Using Novus Odds

Breakdowns guide

Every lab results panel — math advantage, odds buckets, strategy overlays, hot streaks, consistent bands, horse-field frequencies, casino house-edge tables, and more — is explained in the Breakdowns guide, with practical educational use cases.

• Open /help/breakdowns for the full guide (also linked from Learn, Experiments, and each lab header).
• Use the Help control on any breakdown panel to jump to that section’s anchor.
• The guide is educational only: synthetic data, no live odds, not betting advice.

Choosing a lab

• Odds Lab — synthetic two-outcome markets, filters, buckets, and A/B compare.
• Strategy Lab — stake rules and selection heuristics on synthetic tapes.
• Sports Lab — fictional matchups or multi-runner fields.
• Props Lab — fictional player distributions and line sweeps.
• Casino Lab — house-edge models for roulette, blackjack, baccarat, craps, and slots.
• Datasets — generate and export fictional multi-season universes.

Running an Odds Lab experiment

• Choose the number of events. Larger experiments reduce sampling noise but require more browser processing.
• Select the minimum and maximum American odds included in the generated market.
• Set the simplified market overround used to derive the simulation's baseline fair probability.
• Set pricing uncertainty to control variation between the market estimate and simulated true probability.
• Optionally restrict selection to favorites, underdogs, odds thresholds, or mispriced markets.
• Enable A/B compare to contrast two selection rules on the same synthetic tape.
• Enter a random seed. Reusing the same settings and seed reproduces the experiment.
• Select Run Experiment and review the summary, charts, example tickets, and odds bucket table.

Why did two experiments produce different results?

Changing the random seed creates a different sequence of generated markets and outcomes. Even when the underlying assumptions are identical, random variance means observed profit, streaks, and win rates may differ.

Use the same seed when you need a reproducible baseline. Change the seed when you want to explore how the same assumptions behave across different simulated histories.

Where are my results stored?

Recent experiment results are stored locally in your browser. They are not synchronized between devices and are not currently associated with an online user account.

Performance

The current experiment engine supports up to one million events per run. Very large experiments perform substantial computation in the browser and may take longer on lower-powered devices.

Reporting a problem

For product or technical issues, contact Novus Stream Solutions through the main Novus Stream Solutions website. Include the experiment settings, browser, device type, and a description of the unexpected behaviour when possible.