In 2016, the Texas Rangers finished 95-67, the best record in the American League. They scored only eight more runs than they allowed all season. Those two facts don’t usually go together, and what happened next is one of the cleanest lessons in sports data.
The Record and the Run Differential
A team’s win-loss record tells you what happened. Its run differential, runs scored minus runs allowed, tells you how well it actually played across every inning. Most of the time the two line up. When they don’t, the record is usually the one that is lying.
The 2016 Rangers were the extreme case. MLB.com noted that their +8 differential ranked only 14th in the majors, behind teams that finished well below them in the standings. The Chicago Cubs led baseball at +252 and won the World Series. The Rangers’ edge came almost entirely from close games: they went 36-11 in one-run games, a .766 winning percentage, according to the Dallas Morning News.
The Formula Behind the Doubt
Baseball analyst Bill James popularized a simple way to turn runs into expected wins, known as the Pythagorean expectation: square a team’s runs scored, then divide by the sum of runs scored squared and runs allowed squared. The result estimates the winning percentage a team “should” have, based on how many runs it produced and gave up.
By that logic, a team that barely outscored its opponents should finish near .500. The Rangers finished 28 games over. The gap had to come from somewhere, and one-run games are where luck lives in baseball. A bloop single, a bad hop or a wind-blown fly ball decides them more often than talent does.
What Happened Next
Close-game magic rarely carries over. The Dallas Morning News found that of 100 teams through 2015 with the best one-run records, only three matched or beat that mark the next season.

The Rangers didn’t. By July 2017 they were 6-14 in one-run games, the worst mark in the majors, and on pace for the biggest drop in one-run winning percentage in the modern era. They finished the season 78-84 and missed the playoffs.
| Measure | Result |
|---|---|
| Rangers in one-run games, 2016 | 36-11 (.766) |
| Rangers in one-run games, 2017 (as of July 6) | 6-14 (.300), worst in the majors |
| Top 100 one-run teams through 2015 that matched or beat it the next season | 3: the 1906-07 Cubs, 1953-54 Brooklyn and 1980-81 Cincinnati |
| Modern-era drop the 2017 Rangers were on pace to pass | 1934 Yankees, .650 to .341 the next season |
| Season | Record | Runs Scored | Runs Allowed | Expected Record* |
|---|---|---|---|---|
| 2016 | 95-67 | 765 | 757 | 82-80 |
| 2017 | 78-84 | 799 | 816 | 79-83 |
Whatever else changed in 2017, the one-run luck ran out. That is what statisticians mean by regression to the mean, and it is one of the most reliable patterns in sports.
Why the Coverage Usually Misses It
A 95-win team gets covered as a 95-win team. Close wins get described as clutch, resilient or well-managed, all words that suggest a skill that will last. During the 2016 award season, MLB.com ran a column headlined “Banister beat odds, deserves Manager of Year Award,” crediting Rangers manager Jeff Banister.
Some of that credit may have been earned. But a run differential of +8 over 162 games is a warning light, and it was visible in the standings all season for anyone who looked.
How to Use It
The tool works in every sport, under other names. Point differential in the NFL and NBA, goal difference in football and hockey. Any time a team’s record runs far ahead of its scoring margin, expect the record to come back toward the margin, not the other way around.
So the next time a team is praised for “finding ways to win” close games, check one number before you believe the story: did it actually outscore its opponents by much?
