What you will learn
Discrete events, conditional probability, Bernoulli, binomial and Poisson models. This lesson builds transferable statistical literacy for evaluating sports research, not a promise of profitable wagering.
Why this matters
A sports prediction is an uncertain statement about a precisely defined event. Before calculating anything, identify the outcome, unit of analysis, timeframe, population, and source timestamp. Odds and model outputs may look precise while depending on weak assumptions. Students should separate mathematical identities from empirical estimates and distinguish theoretical value from an executable quote.
Core concepts
Discrete events, conditional probability, Bernoulli, binomial and Poisson models. Start with the event definition, then write the model assumptions in ordinary language. For each estimated input, ask whether the sample was representative, whether the observation was available before the event, and how much results change if the input is wrong.
Worked classroom calculation
A shooter makes each of 5 attempts independently with probability 0.4. Probability of exactly 2 makes = C(5,2)×0.4²×0.6³ = 0.3456.
The numerical inputs here are hypothetical and chosen for instruction. Show each intermediate operation, keep probability values between zero and one, and label any implied probability as a price-derived quantity rather than an objectively known chance.
Interpreting the result
Why independence is an assumption; binomial variance np(1-p)=1.2; count data may be overdispersed.
A calculated number is only as defensible as its assumptions. A simple baseline can be more useful than a complex model that was tuned repeatedly against the same results. When comparing methods, use held-out observations and predeclared metrics. When evaluating an offered price, account for fees, liquidity, accepted stake, timing, and settlement rules.
Practical workflow
- Define the event and information cutoff.
- Document the inputs and whether each is observed or estimated.
- Calculate the baseline by hand.
- Run at least two sensitivity scenarios.
- Compare with a naive benchmark on future or held-out data.
- Record failures and revisit assumptions rather than retroactively rewriting them.
Common errors
- Treating independence as automatic when observations share game state or common causes.
- Using information that was unavailable at forecast time.
- Confusing association with causation.
- Equating a model's fair price with a certain true probability.
- Ignoring variance, multiple testing, or the uncertainty of the model itself.
- Presenting an educational simulation as a real, available trading opportunity.
Practice
Exercise A: Reproduce the classroom calculation with every intermediate step.
Exercise B: Compute probability of zero makes: 0.6^5 = 0.07776.
Exercise C: Identify a violated assumption that could materially change the result and propose a sensitivity test.
Exercise D: Write a research log entry with event, as-of time, model version, data sources, and intended evaluation metric.
Answer key
A: The worked example supplies the correct result under its stated assumptions; students should reproduce it, not merely quote it.
B: Compute probability of zero makes: 0.6^5 = 0.07776.
C: Strong answers name a specific assumption, such as independence, representativeness, stable minutes, or accurate probability estimation, and describe an alternative scenario.
D: A complete log preserves the original forecast and prevents hindsight edits.
Sharp Academy implementation
Publish this as a sequential written tutorial with an interactive calculator, a downloadable worksheet, and an ARBAI follow-up prompt for clarification. Avoid using live prices unless timestamp, source, market definition, and executable status are verified. Review mathematics, legal/jurisdictional issues, and educational claims before release.
Try it in Sharp App: Calculators
Questions? Ask ARBAI about this lesson · Ask our Sharps in Discord →
Sharp App surfaces +EV opportunities from public sportsbook data. We do not place, accept, or facilitate wagers. 21+ where legal. If betting stops being fun, call 1-800-GAMBLER or visit ncpgambling.org.
