Does an MLB batter starting actually matter for his prop?

Almost every player prop screen shows the same thing: how often a hitter went over a line in his last ten games. Almost none of them tell you whether he started those games.

That distinction is worth more than it sounds, and the reason has nothing to do with ability.

The number the whole thing rests on

Joining our game logs to MLB boxscores across 2,084 games gives 29,019 player-games where we know for certain whether the batter was in the starting lineup:

A starter gets roughly four trips to the plate. A bench bat who enters late usually gets one.

A 0.5 line asks whether something happens at least once. So before anything about swing quality, contact or matchup, the bet is largely a question of how many chances there are for it to happen. Four chances versus one is most of the answer.

What that does to each market

Across the same 29,019 games:

Hits and total bases being identical is the arithmetic, not an error. Total bases only accrue from hits, so a hitter has at least one total base exactly when he has at least one hit. On the 0.5 line they are the same event, which is worth knowing for its own sake and which we covered separately in hits versus total bases.

The part most articles would stop at, and should not

Those gaps are real but they are not a clean measure of what starting is worth, and it would be easy to imply otherwise.

The problem is that the two columns contain different people. The players who come off the bench are, on the whole, not the same players who start every day, and they are not as good. So a 40 point gap is measuring two things at once: fewer chances, and a weaker population of hitters.

To separate them you have to compare players against themselves.

Comparing players to themselves

Taking only the 218 players with at least ten classified games and at least three in each role, and comparing each player’s started-only rate to his own blended all-games rate:

About five points, typically. That is the honest figure for what blending costs on a mixed-role hitter, and it is a lot smaller than 40.

Five points is still enough to matter on a market priced in the fifties and sixties. But the average hides the shape of it, because this effect is concentrated rather than spread evenly.

Where it actually bites

The players with the most mixed roles are distorted far past the average. Total bases over 0.5, started-only against the blended number the screens show:

Derek Hill’s screen number says 39 percent. In the games he actually started it was 64 percent. Those are not the same player as far as a prop is concerned, and no amount of looking at his last ten games tells you which one is playing tonight.

Whitcomb is the case that started this. Someone on our side pulled up his page, saw a total bases hit rate in the low twenties, and noticed more than half the games behind it were single trips to the plate off the bench.

So how much should this change what you do

Less often than the table suggests, and more when it applies.

Bench appearances were 9.8 percent of all classified player-games. For a regular who starts every night this changes nothing, and most of the board is regulars. The effect lives entirely in the players with genuinely split roles, and for those players the posted rate can be twenty points away from the one that applies.

Two honest limits worth stating plainly:

This is a description, not a system. These are the games that happened. We have not published a hit rate to bet, an edge, or a return, and nothing above should be read as one. What it changes is the denominator you are reading, not whether the price is wrong.

Lineups post late. Before they are out, whether a given hitter starts is unresolved for you and for the book. A line priced pre-lineup is priced on that same uncertainty. Knowing which side of this split a player usually falls on is useful; assuming tonight is not.

Seeing it on a player

Every player page in the Lab now splits the chart three ways: all games, started, and bench. It uses the same MLB boxscore flag as the numbers above, so the started-only rate you see is computed on exactly the games the hitter was in the lineup.

If a hitter has a mixed role, the three numbers will not agree, and the gap between them is the thing this article is about.

Open the Lab and pick any hitter with an unclear role. The toggle is on the chart.

If you have just read that a ten-game sample is thin to begin with, that is a separate and equally real problem, and the two compound: a short window made of the wrong kind of games is worth very little.