Fielding Percentage, Range Factor and Why Older Fielding Stats Mislead
The formulas for fielding percentage and range factor, worked with real shortstop seasons from the historical record, and the reasons both can rank defenders in the wrong order.
Fielding percentage is among the oldest defensive numbers in baseball, and it has a built-in flaw that is easy to state: it can only count what a fielder touched. A shortstop who reaches nearly every ground ball and makes a few throwing errors along the way will score worse than one who never gets to the ball in the hole and therefore never has a chance to make a mistake on it. The statistic measures reliability on the chances a player accepted, not how many chances he created.
Range factor was invented to fix exactly that problem. Instead of dividing by chances, it counts the plays a fielder made, such as putouts and assists, per game or per nine innings. It rewards the shortstop who reaches more balls. It also brings new problems of its own, because the number of balls a fielder reaches depends heavily on the pitchers behind him, the park and the manager's lineup, none of which the fielder controls.
This article gives the formulas, works through them with four real seasons from the historical record, and explains what each number hides. It is aimed at fans who read a back-of-the-card fielding line or a Hall of Fame argument and want to know how much weight that line can bear. The short answer is some, and less than most people think.
The formulas
Both statistics start from the same three counts in a fielder's record: putouts (PO), assists (A) and errors (E). A putout is credited to the fielder who records an out, whether by catching a fly ball, tagging a runner or taking a throw to touch a base. An assist is credited to each fielder who throws or deflects the ball in a way that leads to an out. An error is charged by the official scorer when a fielder fails to make a play that an ordinary effort would have made.
From those counts come two more:
- Total chances (TC) = PO + A + E. A chance is any play in which a fielder had a real opportunity to record an out.
- Fielding percentage (FPct) = (PO + A) / (PO + A + E). It is the share of chances handled without an error.
- Range factor per game (RF/G) = (PO + A) / games played at the position.
- Range factor per nine innings (RF/9) = 9 × (PO + A) / innings played at the position.
The difference between the two is the denominator. Fielding percentage divides by what the fielder had a chance to do, which includes errors. Range factor divides by time, which makes it a rate of plays made per game or per nine innings, and the numerator leaves out errors entirely. The per-nine-innings version avoids a distortion that games produce: a player who appears in 150 games but plays only part of each one will look worse by RF/G than he should. The historical record shows the outs a fielder was on the field for most modern seasons, which allows RF/9 to be computed.
A tiny example makes the arithmetic concrete. Suppose a second baseman has 220 putouts, 380 assists and 6 errors in 140 games. That is 606 total chances. His fielding percentage is 600 divided by 606, which is .990. His range factor per game is 600 divided by 140, or 4.29. Neither number says anything about how many balls went through the infield for hits.
What counts as a putout, an assist and an error
The raw counts shape both statistics, so it is worth knowing where they come from. On a ground ball to the shortstop who throws to first, the shortstop gets an assist and the first baseman a putout. On a fly ball caught by an outfielder, the outfielder gets a putout and nobody gets an assist. On a strikeout, the catcher is normally credited with the putout, which is why catchers have enormous putout totals and why their range factors look huge. A double play can give three fielders credit.
These assignments mean that comparing positions with each other is meaningless. First basemen collect most of their putouts on throws from other infielders, so a first baseman's range factor depends on how well his teammates field. Catchers' totals depend on how often their pitchers strike batters out. Pitchers' assists depend on how many balls they field. Within a position, the figures are comparable; across positions, they are not.
Errors are the subjective part. The official scorer decides whether a play could have been made with ordinary effort. A ball hit to a fielder's glove that he drops is an error. A ball hit just out of reach, so that he has to dive to knock it down, is usually scored a hit, even if he gets a glove on it. A ball that goes through the infield without a fielder touching it is a hit. This is the root of the central flaw in fielding percentage: errors are charged only on plays that the scorer believes a fielder should have made, so a fielder who limits the number of such plays also limits his error count.
