How Many MLB At-Bats Before You Can Trust a Stat

It depends on the stat. A hitter's strikeout rate starts to mean something after about 60 plate appearances, roughly 14 games for an everyday player. Batting average needs about 910 at-bats, which is closer to a season and a half. Those are the stabilization points FanGraphs lists from Russell Carleton's research, and they're why a hot April tells you less than it feels like it does.

The trust points for hitters

Here are the main ones, with how long each takes for an everyday player at 4.2 plate appearances a game, the 2026 average for a spot in the order, per Baseball-Reference.

The fast stats are the ones the hitter and pitcher decide between themselves, and anything that waits on where a ball lands takes longer.

Carleton's line is the point where a sample of that size lines up with another sample of the same size at a 0.7 correlation, which works out to roughly half signal and half noise. FanGraphs stresses that nothing flips at the line: a stat keeps getting more stable the more of it you see.

Batting average is slow because balls in play are noisy

Batting average has to wait on every ball in play, and whether one drops for a hit depends a lot on where it goes and who's standing there. The same FanGraphs list puts BABIP, a hitter's average on balls in play, at 820 balls in play before it means much. Strikeouts and walks don't involve a fielder at all, which is why they settle first.

FanGraphs gives an example. Every April some hitter with a poor track record posts a .380 wOBA over three weeks, about 85 plate appearances, and people call it a breakout. FanGraphs' answer is that a .310 wOBA hitter can do that over a month "no problem," with a couple of lucky bounces and a well-timed cluster of hits.

FanGraphs' advice is to update a little. Those 85 plate appearances give you some idea the hitter might be improving, and they aren't enough to change your mind completely. Blend the new numbers with the track record, and let the track record carry more of the weight until the new sample clears the lines in the table.

Even the best hitters fail most of the time

The league's on-base percentage in 2026 was .318, per FanGraphs, so the average plate appearance ended without the hitter reaching base about 68% of the time. Yordan Alvarez led all qualified hitters in both batting average (.316) and on-base percentage (.432), and he still failed to reach in about 57% of his trips.

That failure rate is what makes single games so swingy. Even a true .300 hitter goes 0 for 4 about one game in four, just from the math of four at-bats (0.7 multiplied by itself four times is about 0.24). A good or bad week of those games is mostly noise.

Hot streaks usually fade

A hot week is about 25 plate appearances for an everyday player, less than half of what even strikeout rate needs.

Ozzie Albies is a 2026 example. Through April he was hitting .323, 40 for 124, per Baseball Savant. Statcast's expected batting average, which grades each ball by how hard and at what angle it was hit, had him at .264 over that stretch. By season's end he was at .237. The contact never backed the average, and the average came down to meet it.

It works in reverse, too. Alec Bohm hit .151 through April, 16 for 106, while his expected average on Savant sat at .218. He climbed to .249 by the end of the year.

So during a streak, check the numbers that settle fast. If a hitter's strikeout rate has changed over 60-plus plate appearances, take it seriously. If only the batting average moved, wait. For power, barrel rate tells you whether the home runs are backed by hard contact. Separating a real change from a lucky week is most of the craft of predicting players.

For pitchers, strikeouts settle in weeks and runs take years

The FanGraphs list covers pitchers too, counted in batters faced. Strikeout rate settles at about 70 batters and walk rate at about 170. Starters faced 21.4 batters a start in 2026, per FanGraphs, so that's three or four starts for strikeouts and about eight for walks. Home run rate needs 1,320 batters faced, more than 60 starts. BABIP allowed needs 2,000 balls in play, which FanGraphs says "can take three years."

That's why a pitcher's April ERA tells you so little. ERA rides on balls in play and home runs, the slowest-settling pieces on the list, plus the order things happen in. FIP strips most of that out, and it was the better April read on José Soriano: a 0.84 ERA over seven starts through April 2026, but a 2.84 FIP, per the FanGraphs March/April split. His season ended with a 3.91 ERA and a 3.90 FIP.

Strikeout and walk rates are the first parts of a pitcher's line worth believing. A pitcher's swinging strike rate gets there after about 400 pitches, or roughly five starts.

Velocity is the exception to all of this. It's measured on every pitch, so a change shows up at once. If a starter who averaged 95 mph last year sits at 93 over their first two starts, something changed, and Baseball Savant shows it pitch by pitch.

How a hot week changes a hits pick in GAGE

Say a hitter with a .260 average over the last three seasons goes 9 for 20 this week. The GAGE menu has them at "Hits 2+" with YES at 28%. Here's what each sample says, using a rough model of four at-bats in which each one is a hit or it isn't.

The hot week makes YES look like a steal, but 20 at-bats is about a third of what even strikeout rate needs. What could justify YES is a reason the odds might be missing, like a move to the top of the order or a hitter-friendly park. Without one, the card's 28% is about right, and YES, at +7.2 if they get two hits and −2.8 if they don't, has no edge. The scoring page has the full rules, and GAGE is on the App Store when you want to make the pick.