Aligned to the 2026 GARP® FRM Curriculum
Free FRM® Practice Questions
FRM Part I and Part II. Original proprietary questions, detailed solutions, lessons that mirror how risk is actually tested. Core prep is free forever — no credit card, no trial period.
JT
Built by Jeffrey Ting, FSA, CFA
Fellow of the Society of Actuaries and CFA charterholder. Risk management is the natural intersection of the actuarial and investment-analyst disciplines — the FRM curriculum is where I spent years on the practice side. I built FreeFellow after paying four-figure prep fees for my own exams and deciding candidates shouldn’t have to. Every question is authored or vetted against the learning outcomes of the current GARP curriculum.
Pick Your Part
FRM Part I
Foundations & Risk Models
100 questions · 4-hour exam
FRM Part II
Applied Risk Measurement
80 questions · 4-hour exam
What’s Free — and What Fellow Unlocks
The core FRM prep — questions, solutions, and lessons — is free forever. Deeper practice tools (flashcards, timed mocks, analytics) unlock when you become a Fellow. Use the free tier as long as you like; upgrade only if you want the extras.
Free forever
$0 · no credit card
- Practice questions across both FRM parts
- Detailed solutions on every question
- Lessons with audio narration on core topics
- Mixed practice sessions across all topics
- Estimated pass likelihood (gated at 50+ Qs)
- Readiness benchmarks vs passing thresholds
Fellow unlock
$59 / qtr · $149 / yr per part
- Flashcards (SM-2 spaced repetition)
- Timed mock exams with weighted scoring
- Topic-level practice to drill weak areas
- Performance analytics · topic mastery breakdown
- Personalized study plan that adapts to your exam date
How FreeFellow Compares
Per-part pricing across major FRM prep providers. GARP exam fees are separate (USD 400 enrollment plus USD 600 early or USD 800 standard per part).
FreeFellow core free
$0 · Fellow $59/qtr
Bionic Turtle
$400–$600 / year
AnalystPrep
$199–$399 / part
Kaplan Schweser
$500–$700 / part
All four prep providers are independent of GARP. The paid providers bundle prep into a single up-front fee; FreeFellow keeps the question bank and lessons free and charges only for deeper study tools. GARP exam fees ($400 enrollment + $600–$800 per part) are paid directly to GARP and are not included in any prep provider’s price.
Sample Questions
Part I · Foundations of Risk Management
Easy
A pension fund holds two equity managers in 2026. Manager A returned 14% with 22% volatility; Manager B returned 11% with 12% volatility. The risk-free rate over the period was 5%. Which manager delivered the higher Sharpe ratio?
Solution
The Sharpe ratio is excess return divided by total risk: (Rp − Rf) / σp. Manager A: (14% − 5%) / 22% = 0.41. Manager B: (11% − 5%) / 12% = 0.50. Manager B delivers more excess return per unit of total volatility, so B has the higher Sharpe ratio. Choice A is the headline-return trap — absolute return ignores risk. Choice C inverts the ratio (Sharpe falls, not rises, with volatility). Choice D is unsupported by the arithmetic. The same question family on Part I will sometimes ask for Treynor ratio (uses β instead of σ) or Information ratio (uses tracking error) — the choice of risk measure in the denominator is what the question is testing.
Part I · Valuation and Risk Models
Medium
A trading desk holds a portfolio with daily returns that are approximately normally distributed with mean zero and daily volatility of $2 million. Using the parametric (delta-normal) approach, what is the one-day 99% Value-at-Risk (VaR)?
Solution
For a normal distribution, parametric VaR at confidence level 1 − α is zα × σ, where zα is the standard normal quantile and σ is the portfolio’s standard deviation. For 99% VaR, the one-tailed quantile is z0.01 = 2.326. So VaR = 2.326 × $2 million ≈ $4.65 million. Choice A uses 1.645 (the 95% quantile, not 99%). Choice C uses 2.58 (the two-tailed 99% quantile, which is wrong for one-tail VaR). Choice D drops the multiplier entirely. The exam tests two pieces here: that 99% one-tail uses 2.326 (not 2.58), and that VaR scales linearly with σ under the delta-normal assumption. For multi-day horizons, scale σ by √t under the i.i.d. assumption.
Part II · Market Risk Measurement
Hard
A risk manager uses 250 trading days to backtest a one-day 99% VaR model. Over the period, the model produced 7 exceptions (days when actual loss exceeded VaR). Using the Kupiec proportion-of-failures (POF) test framework at the 5% significance level, the most likely conclusion is:
Solution
For a 99% VaR model over 250 days, the expected number of exceptions is 250 × 0.01 = 2.5. Observing 7 exceptions is roughly 2.8× the expected count, suggesting the model under-estimates losses. Kupiec’s proportion-of-failures (POF) likelihood ratio test computes LRPOF = −2 ln[(1 − p)T−N pN / (1 − N/T)T−N (N/T)N], which is χ2(1)-distributed under the null. Plugging T = 250, N = 7, p = 0.01 yields LR ≈ 6.5, well above the 3.84 critical value at 5% significance. The null of correct calibration is rejected. Choice A is wrong — expected is 2.5, not 7. Choice B inverts the direction (too few exceptions would suggest conservative; too many suggests under-estimation). Choice D is the “throw up your hands” trap; Kupiec is specifically designed for samples this size and Basel uses a 250-day window in its traffic-light backtesting framework. The exam tests both the arithmetic (expected exceptions = T × p) and recognition of which test applies.