Collect current components
Use available current market projections for hitter events that map to the named fantasy scoring system. Missing components are not silently treated as zero.
Method card · Fantasy v1
How the estimate earns its range.
How Propeller estimates MLB hitter fantasy points, builds the floor and ceiling, labels older data, and explains the method’s limits.
Produced through an automated AI workflow. Scott Olmer is the designated author; this byline does not imply personal drafting or human review. Read our editorial policy.
Direct answer
How does Propeller project MLB fantasy points?
Propeller converts available current hitter-stat market components into DraftKings MLB fantasy points, then uses recent game form for scoring components without current market coverage. The central estimate is surrounded by a lower and upper range derived from recent outcome variation. The range expresses uncertainty; it does not guarantee coverage.
Pipeline
From market components to one scoring identity.
Fill uncovered scoring events
Use recent game form for fantasy-scoring components not represented in the current market inputs. At least five recent games are required to form a range.
Convert and rank
Apply DraftKings MLB hitter scoring to the available components, produce the central estimate, then rank eligible hitters on a common scale.
Range construction
Floor and ceiling are scenarios, not walls.
Recent variation sets the distance
The method examines the recent fantasy-point distribution around its median. The distance to a lower P20-style outcome and an upper P80-style outcome is applied around today's central projection.
The interval is deliberately wider
A fixed 1.25 widening multiplier is applied to the recent-variation distances. This is intended to reduce false precision, but it is not a guarantee of nominal coverage.
“P20/P80-style” describes how the historical distances are constructed. It does not mean exactly 60% of future scores will fall inside every published interval.
Reading the range
Why isn't the projection halfway between floor and ceiling?
Propeller's fantasy projection is the central estimate supplied by its method; it is not calculated by averaging the displayed floor and ceiling. The lower and upper ranges can extend different distances from that estimate. Read all three numbers together, and do not replace the projection with their midpoint or treat either endpoint as a guaranteed score.
In an AI-operated inspection of Propeller's public fantasy response on September 12, 2026, a Jackson Merrill row listed a 7.1 floor, 11.5 projection and 32.1 ceiling in DraftKings MLB fantasy points. The response was labeled stale. We use that dated output only to examine how the three numbers relate, not as a current player recommendation.
- Floor
- 7.1
- Supplied projection
- 11.5
- Ceiling
- 32.1
| Calculation | Arithmetic | Meaning |
|---|---|---|
| Distance below estimate | 11.5 − 7.1 = 4.4 | Lower side of this published range |
| Distance above estimate | 32.1 − 11.5 = 20.6 | Upper side of this published range |
| Endpoint midpoint | (7.1 + 32.1) / 2 = 19.6 | A calculated midpoint, not the supplied 11.5 projection |
The midpoint is 19.6; the supplied projection is 11.5. We calculated the midpoint only to show the middle of the displayed interval. It is not a projection field or a forecast.
The published method combines available market components with recent form to produce its central estimate. It applies separate recent-variation distances around that estimate, so the two sides can differ. This arithmetic is consistent with that description; it does not reproduce the internal model or establish how often future scores will fall within the range.
Keep the supplied estimate and inspect both endpoints on the same scoring basis. A larger upper distance neither erases the lower scenario nor proves a more likely high score. Actual scores can fall outside the range. Inspect the current MLB fantasy board for current research, checking its freshness label before using a row.
Historical source: public fantasy response endpoint, inspected September 12, 2026. The endpoint may now return different data; this case preserves the dated stale response.
Scoring
One label must mean one formula.
| MLB hitter event | DraftKings points |
|---|---|
| Single | 3 |
| Double | 5 |
| Triple | 8 |
| Home run | 10 |
| Run batted in / run scored | 2 each |
| Walk or hit by pitch | 2 |
| Stolen base | 5 |
The feature labels this identity “DraftKings MLB.” Confirm current contest scoring on the platform where you play; platform rules can change.
Freshness
Current, latest available, or unavailable.
Current
The source has eligible projections and its update timestamp is no more than 60 minutes old.
Latest available
Eligible rows exist, but the source update is older than 60 minutes. The page labels that state rather than calling it live.
Unavailable
No eligible rows or no readable source response. The page retains only explicitly labelled examples and does not relabel them as today's projections.
Limitations
What v1 does not know.
Scope limitations
Public v1 covers MLB hitters only. It does not publish pitcher or non-MLB fantasy projections and does not model a user's salary cap, league settings, roster, or exposure.
Input limitations
Market components can be missing, stale, or revised. Recent form may not represent a new role, injury, lineup spot, weather change, or unusual matchup.
Outcome limitations
Fantasy scoring is volatile. Actual scores may fall below the floor or above the ceiling. A higher ceiling is not automatically a better decision.
Evidence limitations
No public accuracy or calibration claim is made in v1. Such claims require a reproducible, time-ordered evaluation with documented exclusions and scoring sources.
Questions
Method details in plain language.
How does Propeller project MLB fantasy points?
Propeller converts available current hitter-stat market components into DraftKings MLB fantasy points and uses recent form for scoring components without current market coverage. It then estimates a lower and upper range from recent outcome variation.
Are floor and ceiling guaranteed percentiles?
No. They are P20/P80-style range estimates, not guaranteed bounds or a claim that exactly 60% of future outcomes will land inside the interval.
Should I average the floor and ceiling to get the projection?
No. Use the supplied projection. Averaging the endpoints gives their midpoint, which can differ from the method's central estimate when the lower and upper distances are unequal.
Does Propeller publish fantasy projection accuracy?
Not yet on this page. Propeller will publish calibration or accuracy claims only after the evaluation cohort, timing, exclusions, scoring source, and reproducible calculation are documented for public review.
Can I reproduce the current feature exactly?
The public method and scoring identity are documented, but raw market inventory and all internal processing are not exposed. The machine-readable method file identifies the current version and public inputs, outputs, and limitations.
For a shorter explanation, read how Propeller fantasy projections work. Explore the fantasy feature overview for scope and available research tools.