The flaw in fielding percentage
Consider what fielding percentage rewards. It rewards a fielder who gets to few balls, because each ball is a risk. It rewards a fielder whose pitchers strike out a lot of hitters, because those hitters never put the ball in play. It treats a throwing error on a difficult backhand play at the edge of range the same as a dropped routine fly ball, if the scorer calls both errors, and it ignores the plays that a fielder did not get to at all.
It also barely separates good fielders from average ones, because the differences rest on a small number of errors. A shortstop with 12 errors in 700 chances finishes at .983, and with 18 errors he finishes at .974. A difference of six errors in a season, which one bad stretch can produce, is nine points of fielding percentage, and over a different season the order could reverse.
The result is that fielding percentage is a good measure of one narrow thing, which is how often a player makes an error when he gets to a ball, and a poor measure of defensive value. It is the stat that makes an immobile fielder with soft hands look excellent.
Four seasons that show the tradeoff
The historical record lets us see the tradeoff in the careers of two Hall of Fame shortstops, Cal Ripken and Ozzie Smith. The table shows four of their seasons, with every figure taken from the database.
| Player and season | Games | PO | A | E | Total chances | FPct | RF/G |
|---|---|---|---|---|---|---|---|
| Cal Ripken, 1984 | 162 | 297 | 583 | 26 | 906 | .971 | 5.43 |
| Cal Ripken, 1990 | 161 | 242 | 435 | 3 | 680 | .996 | 4.20 |
| Ozzie Smith, 1980 | 158 | 288 | 621 | 24 | 933 | .974 | 5.75 |
| Ozzie Smith, 1987 | 158 | 245 | 516 | 10 | 771 | .987 | 4.82 |
Ripken's 1990 season is a striking one. He made 3 errors in 161 games and posted a .996 fielding percentage, the best of his full seasons from 1983 to 1996. In 1984, with the same games played, he made 26 errors and posted .971. By the fielding-percentage ranking, 1990 was his far better season. By range factor, 1984 was: 880 plays made against 677, a difference of more than 200 plays over a season. His RF/G of 4.20 in 1990 is the lowest of any of his full seasons from 1983 to 1993.
Does that mean Ripken was a worse shortstop in his best fielding-percentage year? Not necessarily. The database cannot say why the number of plays fell. It could be a change in his positioning, a pitching staff that struck out more hitters or allowed fewer ground balls, or his own range. The honest reading is that the two statistics disagree and neither settles the question. What the table does show is that a single number can point in opposite directions in consecutive stretches of the same career.
Smith's two seasons repeat the pattern. In 1980 he handled 909 plays and committed 24 errors; in 1987 he handled 761 plays with 10 errors. A fan looking only at fielding percentage would conclude that 1987 was far better. Someone using range factor would say that 1980 was the busier and more productive season, and would be right to ask whether the later season simply had fewer plays available. Both readings are partial.
What range factor fixes, and what it does not
Range factor corrects the obvious problem. A fielder who reaches more balls gets more credit, and a fielder who makes errors while trying for them is not punished as severely. The metric needs only putouts, assists and playing time, which are available for most of baseball's history, so it can be applied almost everywhere. In the decades since it was popularized, it has been a staple of analysts who wanted something better than fielding percentage and could not yet measure where balls were hit.
But range factor assumes that every fielder gets the same number of opportunities per game, which is not true. Several things determine how many balls come to a position:
- Strikeouts. A staff that strikes out more batters generates fewer balls in play, so each fielder gets fewer chances. A hypothetical staff that records 1,400 strikeouts instead of 1,000 removes about 400 plate appearances from the field in a season, and those plays are spread across the lineup.
- Ground-ball and fly-ball tendencies. A sinkerballer's infielders collect more assists than a fly-ball pitcher's, and the outfielders on the second staff collect more putouts.
- Handedness. The proportion of left-handed and right-handed hitters a team faces shifts how many balls go to the left and right sides of the field.
