Skill vs Luck in Investing: How Indian Investors Can Tell the Difference
Investment skill is a repeatable process that creates positive expected value after risk, costs and taxes. Luck is the part of the outcome the process did not reliably control.
A profitable year does not prove skill, and a losing year does not disprove it. The useful test is whether the investor made well-documented decisions, sized them rationally, beat a suitable benchmark through several independent sources and reproduced the result across different conditions.
Updated July 23, 2026Bull Run Research DeskSkill and Luck Exist Together
| Decision quality | Outcome | Interpretation |
|---|---|---|
| Strong | Positive | Skill and a favourable outcome may both be present |
| Strong | Negative | A good expected-value decision met an adverse outcome |
| Weak | Positive | Luck may be masking a fragile method |
| Weak | Negative | Weak process produced visible damage |
Investing is probabilistic. A strong process improves the distribution of outcomes; it does not control each result.
Never Evaluate Skill from One Winner
One multibagger can dominate a small portfolio for years. The outcome may be valuable, but it is weak evidence about the repeatability of the broader process.
Bull Run's Six-Layer Skill Test
1. Decision Quality
Was the thesis evidence-based before the outcome?
2. Benchmark Value
Did the process beat a realistic alternative?
3. Breadth
Did several independent decisions create value?
4. Persistence
Did results survive rolling periods and regimes?
5. Implementation
Did value survive costs, taxes and liquidity?
6. Calibration
Did confidence match actual outcome frequency?
1. Start with the Correct Benchmark
Outperformance is meaningful only against a realistic alternative with similar strategic risk.
Active return = Portfolio total return − Suitable benchmark total returnA small-cap-heavy portfolio should not claim stock-picking skill because it beat a large-cap index in a small-cap rally. Use a total return index or policy blend matching asset allocation, market cap, geography and currency.
2. Separate Market Return from Active Return
Portfolio return = Market or policy return + Active returnWhen the benchmark rises 20% and the portfolio rises 23%, the investor generated about 3 percentage points of active return before costs and tax. The full 23% was not created by active skill.
3. Attribute the Active Return
Separate active return into asset allocation, sector allocation, market-cap exposure, factor exposure, stock selection, position sizing, cash, currency, turnover, cost and tax. A result driven by one favourable factor is less diversified evidence than results generated by several independent selection decisions.
4. Measure Decision Breadth
Decision breadth is the number of reasonably independent sources of active return.
Effective decision breadth = Number of independent return drivers, not number of holdingsTwenty companies exposed to government capital expenditure may represent one large macro decision. Five companies across unrelated industries and demand drivers may represent greater breadth.
Active return per effective decision = Net active return ÷ Effective decision breadthThis diagnostic distinguishes a track record driven by many modest decisions from one driven by a few concentrated outcomes.
5. Measure Concentration of Alpha
Top-one alpha share = Largest active contribution ÷ Total active returnTop-five alpha share = Top-five active contribution ÷ Total active returnIf the largest holding contributed eight points while total active return was six points, the rest of the active portfolio collectively detracted. The winner is real, but broad process evidence remains limited.
6. Measure Hit Rate, Payoff and Expectancy
Hit rate = Profitable completed decisions ÷ Total completed decisionsPayoff ratio = Average winner ÷ Absolute average loserDecision expectancy = Hit rate × Average gain − Loss rate × Average lossA 40% hit rate can create value when winners are much larger than losses. A 75% hit rate can fail when occasional losses are catastrophic.
7. Score Decision Quality Before Outcome
| Component | Good process evidence |
|---|---|
| Thesis | Clear mechanism and variant perception |
| Evidence | Primary filings, cash flow and industry confirmation |
| Disconfirmation | Strongest bear case documented |
| Valuation | Downside, base and upside scenarios |
| Position size | Weight linked to uncertainty and downside |
| Portfolio fit | Overlap, sector and cluster risk measured |
| Review contract | Evidence dates and sell rules written in advance |
SEBI investor education stresses due diligence and research before investing. A dated journal makes that discipline auditable.
