About this report
Yesterday, I took a look at what would happen if you applied the salary cap and floor the MLB owners proposed on May 28, 2026 to the regular season as of that date. While I initially planned to look at it in terms of how much payroll clubs would have to slash or add, it turned into an analysis of competitive balance because the data showed that teams below the floor were very competitive and there really wasn't that big of a difference between clubs that were above the cap to ones that were below the floor.
Today I want to dig deeper into correlations between spending and success. So, this report compares MLB opening day payrolls with regular season team records from 2015 through 2025.
The question is simple:
Does spending more money reliably produce more wins?
This time the answer is really messy and so this report is actually quite hard to write. Payroll clearly matters. The Dodgers, Yankees, Astros, Red Sox, Mets, Phillies, and other high-spending teams have been successful. Having money helps teams keep stars, pretend their mistakes didn't happen, add depth, and fill roster holes seemingly at will.
But payroll does not explain everything. Low-payroll teams win. High-payroll teams lose. Some teams turn modest payrolls into playoff seasons while others drag enormous payrolls into the cellars of their respective divisions. You don't have to think back far to find examples of this. It's even happening this season - the Tampa Bay Devil Rays are first in the American League despite having one of the lower payrolls and the Mets are near the bottom of the National League despite having one of the highest.
So that is why this report looks at payroll, win percentage, raw wins, payroll rank, record rank, playoff results, cost per win, and wins per $100 million. No number really answers the question and the summary stats just make it a lot messier.
Why 2020 has a toggle
The 2020 MLB season was real baseball history, but it was not a normal year by any means.
It had a 60-game regular season. It used an expanded playoff format. Payrolls were much smaller because of the shortened season, and a hot month that would even itself out over a 162 game season could make a team look like a machine. Whereas a lot of good teams got hit by the injury bug or got cold at the wrong time and ended up looking a lot worse than they could have.
So this report includes 2020 by default, but gives you a toggle to remove it.
That matters because 2020 can distort:
- cost per win
- playoff team counts
- best and worst single-season examples
- raw payroll/wins correlations
- wins per $100 million
- top-10 and bottom-10 payroll playoff rates
The toggle doesn't hide the data, but if you watch the summary stats carefully while you toggle it on and off, you'll see what a big impact it has on the data. Without it though, you have ten complete seasons of data.
How to read the summary
The summary section is calculated from the active dataset. If 2020 is included, the summary uses 2015 through 2025. If 2020 is excluded, the summary uses every season in that range except 2020. You likely figured that out on your own but I've always had a problem with brevity. It all started one day, it was warm, Billy Martin got mad and...
The main summary fields are:
- Payroll/win percentage correlation: how strongly payroll and winning percentage move together.
- Payroll rank/record rank correlation: how strongly payroll rank and record rank align.
- Team-seasons analyzed: the number of team-season rows currently included.
- Average opening day payroll: the average payroll across the active rows.
- Average wins: the average raw win total across the active rows.
- Overall cost per win: total payroll divided by total wins.
The headline correlation uses win percentage rather than raw wins because 2020 had a 60-game schedule. Raw wins are still useful, especially for cost per win and wins per $100 million, but they make the shortened season act really weird when it's compared with all the other years. Or maybe I just screwed something up. Who knows?
Correlation is useful, but as I learned in University, correlation is not pause Asian. Or something. I skipped a lot of class. It's also not magic. And just because two things are strongly correlated it doesn't mean that one causes the other. They just move together. They could be caused by something else.
A positive correlation means payroll and winning percentage tend to move together. It doesn't mean that teams succeed because they spend too much. And it doesn't mean that teams should spend less money (unless they're the Los Angeles Dodgers) because it's just as easy to get into the playoffs if you spend $80,000,000 as if you spend $300,000,000.
How to read the contradiction cards
The second row of summary cards looks for useful friction in the data.
The key questions are:
- How often did top-10 payroll teams make the playoffs?
- How often did top-10 payroll teams have losing records?
- How often did bottom-10 payroll teams make the playoffs?
- How often did bottom-10 payroll teams have winning records?
This is why baseball is really cool.