- The park. Foul territory, fence height, turf and grass all change the chances at different positions.
- Teammates. A first baseman's putouts depend on infielders' throws, and an infielder's assists depend on the fielder next to him.
Range factor, in other words, mixes a fielder's ability with the opportunities his team gave him. Analysts developed adjusted versions that compare a fielder's plays with what his team's pitchers produced, and the idea led to more sophisticated methods that track where each ball landed. Those methods, including outs above average, defensive runs saved and UZR, are explained in a separate article, and they exist because range factor could not fully separate fielder from context.
A century of errors
Fielding percentage has another quirk: its meaning has changed over time. The league-wide figure has climbed almost every decade. The table below uses all putouts, assists and errors from major league fielders (including pitchers and catchers) in the historical record, grouped by decade.
| Decade | League fielding percentage | Errors per 100 chances |
|---|---|---|
| 1900s | .953 | 4.7 |
| 1920s | .968 | 3.2 |
| 1950s | .977 | 2.3 |
| 1980s | .979 | 2.1 |
| 2000s | .983 | 1.7 |
| 2010s | .984 | 1.6 |
The 2020 to 2025 seasons, which cover only part of a decade, stand at .985. The drop in errors likely reflects changes that have little to do with individual skill: better gloves, better field surfaces and groundskeeping, and changes in how scorers apply the standard. In the game's early decades, fielders worked on rougher infields with smaller gloves, and errors were frequent enough that unearned runs made up a large share of the scoring.
This matters for comparisons across eras. A shortstop at .971, as Ripken was in 1984, was not far from the .979 league figure for the 1980s. A shortstop at .971 today would sit well below the .985 figure for 2020 to 2025. The league figure includes easier positions such as first base, so it is only a rough guide, but the point stands: fielding percentage has to be measured against the league of its year, not against an absolute scale.
How to use the older fielding stats
Treat fielding percentage and range factor as two imperfect clues. Fielding percentage tells you about reliability on balls fielded, which is a real part of defense, especially at first base and catcher, where range matters less. Range factor tells you about volume, which is useful when two fielders played on similar teams with similar staffs.
Here is a short checklist for reading either number:
- Compare with the league figure for the same year and the same position, never with an absolute standard.
- Look at the size of the sample. A single season has a small number of errors, and a career has a better chance of showing a pattern.
- Ask who pitched. A fielder behind a staff that strikes out a lot of batters will look worse by range factor than he is.
- Be careful with first base, catcher and pitcher, whose counts depend on teammates and pitching rather than range.
- Do not use either number to rank fielders closely. A gap of a few points of fielding percentage is not evidence of a better defender.
For modern players, prefer the newer metrics, which use batted-ball location and tracking data. For players before those data existed, the older numbers are the evidence that exists, and using them carefully, with the checklist above, is better than ignoring them.
Common questions
Is a higher fielding percentage always better? Not as a ranking of defenders. It says a player rarely made errors on the chances he accepted. It cannot say whether he reached balls that others missed.
Why do catchers have such high range factors? Because strikeouts count as putouts for the catcher. A catcher behind a strikeout-heavy staff will have a big putout total that tells you about the pitchers.
Can I compare range factor across positions? No. The plays at different positions are different kinds of events, and a catcher's total is driven by strikeouts while a first baseman's is driven by his teammates' throws.
How are errors decided? By the official scorer, who watches the play and decides if ordinary effort would have produced an out. A scoring decision can be appealed to the league office, and the standard of ordinary effort is not a bright line.
Do these numbers appear in the database? Yes. The historical fielding table lists games, innings played (in outs), putouts, assists and errors for each player, position and season, and the figures in this article were computed from it.
Where to go next
For the modern way of measuring defense, read about outs above average, defensive runs saved and UZR and about catcher framing, which measures a part of defense that fielding percentage ignores. To see how fielding credit appears in a game record, read how to read a box score. You can work out any of the rates in this article with the stat calculator, and the stats and analytics topic page collects current coverage.