FAQ / How It Works

Guidance for using Hoop Props and interpreting the slate without statistical jargon.

Back to Today's Slate

What Hoop Props Is (and Is Not)

Hoop Props is an NBA player prop analytics tool that turns posted lines into estimated probabilities and highlights where model estimates differ from those lines.

  • It helps you spot when a prop looks meaningfully different from the model (and when to pass).
  • It does not guarantee outcomes, and it is not financial advice.
  • It is not betting advice and it is not a replacement for responsible decision-making.

How Probabilities Are Generated (High Level)

Hoop Props estimates the chance that a player goes over or under a line by looking at how a player has performed in similar situations and combining multiple independent model signals into a single estimate.

The exact math is intentionally not shown, but you can think of it like this: multiple “opinions” (recent form, longer-term history, distribution-based estimates, simulations) are blended to produce a stable probability estimate.

How Calibration Works on /picks

Calibration is a post-processing step. It does not create picks by itself. Instead, it checks whether past model probabilities matched reality over time. If the model has been systematically too aggressive or too conservative, calibration remaps those probabilities closer to observed hit rates.

On /picks, the default view uses the latest stored base model probabilities from the pipeline. Choosing a calibration run from the dropdown does not fit anything live. It simply swaps in a saved historical calibration run for the current slate when one exists.

  • Use the default view when you want the current model output exactly as generated.
  • Use a calibration run when you want those probabilities remapped using resolved historical results.
  • If a selected run has no calibrated value for a pick, the app falls back to the base model probability.

When To Recalibrate

Recalibration is not something you need to do every page refresh or every pipeline run. A better cadence is after a meaningful batch of picks has fully settled.

  • As a practical default, recalibrate about once per week during the season, or after a solid block of new settled results.
  • Recalibrate immediately after material model changes, feature changes, source changes, or line-mix changes.
  • Do not replace the current run just because a newer one exists. Keep the new run only if it improves reliability on holdout data.

In plain terms: recalibration should follow new evidence, not new page views.

How to Read the Props Table

Key columns

  • Decision: Which side (Over or Under) the model leans toward.
  • Recommendation: Whether the model flags the entry as higher signal or “No Bet” when the signal is weak.
  • Win Probability: The estimated chance that this outcome (Over or Under) occurs, based on all models combined.
  • Edge (%): The difference between the model probability and the implied probability from the line.
  • Confidence Score: A reliability score based on model agreement, data volume, and recent player performance.
  • Simulation Prob: Win probability estimated using Monte Carlo simulations of player outcomes.
  • Model Estimate: Average win probability across analytical models (historical, Bayesian, KDE, etc.).

A simple workflow

  1. Start at the top of the slate (highest-ranked entries first).
  2. Focus on entries with a clear signal (not “No Bet”).
  3. Use Edge (%) to compare differences across props, not to force action.
  4. If the slate is weak, passing is a valid decision.

Use the Tool Responsibly

  • Models can be wrong—injuries, rotations, pace, role changes, etc.
  • A “high probability” projection can still be wrong. Variance is real.
  • Use Hoop Props as one input in a disciplined process, not a promise of results.

Common Misconceptions

  • “The model says 70%, so it should be correct.” A 70% estimate is expected to be wrong about 30% of the time over the long run.
  • “More entries means better accuracy.” More volume can mean more risk. Quality matters.
  • “If it missed yesterday, it’s ‘due’.” Past misses do not create future wins.

Disclaimer

Hoop Props is for informational and entertainment purposes only. It is not betting advice.

Hoop Props provides estimates based on historical data and model assumptions. No projection is guaranteed. Sports outcomes are uncertain and any wagering involves risk, including the risk of losing your entire stake.

By using Hoop Props, you acknowledge that you are responsible for your own decisions and results.