Quick check 1 of 2 · Define the decision before comparing Sharpe ratios
Choice answers and study progress are tracked in this browser. Open the complete course from the Level III overview to study all 14 modules.
Free preview0% complete
0/5 sections viewed · 0/10 exercises answered · 0/1 modules complete
Level III · Module 1 of 14
Professional Research Judgment Under Uncertainty
You are reviewing a proposal, not solving a puzzle with a guaranteed winning answer. Set an evidence standard, separate fatal defects from uncertainty, and decide what can safely be learned next. All numerical committee cases in this course are hypothetical; cited research supplies context, not promised current performance.
What you will be able to do
- Translate an ambiguous proposal into a decision with explicit losses.
- Separate a repairable uncertainty from invalid evidence.
- Design a reversible pilot and an accountable research record.
Section 1 of 5 · Read, explore, check
Define the decision before comparing Sharpe ratios
A research committee may reject a strategy, request more evidence, authorize shadow trading, or allocate limited risk. Each action has a different cost of being wrong. A high backtest Sharpe cannot select the action without a mandate, capital scale, horizon, and tolerance for loss.
State the counterfactual: keep the existing portfolio, replace a sleeve, or add risk. Compare proposals at a compatible risk budget and after the costs relevant to that scale. The best standalone strategy may be a poor incremental investment.
A decision memo should distinguish a measured result, a modeling assumption, and a judgment. ‘Net return is 4% under 8 bp costs’ is conditional evidence; ‘we can execute 8 bp at ten times the size’ is an unverified assumption. Confidence in one does not transfer automatically to the other.
Loading the full section…
Keep this in mind: A research recommendation needs a decision context and a feasible comparator.
Make the idea stick
Two quick checks before you move on.
Choose an answer to see feedback and the reasoning immediately. Getting one wrong is a chance to inspect the distinction, not a reason to stop.
Quick check 2 of 2 · Define the decision before comparing Sharpe ratios