If every top-10 payroll team won and every bottom-10 payroll team lost, the game would be really boring. Spending would explain nearly everything and an owner with deep enough pockets could purchase a very short World Series window. But the data doesn't show that. The expensive teams are often good, but they are not immune to disastrous seasons. The cheap teams are often constrained, but they're not doomed.
And so somehow, fans suffer. From watching your expensive team get whooped by the cheapest, to a brief glimmer of hope when your low budget team makes the playoffs one year. And then sucks for the next three seasons. Baseball can be cruel.
How to read the table
The full table uses one row per team-season.
The main fields are:
- Season: the MLB season.
- Team: the current normalized team name, with historical name where useful.
- League: American League or National League.
- Record: regular season win-loss record.
- Win %: regular season winning percentage.
- Payroll rank: rank within that season, where 1 is the highest payroll.
- Record rank: MLB-wide regular season rank based on winning percentage, wins, and run differential.
- Opening day payroll: payroll from the normalized SteveTheUmp data.
- Cost per win: payroll divided by wins.
- Wins per $100M: wins scaled by payroll.
- Payroll-record gap: payroll rank minus record rank.
- Run diff: runs scored minus runs allowed.
- Playoff result: whether the team made the playoffs or won a title.
A negative payroll-record gap means the team ranked higher in payroll than it ranked in record. A positive gap means the team ranked better in record than it ranked in payroll.
For example, a team with payroll rank 2 and record rank 16 has a payroll-record gap of -14. That means it spent like one of the biggest teams but finished closer to the middle of the league. A team with payroll rank 27 and record rank 4 has a payroll-record gap of +23. That means it performed far better than its payroll rank.
And so with this in mind, I would like to propose a new drinking game. If your favourite team has -15 or lower in a season, drink. If your favourite team has a +15 or higher, you buy.
Methodology
The report uses two main sources:
- SteveTheUmp MLB opening day payroll data.
- Lahman Teams.csv regular season and playoff data.
The data was normalized into one row per team-season. Team names were adjusted where needed so historical names and current names can be handled consistently. For example, the Cleveland Indians and Cleveland Guardians are treated as one franchise, while the report can still preserve the historical team name for a specific season. Same with the Oakland Athletics and the Athletics.
The calculations are really simple because I didn't want to spend all day on this. But if you do, open up your browser tools, go to network and grab my json. You can even format data.json into a more readable format here.
Payroll rank comes from the normalized payroll data. Record rank is calculated across MLB within each season using winning percentage, then wins, then run differential. Cost per win is payroll divided by wins. Wins per $100 million is wins divided by payroll in $100 million units. If you hated math class, don't worry I only have one sentence left. The headline payroll correlation uses winning percentage instead of raw wins so that 2020 doesn't completely mess everything up. Wouldn't it be nice if it was actually that simple? :)
This report does not try to measure prospect value, player development, injuries, deferred money, midseason trades, free agent opportunity cost, market size, broadcast revenue, ownership willingness, or whether a front office hated their fans with such vehemence they decided to ruin all their hopes and dreams. It's just money and wins.
Accessibility notes
This report is built around text and tables first.
The summary cards are useful, but the full table contains the real data. The table can be searched, filtered, and sorted. The result count updates after filtering. The controls are normal form fields with labels. The table uses a caption, column headers, and text values instead of requiring chart-only interpretation.
This page was scanned for accessibility with Siteimp and tested with NVDA.
I am still learning accessibility, so mistakes are possible. If something does not work with your screen reader, keyboard, zoom settings, or browser setup, please contact me so I can fix it.
What did we learn?
Payroll matters, but payroll is nowhere near a guarantee.
The highest-spending teams often do well, but they still produce expensive disappointments. Lower-payroll teams often face real constraints, but some of them produce winning seasons, playoff seasons, and excellent wins-per-dollar results.
The practical lesson is not that money is meaningless. The practical lesson is that payroll is one important input in a much larger system.
Baseball teams are not only payrolls. They are development systems, scouting systems, medical systems, decision systems, and coaching systems. Success is about culture, and interpersonal problems between players can derail the most promising clubs. And then there are injuries. Garret Cole has the most 10k+ games since 2020 and he missed most of last season. The Yankees are much more dangerous with him in the rotation. Without him? Their staff sometimes looked exposed.