Hypothesis
A hypothesis is a provisional proposition, explanation, or prediction that can be evaluated against evidence.
Cyber Threat Hunting
A hypothesis is a proposition consistent with known data that has not yet been verified or shown to be false.
- How it is used
- Drives structured threat hunting by defining a suspected threat and directing evidence searches.
- How deep it goes
- Defines hypotheses, gives relevance and testability criteria, and details an iterative testing and response workflow.
- Examples
- Uses PowerShell activity, spearphishing through a malicious Word document, Kubernetes container escape, and public-facing API exploitation.
- What you can do after
- Apply it in structured threat hunting to scope searches, assess evidence, refine inconclusive investigations, and escalate confirmed incidents.
Grokking Statistics
A hypothesis is a proposed explanation or relationship that can be evaluated with evidence.
- How it is used
- Frames data investigations by turning a proposed relationship or claim into something evidence can evaluate.
- How deep it goes
- Defines hypotheses through workflows and examples, then develops null and alternative claims for formal testing without deriving test formulas.
- Examples
- Uses temperature and sports drink sales, 305 mL soda cans, and toddler meltdowns after missed naps.
- What you can do after
- Evaluate/understand proposed claims by selecting variables, gathering data, testing models on new data, and qualifying conclusions.
Effective Data Analysis
A hypothesis is a testable proposition about a population, relationship, or data-generating process.
- How it is used
- Guides analytics questions, research design, method selection, and evaluation across experiments, metric comparisons, and predictive modeling.
- How deep it goes
- Explains evidence gathering, construction, null and alternative forms, direction, and quantification, while deferring statistical tests to later treatment.
- Examples
- Uses tutoring comparisons, e-commerce layout experiments, A/B tests, and weather-based rat-sighting prediction to illustrate hypotheses.
- What you can do after
- Evaluate questions by structuring testable propositions, documenting assumptions, selecting methods, and defining evidence-based decision criteria.
Statistics Slam Dunk
A hypothesis is a stated proposition about a population, process, or relationship that can be evaluated with evidence.
- How it is used
- Frames NBA questions as testable starting points for data preparation, statistical analysis, visualization, and interpretation.
- How deep it goes
- Explains null hypotheses and assumptions, then demonstrates decisions through tests, summaries, charts, and comparisons rather than formal proofs.
- Examples
- Uses draft segments and career games, officiating bias, fourth-quarter performance, and standardized scoring as concrete analytical questions.
- What you can do after
- Evaluate/understand NBA propositions by preparing data, comparing summaries, visualizing results, and interpreting statistical test outcomes.
The Art of AI Product Development
A hypothesis is a provisional, testable proposition about a product change, user behavior, or expected outcome.
- How it is used
- Uses provisional propositions to discover and validate AI product opportunities through rapid experiments.
- How deep it goes
- Explains formulation through small product changes and validation, but does not specify experiment metrics or statistical methods.
- Examples
- Uses prompt changes, local layout changes, prompt suggestions, and chatbot experimentation as concrete hypothesis tests.
- What you can do after
- Apply via experiments that implement, deploy, and test incremental product changes without requiring complete prior data.
Read how The Art of AI Product Development explains hypothesis →
The Well-Grounded Data Analyst
A hypothesis is a specific, testable proposition about a relationship or difference in a domain.
- How it is used
- Focuses survey analysis on evidence that supports or contradicts stakeholder theories and informs a potential AI-product market gap.
- How deep it goes
- Moves from proposition to variable selection, exploratory comparison, and possible statistical testing, but does not specify the test procedure.
- Examples
- Uses developer experience, generative AI usage, usefulness, favorability, and trust responses from the Stack Overflow Developer Survey.
- What you can do after
- Evaluate whether a proposed relationship is supported by survey evidence and decide which analytical paths serve the stakeholder question.
Read how The Well-Grounded Data Analyst explains hypothesis →
Sutskever's List
A hypothesis is a proposed explanation or generative process whose plausibility is evaluated against observed data.
- How it is used
- Frames inductive inference as evaluating possible data-generating processes, grounding a theoretical account of intelligence.
- How deep it goes
- Defines compatibility and simplicity through intuitive examples, then sets up consideration of computable programs without detailing the weighting.
- Examples
- Uses a cookie-jar mystery and the sequence 1, 3, 5, 7 to contrast compatible explanations and simpler predictions.
- What you can do after
- Evaluate competing explanations by retaining those consistent with evidence and preferring simpler ones.
Causal Inference for Data Science
A testable proposition about a causal relationship between an intervention and an outcome.
- How it is used
- Frames A/B tests and randomized experiments around a causal question about an intervention and an outcome.
- How deep it goes
- Defines the experimental role of a hypothesis and sketches design, execution, and analysis while deferring detailed explanation.
- Examples
- Uses a new-versus-old disease treatment comparison and binary patient improvement outcomes.
- What you can do after
- Apply it via an A/B test by stating a causal proposition, designing an experiment, protecting randomization, and comparing responses.
Read how Causal Inference for Data Science explains hypothesis →
Machine Learning System Design
A hypothesis is a testable proposition about an expected relationship or effect in an experiment.
- How it is used
- Frames expected effects before offline evaluation and A/B testing.
- How deep it goes
- Defines components and planning parameters, but defers detailed sample-size mathematics.
- Examples
- Uses retail stockouts and sales prediction, plus offline estimates preceding online experiments.
- What you can do after
- Apply it via an A/B test by specifying metrics, uplift, error thresholds, duration, and reporting expectations.
Read how Machine Learning System Design explains hypothesis →
Related concepts
- alternative hypothesisincludes · An alternative hypothesis is a statistical proposition that contrasts with the null hypothesis by asserting an effect, difference, relationship, or other departure.
- null hypothesisincludes · A null hypothesis is a designated baseline statistical claim, typically asserting that an effect, difference, or association is absent.
- threat huntingused by · Threat hunting is the proactive investigation of systems and networks to discover malicious activity that automated detection may have missed.
- threat-hunting playpart of · A repeatable operational procedure for conducting a particular threat hunt and recording its investigative scope and logic.
- red-teaminguses · Red-teaming is an adversarial assessment practice that deliberately simulates attacks or uses challenging inputs to identify a system’s vulnerabilities, failures, or unsafe behavior.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.