are candidates for regression, either positive or negative, providing bettors with another piece of data to use in their analysis.

If you want to bet on some MLB win totals before the season starts, be sure to click the link below to sign up for an account at DraftKings Sportsbook. If you do, you'll receive $150 in bonus bets if when you place your first $5 wager.