GARP FRM Part I Quantitative Analysis Study Guide

Table of Contents

• • Where Quantitative Analysis Fits in FRM Part I
• • What Is the Official FRM Part I Exam Format?
• • How the 15 Chapters Build Quantitative Skill
• • Formula Learning Without Blind Memorisation
• • A Six-Stage Quantitative Analysis Study Plan
• • Common Quantitative Analysis Exam Traps
• • When Are You Ready to Practise Across All Four Modules?

Quick answer: FRM Part I has 100 equally weighted multiple-choice questions in four hours across four printed-book domains. Quantitative Analysis is Book 2, included with the other three books as a module inside the single Exams Academy FRM Part I course.

The GARP FRM Part I Quantitative Analysis book supplies the probability, statistical and predictive tools used across risk management. Its path is deliberate: begin with events and random variables, build estimation and hypothesis testing, extend those ideas into regression and time series, and finish with return measurement, simulation and machine learning.

The challenge is not only remembering formulas. Candidates need to recognize the assumptions that make a formula valid, distinguish similar concepts, and interpret a numerical result in risk language. The official GARP FRM study-material structure places Quantitative Analysis alongside three other Part I books. Keep the book boundary clear while studying, then integrate the tools across the qualification.

Where Quantitative Analysis Fits in FRM Part I

Part I is organized into four printed-book domains:

1. Foundations of Risk Management
2. Quantitative Analysis
3. Financial Markets and Products
4. Valuation and Risk Models

The combined Exams Academy course preserves this structure as four clearly labelled modules. A candidate can see which book and chapter has been completed, diagnose weaknesses at chapter level and then move into integrated Part I practice without purchasing or switching between separate courses.

Quantitative Analysis is also a bridge. Probability and distributions support risk models. Regression estimates factor exposure and hedge relationships. Time-series methods support volatility and return forecasts. Simulation helps when analytical solutions are unavailable. Machine-learning methods add flexible prediction while reinforcing the need for clean validation and out-of-sample testing.

What Is the Official FRM Part I Exam Format?

The current GARP program and examinations page states:

Item	Official Part I detail
Questions	100
Question type	Multiple choice
Weighting	Equally weighted questions
Duration	Four hours
Curriculum scope	All four Part I book domains

That gives an average of 2.4 minutes per question. Use that average to build pacing discipline in integrated practice across all four course modules.

GARP does not publish a fixed candidate pass percentage on the program page. A platform may use 70% as an internal mastery signal, but that figure should guide revision rather than be presented as GARP’s pass mark or as a prediction of the final result.

How the 15 Chapters Build Quantitative Skill

The first four chapters establish probability models. Fundamentals of Probability introduces event spaces, conditional probability, independence and Bayes’ rule. Random Variables converts outcomes into numerical objects described by PMFs, PDFs and CDFs. Common Univariate Random Variables covers the named distributions used in risk work, including Bernoulli, binomial, Poisson, normal, lognormal, Student’s t, chi-squared and F. Multivariate Random Variables then adds joint distributions, covariance, correlation and conditional expectation.

Chapters 5 and 6 move from populations to evidence. Sample Moments explains estimators, bias, efficiency, consistency, the law of large numbers and the central limit theorem. Hypothesis Testing turns sampling uncertainty into null and alternative hypotheses, rejection regions, p-values, confidence intervals, Type I errors, Type II errors and power. A candidate should always state what is assumed under the null and why the selected test statistic has its reference distribution.

Chapters 7 to 9 develop regression. Simple Linear Regression explains OLS coefficients, assumptions, standard errors and R-squared. Multiple regression changes a coefficient into a partial effect while introducing joint tests and adjusted R-squared. Regression Diagnostics then asks whether the model is trustworthy by examining omitted variables, heteroskedasticity, multicollinearity, residual patterns, outliers and influential observations.

Chapters 10 and 11 separate stationary from non-stationary time series. A stationary model has stable moments and dependence across time. AR, MA and ARMA models describe that dependence and support forecasts. A unit-root process behaves differently: shocks can be permanent, variance grows with forecast horizon and ordinary regressions can become spurious. Detrending and differencing solve different problems, so identify the trend type before transforming the data.

