Community Q&A
Expert-verified answers to your financial certification questions. Ask, learn, and connect with fellow candidates.
Updated
What makes the Weibull distribution useful for modeling time-to-failure?
Weibull generalizes exponential with shape beta controlling whether hazard rises or falls over time. beta greater than 1 means wear-out, less than 1 means infant mortality.
How do I read a confusion matrix for a financial classification model?
Confusion matrix for Quellmont Bank: 143 TP, 53 FN, 411 FP, 7593 TN yields 94.3% accuracy, 25.8% precision, 73% recall. Economic analysis shows $199M net benefit...
How do Sharpe ratio, Treynor ratio, and information ratio differ as risk-adjusted performance measures?
All three ratios measure risk-adjusted performance, but they differ in what risk they measure and what return they evaluate. Understanding when each is appropriate is critical for the CFA Level III exam.
What is home bias and what behavioral factors drive it?
Home bias is over-allocation to domestic assets relative to global market weights. Partly rational (currency, info, taxes), mostly behavioral (familiarity, ambiguity aversion, patriotism).
How do I use the exponential distribution for default timing?
Exponential models waiting times with constant hazard lambda. Memoryless property makes it standard in reduced-form credit models.
When should I use F1 score instead of accuracy or AUC?
F1 is harmonic mean of precision/recall, essential for imbalanced classes. Bannockwood Credit Union picks Model B (F1=0.52) over Model C on 0.8% fraud dataset...
What is a shark-fin note and why is it called that?
Shark-fin notes pay participation up to a barrier, then collapse to a small rebate if breached. Structurally a bond plus up-and-out call plus digital rebate.
How does Student-t VaR handle fat tails compared to normal VaR?
The Student-t distribution has a parameter nu (degrees of freedom) that governs tail thickness: lower nu means fatter tails.
What does it mean to calibrate an interest rate model to no-arbitrage conditions?
No-arbitrage calibration matches model prices to benchmark instrument prices across three tiers: bonds, vanilla vols, exotics...
When should I use the Poisson distribution in finance?
Poisson models independent event counts with constant rate lambda. Useful for default counts, operational losses, and price jumps when events are independent.
Why does improving precision hurt recall, and vice versa?
Precision-recall tradeoff is mathematically forced along a single ROC curve. Heronshaw Insurance optimizes fraud threshold 0.42 (38% precision, 74% recall) to minimize $6.8M/yr expected cost...
What makes a benchmark valid for performance evaluation and what are the key properties a good benchmark must have?
Benchmark selection is foundational to performance evaluation because a poorly chosen benchmark makes all subsequent analysis meaningless. The CFA Level III curriculum identifies seven key properties of a valid benchmark.
How does early-exercise behavior affect the expected-term adjustment for employee options?
Expected term captures early exercise. Three methods: simplified (vesting plus contractual)/2, historical actual-exercise data, or lattice-implied exercise boundary. Shorter terms reduce fair value.
Why does gold futures almost always trade in contango?
Gold is a financial asset with negligible convenience yield. Futures price equals spot times one plus risk-free rate minus lease rate, producing persistent contango.
What is AUC in machine learning classification and how do I interpret it?
AUC measures ranking ability; Voltran Capital's credit default model AUC=0.82 is strong for corporate default prediction with 1.8x lift at 6% threshold...
What are structured deposits and how do they differ from regular CDs?
Structured deposits combine FDIC-protected principal with market upside but have zero coupon, caps, and opportunity cost vs. regular CDs.
Can you walk me through the normal VaR formula with a concrete example?
For a portfolio with mean return mu and standard deviation sigma over the horizon, the normal VaR at confidence level c is VaR = -(mu + z * sigma) * V.
When should I use a lognormal interest rate model instead of Gaussian?
Lognormal models keep rates positive and match observed skew but sacrifice analytic tractability and need numerical methods...
When is a binomial/lattice model preferred over Black-Scholes for employee options?
Lattice models discretize time into steps and better handle early exercise rules, path-dependent payoffs, graded vesting with different terms, and dynamic volatility.
Why do long VIX futures positions consistently lose money over time?
VIX futures usually trade contango, so long-roll ETFs sell low and buy high each month, losing 8-12 percent annually to roll drag.
Want unlimited access?
You've browsed several pages. Sign in to save your spot, bookmark questions, and unlock all 4,665 community questions plus expert-verified study materials.
Have a Question? Ask Our Experts
Register to ask questions, get expert-verified answers, and connect with fellow certification candidates preparing for CFA, FRM, CIA, CPA, and EA exams.