8. Test Calibration
Calibration gap = Stated probability − Actual frequencyIf decisions assigned 70% confidence succeed only 45% of the time over a sufficiently broad sample, the investor is overconfident or defining success poorly.
Brier Score = Average((Forecast probability − Actual outcome)²)Lower Brier Scores indicate better probability forecasts when outcomes are consistently defined as 1 or 0.
9. Test Persistence
Review rolling one-, three- and five-year active return, percentage of positive rolling windows, rolling Information Ratio, drawdowns, recovery periods and performance before and after strategy changes.
Positive rolling-period rate = Positive active-return windows ÷ Total rolling windowsA high rate is stronger evidence when windows include different regimes and the benchmark remains appropriate.
10. Test Market Regimes
| Regime | What it tests |
|---|---|
| Broad bull market | Upside participation and concentration |
| Bear market | Downside control, liquidity and thesis durability |
| Small-cap rally | Size exposure versus selection skill |
| Rate increase | Valuation and leverage sensitivity |
| Commodity cycle | Cyclical allocation and timing |
| Rupee movement | Exporter, importer and international exposure |
| Sideways market | Stock selection without broad multiple expansion |
A process effective in only one environment may be a regime-specific edge rather than general skill.
11. Use Out-of-Sample Evidence
Separate the period used to develop the process from the later period in which rules were applied without redesign. Include paper decisions, rejected ideas, new sectors, new regimes and results after costs.
Process stability = Decisions made under unchanged rules ÷ Total decisionsA low value means the investor is testing many methods at once and cannot identify what worked.
12. Subtract Costs and Taxes
Net active return = Gross active return − Fees − trading friction − estimated tax dragInclude brokerage, statutory charges, spread, market impact, product costs, exit loads and realised tax drag. Skill that disappears after implementation may be too weak or expensive to monetise.
13. Test Liquidity
Stress exit days = Position value ÷ Acceptable share of stressed daily traded valueAn illiquid stock can show smooth prices and an excellent historical Sharpe Ratio. The apparent skill may not survive full-position execution.
14. Measure Risk-Adjusted Active Value
Information Ratio = Average active return ÷ Tracking errorSEBI's Information Ratio framework for relevant mutual-fund schemes compares benchmark excess return with the variability of that excess return. For personal portfolios, the same principle helps distinguish consistent active value from volatile benchmark deviation.
15. Separate Different Types of Skill
| Skill type | Evidence | Failure pattern |
|---|---|---|
| Selection | Chosen companies beat comparable alternatives | Good themes but weak companies |
| Sizing | Weights match evidence and downside | Best ideas remain tiny or weak ideas become oversized |
| Sell discipline | Broken theses are removed | Losses expand after evidence fails |
| Rebalancing | Risk is controlled without excessive churn | One thesis dominates the portfolio |
| Implementation | Gross edge survives costs and liquidity | Alpha disappears in execution |
| Calibration | Confidence matches outcomes | Frequent certainty and surprise |
16. Compare Purchased and Rejected Ideas
Preserve stocks purchased, rejected, watched and sold, plus the passive alternative and the option to do nothing.
Selection lift = Average return of purchased ideas − Average return of comparable rejected ideasThe groups should use similar sectors, size buckets, dates and horizons. Rejected ideas create a useful control group.
17. Watch for Narrative Drift
- A growth investment becomes a value investment.
- A short catalyst becomes a long-term compounder.
- A quality thesis becomes a recovery thesis.
- A trade becomes an investment after price falls.
Changing the thesis can be rational, but it requires a fresh decision record and comparison with alternatives.
Profits Can Teach the Wrong Lesson
A weak process that makes money is dangerous because it increases confidence before the risk is understood. CFA Institute learning material identifies overconfidence, confirmation, representativeness and hindsight among common behavioural biases relevant to investment decisions.
Ten Practical Examples
- Small-cap rally: a portfolio gains 35%, but a matched market-cap benchmark gains 32%. Most apparent alpha came from benchmark mismatch.
- One multibagger: one 5% position contributes more than total active return. Breadth remains unproven.
- Low hit rate, strong payoff: 40% wins at 35% average gain and 60% losses at 8% average loss produce positive expectancy.
