Why on-base percentage beats batting average for lineup decisions

Batting average is the number everyone knows and the number most lineups get built from. It also throws away walks, which is a strange thing to do when the whole job of a batting order is avoiding outs.

The two statistics, precisely

The difference is entirely in the denominator, and it is worth being exact about it.

Batting average is hits divided by at‑bats. Walks are not at‑bats. Neither are hit‑by‑pitches or sacrifices. So a walk does not appear in batting average at all — not in the top, not in the bottom. It is as though the plate appearance never happened.

On‑base percentage is times reached — hits plus walks plus hit‑by‑pitches — divided by plate appearances. Every trip to the plate counts, and every way of reaching base counts.

Why this changes the order

Consider two hitters over 100 plate appearances.

• Hitter A hits .300 but almost never walks. In 100 plate appearances he reaches base about 33 times.
• Hitter B hits .250 and walks often. In the same 100 plate appearances he reaches base about 38 times.

Ranked by batting average, A is clearly better and bats higher. Ranked by how often he avoids ending an inning, B is better by five extra baserunners per hundred trips — and it is the second question that determines how many runs your lineup scores.

A batting order is a machine for not ending innings. The statistic you sort it by should be the one that measures not ending innings.

This is not a marginal effect in youth and high school baseball. Walk rates at those levels are high, because pitchers are still developing command. The gap between batting average and on‑base percentage is often enormous — far wider than in professional baseball — which means sorting by average discards more information, not less, the younger the league.

Where slugging comes in

On‑base percentage tells you how often a hitter avoids an out. It says nothing about what happens when he does not. That is slugging percentage: total bases divided by at‑bats, so a double counts twice a single and a home run four times.

The two answer different questions and a lineup needs both:

• On‑base percentage decides who should get the most plate appearances and who should bat in front of your best hitters.
• Slugging percentage decides where your damage is worth the most — which is to say, the slots most likely to come up with someone already on base.

This is why Diamond Data PRO asks for exactly these two numbers per hitter and nothing else. They are sufficient to sequence an order, and they are numbers a coach already has.

What about OPS?

OPS adds on‑base percentage and slugging percentage together. It is a genuinely useful single number for ranking hitters at a glance, and it is reported in the app for that purpose.

For sequencing, though, adding them together loses the thing you need. Two hitters can have identical OPS with completely different shapes — one a high‑on‑base, low‑power hitter, the other the reverse — and those two belong in different slots. Keeping the components separate is what lets the order put each skill where it pays.

The common objections

“Walks are not the same as hits.” Correct — a walk cannot drive in a runner from second, and slugging percentage exists precisely to capture that difference. But for the purpose of not making an out and passing the at‑bat to the next hitter, a walk and a single are close to identical. Both statistics are in the model for this reason.

“My players should be swinging, not taking pitches.” This is a real coaching concern and it is not answered by dismissing on‑base percentage. A hitter who takes good pitches is not being taught passivity; he is being taught the strike zone. Rewarding plate discipline in the batting order is consistent with teaching aggressive swings at good pitches.

“Our sample is too small for any of this.” Partly fair. Rate statistics are unstable over twenty plate appearances, and on‑base percentage is somewhat more stable than batting average at the same sample size because it counts more events. Small samples argue for carrying prior numbers forward and re‑checking often — not for switching to a statistic that measures the wrong thing.

What to do with this

If you take one thing from this piece: stop sorting your lineup card by batting average. Sort by on‑base percentage, keep slugging in view for the middle of the order, and stop burying high on‑base hitters at the bottom because their average looks ordinary.

That single change captures most of the available gain before any software is involved. The optimizer exists for the rest of it — the sequencing decisions that are genuinely too complex to eyeball. How many orders your roster actually has.