Two games, one roster
Every time you open a DFS lobby, you face the same fork: cash games or tournaments. They look like variations of the same contest — same slate, same scoring, same format — but playing them the same way is one of the most expensive habits in daily fantasy baseball.
Cash games pay about the top half of the field. You need to beat the median. A safe lineup that consistently lands in the upper half makes money over time; a spectacular lineup that crashes when its stack goes 0-for-4 does not.
Tournaments (GPPs) are the opposite structure. A tiny fraction of entries split most of the prize pool. Beating the median is irrelevant — you need to be in the top one or two percent to cash meaningfully, which means you need a ceiling, not a floor. A safe lineup that finishes 40th percentile doesn't pay you anything.
The same player can be a cash-game anchor or a GPP liability depending on which game he fits. Getting this backward isn't a strategy failure — it's a format failure, and it happens every night.
The floor — what the numbers actually say
We grade our projected DraftKings points against real box scores every night. Across more than 24,000 graded hitter results — players who were salary-slotted in a real lineup and appeared in the lineup confirmation — the distribution says something most DFS advice buries:
27 percent of those hitters scored zero.
Not zero as in a rough day at the plate. Zero as in no points — a scratch, an early exit, a manager's day-off decision made at 4:30 PM. The salary was printed, the player was on your roster, and the points never showed. The 15th percentile is also zero. A quarter of the field is below three points on a typical night.
The median hitter scores five points. That's the middle of the field — not the floor, not the ceiling, just the realistic expectation for a player who plays his normal game.
This is why picking five hitters you feel good about fails cash games. Even with reasonable projections, one zero in your lineup puts you at a structural disadvantage. Two zeros and you're probably out of the money. The cash game question isn't 'who do I like?' — it's 'who can't possibly score zero?' That's a much shorter list.
Starting pitchers are the real floor play
Across more than 2,500 graded starting pitcher results, the median score is 13.4 DraftKings points — more than twice the median hitter. That gap isn't an accident.
Strikeouts pay two points each, and an average start generates five or six of them. Innings compound: six strong innings at 2.25 points each is 13.5 before the strikeouts even count. A hitter who goes 1-for-4 with a run scored is having a decent day — seven points, you'll take it. A starter who goes six innings and strikes out six has 26.5 on his own, and that's a quiet outing.
The 85th-percentile starting pitcher scores 24.6 points. The 85th-percentile hitter scores 14. The better arm is the more reliable path to that number because his scoring is built on volume — innings and strikeouts — rather than a few discrete events that may or may not happen.
There's real downside risk too. About 11 percent of starters post negative scores — a short, rough outing where earned runs subtract faster than innings add. The solution isn't to avoid pitching spend; it's to pick arms with favorable matchups. In our data, the highest-salaried starters averaged 15.4 points against 11.1 for the mid-range and 6.2 for the cheapest arms. Pitching is the one position on the roster where salary consistently predicts production.
The stacking math
Here's why the cash-vs-GPP divide matters so much in baseball specifically.
DraftKings pays a home run 10 points. But a home run almost never arrives alone. Say your cleanup hitter hits a three-run shot with the two batters ahead of him on base. Here's what happens:
The runner who scores from first: two points. The runner who scores from second: two points. The cleanup hitter: 10 for the home run, two for his own run, and two for each of the three RBIs — 18 points from one swing. A single pitch just put 22 DraftKings points on the board, distributed across three roster slots.
That's the stacking argument. When you roster four or five batters from the same team, a big inning pays multiple spots at once. A solo homer from your stacked hitter is worth 12 points. A three-run shot with two of your other hitters aboard is worth 22, spread three ways. The correlation works in your favor instead of against you.
This is also exactly why stacking is the wrong move in cash games. Correlation means those same three players all get zero if the pitcher on the other side shuts the lineup down. A big inning is a low-probability event on any given night. You're building a lineup that wins 15 percent of the time and wins big — which is GPP math, not cash math.
Only 6 percent of hitters in our graded results score 20 or more points in a night. The ones who do are almost always part of a team that had a big inning. Getting one of those nights into your lineup is a tournament strategy, and building lineups with correlated upside is the way to do it.
Building each lineup differently
Cash game target: minimize zeroes. Start with two starting pitchers in good matchups — the salary tends to signal this, and the production numbers back it up. Fill the lineup with hitters who are confirmed in their spots, batting in the top six of the order where they'll see four or five at-bats. Avoid day-to-day players and anyone whose lineup status isn't final before lock. Boring is correct — you're trying to reliably beat the median, not swing for the fences.
Tournament target: maximize ceiling. That means committing to a stack — four or five hitters from a team projected to score, ideally against a starter with recent struggles. It means accepting that most of your entries will finish below the median, because the ones that catch a big inning will finish in the range where the prizes are. Spreading your set across two or three different stacks so one shutout doesn't end your night.
One rule that holds in both formats: don't roster two starting pitchers from the same game. They pull in opposite directions. Every run one gives up helps the other. Every out one gets is one fewer plate appearance for the lineup the other needs to handle. You're capping your own ceiling.
What this doesn't show
A distribution of scored points can't tell you which players tonight will land where in it. The 6 percent who score 20 or more are spread across the whole field on any given night. Identifying the ones with the best chance of being in that group — based on the matchup, the opposing arm, the weather, the lineup spot — is what the projection is trying to do. The distribution is just the shape of the game those projections are targeting.
The ownership dimension is missing from any single-lineup analysis. In a large-field tournament, having the same players as everyone else blunts the upside of being right. A 30-point night from the most popular player in the field moves every lineup, not just yours. Real GPP leverage comes from being right on a player most of the field avoided, which means taking positions that feel uncomfortable at lineup lock.
And these numbers come from a single season of data. Scoring distributions are fairly stable in DraftKings baseball, but the specific percentiles shift a few points depending on slate size, parks, and weather. A 15-game main slate and a 4-game afternoon slate are different environments, even with the same format and the same scoring table.
The floor is real. So is the ceiling. Neither is guaranteed — which is exactly why the format decision matters as much as the player picks.
Sources
Every number here comes from the graded ledger the boards publish. Research from a statistical model, not betting advice; no outcome is guaranteed. Method: how MatchWiz works.