The math is simple. Three earned runs over six innings works out to 4.50 per nine — the formula for ERA, unchanged since Henry Chadwick devised it in the 19th century. A pitcher who posts 4.50 over a full season in 2024 would have been worse than the league average of 4.07, which itself sat just below the 4.39 runs per game MLB teams scored. By any modern standard, 4.50 is a bad ERA.
Yet that same line — six innings, three earned runs — qualifies as a “quality start,” a statistic invented in 1985 by John Lowe, then writing for the Philadelphia Inquirer, who said it “shows exactly how many times a baseball pitcher has done his job.”
The apparent contradiction, raised this week in a widely discussed baseball forum post, has been kicking around since the stat’s invention. Tim McCarver dismissed it on air years ago: “Three runs for six innings … means the ERA would be 4.50 … Rubbish.”
The resolution is mostly about distributions. The quality start’s floor — exactly six innings, exactly three runs — is a boundary, not a typical outcome. Bill James tackled this in his 1987 Baseball Abstract, arguing that the hypothetical worst-case quality start was extremely rare. A computer analysis of every quality start from 1984 through 1991 found the average ERA within those starts was 1.91, not 4.50. Most outings that clear the bar clear it comfortably.
But the floor still creates logical problems. A pitcher who throws nine innings and allows four earned runs posts a 4.00 ERA — better than the 4.50 floor — yet fails to record a quality start. Former pitcher Carl Erskine put it bluntly: “In my day, a quality start was a complete game … you gave everybody a day’s rest.”
The criticism has persisted into the analytics era. In 2022, John Laghezza of The Athletic called the quality start “garbage” and proposed replacements — the “money start” and the “plus start” — with stricter thresholds.
The deeper issue is one that anyone who has filled out a box score understands: single-game lines and season-long rates measure different things. A 4.50 ERA over 180 innings tells you a pitcher is probably getting hit hard most nights. A 4.50 ERA over one start tells you almost nothing — it could be a grind-through-it outing in a 7-3 win, or the back end of a doubleheader where the bullpen was spent. The quality start tries to capture the former: did the starter give his team a chance? The problem is that its minimum threshold happens to sit below the league average, which makes the name feel generous.
In 2024, 11 of 30 teams posted team ERAs above 4.07. The Colorado Rockies, playing at altitude, finished at 5.47. The Chicago White Sox, who lost 121 games, came in at 4.67. A pitcher who matched the quality-start floor every time out would have been worse than 19 teams’ entire staffs.
The stat survives because, in aggregate, it still separates good starters from bad ones. Greg Maddux had 24 quality starts in 25 games in 1994. Framber Valdez rattled off 25 consecutive quality starts in 2022. Nobody is confusing those seasons with a string of 6-and-3s.
But the name oversells the floor. A 4.50 ERA is a 4.50 ERA, whether it spans one start or thirty. The quality start just happens to count the games where a pitcher cleared a bar that, at its lowest point, sits a few inches below average.
Sources
- Wikipedia — Quality Start
- Wikipedia — Earned Run Average
- Baseball-Reference — 2024 MLB Standard Pitching