NBA · Research roadmap
Research roadmap, not a current capability
Basketball has the daily volume that makes baseball attractive, but its markets are dominated by late roster news in a way that rewards speed over analysis. Nothing described here is running.
Current status
Nothing here is running
- Data collection
- None
- Model
- None
- Published record
- None
- Timeline
- Not committed
Roadmap means roadmap
VigPulse does not run NBA. There is no model, no data pipeline, no graded entry, and no committed date. The only sport currently covered is MLB, that coverage is moneyline only, and it is paper-only.
The record
What a published record would have to contain
| Field | MLB | NBA |
|---|---|---|
| Graded entries | 33 | 0 |
| Record | 9W-24L | None |
| Paper units | +2,417.5u | None |
| Pre-game CLV sample | 6 | 0 |
| Stake type | Paper only | Nothing staked |
MLB figures are read from the published ledger. NBA columns are empty because no basketball entry has ever been logged, not because the data is withheld.
Market structure
How NBA differs
Late news moves everything
A star ruled out shortly before tip-off can move a spread by several points. Because that information arrives close to the event and irregularly, the advantage goes to whoever reacts fastest rather than to whoever models best.
Rest and scheduling are quantifiable
Back-to-backs, road trips, and rest differentials have measurable effects that books price explicitly. This is one of the few areas where a public model can be built on well-documented inputs.
Load management is a modelling problem in itself
Teams rest healthy players for reasons that are partly strategic and partly opaque. Predicting availability is a separate problem from predicting performance, and in basketball it may be the larger one.
High scoring compresses variance
More possessions per game means outcomes track underlying quality more closely than in baseball or football. That makes the sport more predictable, which also means the market is more likely to have already priced it.
Method
What would have to change in the method
The sequence on the methodology page assumes information arrives on a schedule, as it does in baseball. Below is what each step would have to become when the dominant input lands minutes before the event instead. It is a description of unsolved work.
Six steps, rewritten for late news
not implementedMarket
Basketball is priced as a spread and a total, and both react to the same input: who is playing. Capturing the price without capturing the roster state that produced it records the answer and discards the question.
Consensus
De-vigging works the same way it does in baseball, but the consensus it produces is unstable in the final hour before tip-off. A number that is correct at 18:00 can be stale by 18:40 for reasons that have nothing to do with mispricing.
Movement
The core problem is attribution. A three-point move caused by a star being ruled out is not the same event as a three-point move with no news behind it, and a model that treats them alike will read the first as a signal when it is only a correction.
Model
Two models are needed, not one. The first predicts availability, which is partly a strategic decision made by a coaching staff. The second predicts performance given availability. In basketball the first may be the harder of the two.
Opportunity
An edge with a half-life measured in minutes is not the same object as an edge that persists for hours. Speed of reaction becomes the binding constraint, and that is a different competition from the one this project is set up to enter.
Risk
Books restrict quickly on markets where late news drives the price. Any assessment has to include whether a position could be taken at a useful size at all, not only whether it would have been correct.
The speed problem
StructuralBaseball posts its biggest information event, the lineup, at a known time before first pitch. Basketball does not. A late scratch can move a spread by several points, and the advantage goes to whoever reacts first rather than to whoever models best.
That is a latency competition against participants with better infrastructure. It is worth naming as the reason this sport is unattractive for a small experiment, rather than discovering it after building.
The predictability trap
StructuralHigh possession counts make basketball outcomes track team quality more closely than baseball outcomes do. That sounds like an advantage, and it mostly is not: a sport that is easier to predict is a sport the market has already priced more accurately.
Difficulties
What makes basketball hard
Known problems
3 listed- Reaction speed to injury news matters more than model quality, which is a different competition than the one this project is set up for.
- Availability prediction is a distinct and difficult problem layered on top of game prediction.
- Books limit or restrict quickly on markets where late news is the main driver.
What building this would require
3 listed- A reliable, fast source of roster and availability information.
- Movement analysis that distinguishes news-driven moves from ordinary repricing.
- Realistic assessment of whether reacting fast enough is achievable at all.
Readiness
Gates before NBA would be anything but a research note
| Gate | Status | Notes |
|---|---|---|
| A fast, timestamped source of roster and availability information | Not started | None exists here. Without it, the sport cannot be modelled at all, let alone traded. |
| Movement analysis that separates news-driven moves from drift | Not started | Not started. This is the substantive research problem for basketball. |
| An availability model distinct from the performance model | Not started | Not started. Predicting who plays is its own forecasting task. |
| An honest assessment of whether reacting fast enough is achievable | Not started | Not started, and the answer may be no. That is a legitimate outcome of the research. |
| Publish a graded record before describing NBA as a capability | Not started | Not started. No NBA entry has ever been logged or graded. |
Comparison
Where basketball sits against the rest
| Sport | Status | Cadence | Primary market | Liquidity |
|---|---|---|---|---|
| MLB | Current · paper | Daily, roughly 15 games | Moneyline | High on major markets |
| NFL | Roadmap | Weekly, roughly 16 games | Point spread | Highest of any US sport |
| NBA | Roadmap | Daily, roughly 8 games | Point spread and total | High |
| WNBA | Roadmap | Seasonal, several games per week | Point spread and total | Lower than NBA |
| Tennis | Roadmap | Near-continuous, year-round | Match winner | Varies enormously by event tier |