The final four chapters connect measurement and modern prediction. Measuring Returns, Volatility, and Correlation compares simple and log returns, time scaling, implied volatility, non-normal tails and dependence measures. Simulation and Bootstrapping distinguishes drawing from an assumed data-generating process from resampling observed data. Machine-Learning Methods introduces data preparation, PCA, clustering, reinforcement learning and NLP. Machine Learning and Prediction then covers regularization, logistic regression, trees, ensembles, nearest neighbours, support-vector machines, neural networks and classification metrics.

Formula Learning Without Blind Memorisation

Use a four-part note for every formula:

1. Definition: write the relationship and define each symbol.
2. Assumptions: state what must be true for the calculation or inference to hold.
3. Interpretation: explain the output in words and units.
4. Failure mode: record the common way the relationship is misused.

For example, the variance of a weighted two-variable sum includes two squared-weight variance terms and a covariance term. The covariance term disappears only when covariance is zero. The standard error of an iid sample mean falls with the square root of sample size, so precision improves slowly. The square-root-of-time volatility rule also needs stable, independent return behavior; it should not be applied mechanically to autocorrelated or changing-volatility data.

This method is especially useful for regression. A coefficient has a calculation, but it also has a conditional interpretation. Its standard error depends on assumptions. R-squared describes sample fit but not causality. A robust standard error can repair inference under heteroskedasticity without changing the OLS coefficient estimate.

A Six-Stage Quantitative Analysis Study Plan

Stage 1 — Secure probability notation. Translate and, or, not and given into intersection, union, complement and conditioning. Practise Bayes’ rule by drawing a partition of mutually exclusive states before substituting numbers.

Stage 2 — Match distributions to data. Learn each distribution’s support, generating process, parameters and financial use. Binary events suggest Bernoulli, repeated independent successes suggest binomial, constant-rate counts suggest Poisson and positive right-skewed levels may suggest lognormal.

Stage 3 — Connect population and sample. Separate a population parameter, an estimator and a realized estimate. Compare unbiasedness, efficiency, mean squared error and consistency. Explain why standard error falls with the square root of sample size.

Stage 4 — Build regression in layers. Master simple OLS before adding regressors. Then study diagnostic failures one at a time: omitted variables, heteroskedasticity, multicollinearity and influential observations. For every issue, record whether it affects coefficients, standard errors, efficiency or interpretation.

Stage 5 — Separate time-series regimes. Decide whether a series is stationary, deterministically trending, seasonal or unit-root driven. Only then select ARMA modeling, trend regression, seasonal controls or differencing.

Stage 6 — Validate simulation and machine learning. Distinguish Monte Carlo sampling error from model error. Keep training, validation and test data separate. Choose classification metrics according to the cost of false positives and false negatives rather than relying on accuracy alone.

Common Quantitative Analysis Exam Traps

• Mutual exclusivity versus independence: positive-probability mutually exclusive events are dependent because one rules out the other.
• PDF versus probability: a continuous density at one point is not the probability of that exact outcome.
• Variance scaling: multiplying a variable by b multiplies variance by b squared, not b.
• Zero correlation versus independence: zero covariance does not generally eliminate nonlinear dependence.
• Failing to reject: insufficient evidence against a null does not prove that the null is true.
• R-squared interpretation: a high R-squared does not establish causality, correct specification or future accuracy.
• Heteroskedasticity response: robust standard errors change inference, not the OLS point estimates.
• ADF direction: the unit root is the null; rejection requires a sufficiently unfavorable statistic using the correct non-standard critical value.
• Simulation precision: ten times smaller Monte Carlo error requires roughly one hundred times as many replications.
• Validation leakage: tuning on the test set makes the final performance estimate optimistic.
• Precision versus recall: the measures use different denominators and reflect different error costs.

When Are You Ready to Practise Across All Four Modules?

You are ready to move beyond book-level practice when you can choose a distribution from its data-generating story, transform moments correctly, explain estimator properties, construct and interpret a test, read regression diagnostics, distinguish stationary and unit-root behavior, scale returns under stated assumptions, design a reproducible simulation and evaluate a classifier with an appropriate metric.

Quantitative Analysis is not the complete FRM Part I curriculum. It is Module 2 of the complete course and the toolkit that makes the other domains easier to analyze. After securing this module, use the same dashboard to continue through the remaining books and begin integrated practice across all four domains.