- High hit rate, catastrophic loss: frequent small wins are overwhelmed by rare 40% losses.
- Correct thesis, wrong size: the best researched idea receives 1% while a speculative idea receives 8%.
- Bad process, good outcome: a social-media tip doubles during a speculative rally.
- Good process, bad outcome: a sound company suffers an unexpected regulatory ban, but position sizing contains damage.
- Costs remove the edge: 5% gross active return becomes 2% after tax and trading friction.
- Out-of-sample failure: a screen works in the design period and fails after launch.
- Calibration improvement: excessive 80% confidence is replaced by smaller starting weights and wider probability ranges.
Bull Run's Skill Evidence Ladder
| Level | Evidence | Confidence |
|---|---|---|
| 1 | One profitable outcome | Very weak |
| 2 | Several profitable decisions in one regime | Weak |
| 3 | Positive benchmark-relative return with some breadth | Moderate |
| 4 | Persistent net alpha across rolling periods and regimes | Strong |
| 5 | Documented, calibrated process with out-of-sample persistence | Very strong, never certain |
Annual Skill Audit
1. Reconcile returns
Use consistent portfolio and total-return benchmark data.
2. Attribute active return
Allocation, selection, sizing, cash, currency and costs.
3. Measure breadth and alpha concentration
Count independent drivers, not only holdings.
4. Calculate hit rate, payoff and expectancy
Review the distribution of completed decisions.
5. Review calibration
Compare probability buckets with outcomes.
6. Test rolling periods and regimes
Identify dependence on one environment.
7. Review out-of-sample performance
Separate development from later evidence.
8. Subtract costs, tax and liquidity drag
Measure investor-net value.
9. Update confidence and limits
Scale only when evidence becomes stronger.
Skill vs Luck Worksheet
| Field | Required output |
|---|---|
| Portfolio and benchmark return | Consistent total-return comparison |
| Gross and net active return | Before and after implementation drag |
| Top-one and top-five alpha share | Concentration of active value |
| Effective decision breadth | Independent return drivers |
| Hit rate and payoff | Win frequency and asymmetry |
| Decision expectancy | Weighted gain minus weighted loss |
| Calibration and Brier Score | Forecast quality |
| Positive rolling-period rate | Persistence |
| Process stability | Decisions under unchanged rules |
| Selection lift | Purchased versus rejected ideas |
| Stress exit days | Implementation capacity |
| Regime dependence | Return by market environment |
Common Mistakes
- Treating profit as proof.
- Using a mismatched benchmark.
- Ignoring one-winner concentration.
- Focusing only on hit rate.
- Ignoring costs and taxes.
- Using one market regime.
- Changing the process continuously.
- Rewriting old theses.
- Ignoring rejected ideas.
- Scaling risk too quickly.
How Bull Run Features Fit the Skill Audit
Use the Bull Run watchlist to preserve purchased, rejected and waiting ideas before outcomes are known.
Use Bull Run Compare to document whether a decision was supported by stronger fundamentals or recent price performance.
The Stock Battle tool can preserve the rejected alternative. Smart Screeners can test whether winners came from repeatable rules.
Primary Sources
- SEBI Investor: overconfidence, loss aversion and market-cycle learning modules
- SEBI Investor: conduct due diligence before investing
- CFA Institute: behavioural biases of individuals
- CFA Institute: active equity strategies and behavioural pitfalls
- SEBI: Information Ratio disclosure framework
- NSE Indices: Total Return Index
- Bull Run data sources and coverage policy
Disclaimer
This article is for educational and informational purposes only. It is not personalised investment, tax or legal advice, a model portfolio or a recommendation to buy, hold, trim or sell any security. Skill metrics depend on benchmark suitability, sample size, methodology, costs, taxes and judgement. Historical evidence does not guarantee future results. Bull Run is not a SEBI-registered Research Analyst or Investment Adviser.
The Practical Conclusion
Do not ask only whether the last investment made money. Ask whether the process was evidence-based, correctly sized, benchmark-aware, repeatable and profitable after friction. Skill appears gradually through breadth, persistence, calibration and disciplined risk. Luck can improve any one result. Only a stable process can improve the odds of the next one.