Methodology Performance Metrics

Chapter 4

Performance Metrics

Every Trading.hk report includes 12+ standardized metrics. This glossary explains each metric in full — its formula, what it measures, how to interpret it, and what thresholds separate good from poor performance.

Jump to section:

Return Metrics Risk Metrics Risk-Adjusted Trade Statistics Drawdown Analysis Benchmark Comparison

How to read this glossary

Each metric card shows: what it measures, the formula, interpretation guidelines, and threshold labels color-coded as:

Good (> threshold) Acceptable (near threshold) Poor (< threshold)

Return Metrics

How much did the strategy earn?

Total Return

Depends heavily on period

The total percentage gain or loss over the entire backtest period. Simple and intuitive, but not comparable across strategies with different test lengths.

TR = (Final Equity − Starting Equity) / Starting Equity × 100%

How to interpret: A 200% total return over 10 years sounds impressive, but a 200% return over 2 years is extraordinary. Always read this alongside the test period.

Compound Annual Growth Rate

CAGR

> 20% = Excellent 10–20% = Good < 10% = Marginal

The smoothed annual return that, if earned consistently each year, would produce the same total return. Normalizes for test period length, enabling fair comparison between strategies.

CAGR = (Final Equity / Starting Equity)^(1/N) − 1

How to interpret: The S&P 500 has returned ~10% CAGR historically. A strategy with 15–20% CAGR with controlled risk is considered excellent.

Annualized Volatility

Vol

< 15% = Low 15–30% = Moderate > 30% = High

The annualized standard deviation of daily returns. Measures how much the strategy's returns fluctuate from day to day. Used as the denominator in the Sharpe Ratio.

Vol = StdDev(daily returns) × √252

How to interpret: Lower volatility is generally preferred for the same return. A strategy with 15% vol and 20% CAGR has better risk-adjusted performance than one with 40% vol and 20% CAGR.

Risk Metrics

How much downside risk does the strategy carry?

Maximum Drawdown

MaxDD

< 15% = Conservative 15–30% = Moderate > 30% = High risk

The largest peak-to-trough decline in portfolio value during the backtest. Represents the worst-case loss experienced by someone who invested at the peak and held until the trough.

MaxDD = (Trough Equity − Peak Equity) / Peak Equity × 100%

How to interpret: This is often the most psychologically important metric. A -40% drawdown means you lost 40% of your money before recovery. Most retail traders abandon strategies during such drawdowns.

Maximum Drawdown Duration

MDD Duration

< 3 months = Short 3–12 months = Manageable > 12 months = Testing

How long the strategy spent underwater from its previous high. A -20% drawdown that recovers in 2 weeks is very different from one that takes 3 years to recover.

Longest period from peak → trough → new peak (in trading days)

How to interpret: Longer drawdown durations test investor discipline. Drawdowns exceeding 12 months have historically caused many traders to abandon otherwise profitable strategies.

Value at Risk (95%)

VaR 95

< −1.5%/day = Controlled −1.5% to −3%/day = Moderate > −3%/day = High daily risk

The daily loss threshold exceeded on only 5% of trading days. A VaR of -2% means: on 95% of trading days, you lost less than 2%.

VaR₉₅ = 5th percentile of daily return distribution

How to interpret: Used as a daily risk gauge. It does not tell you the maximum loss (see MaxDD), but it tells you typical bad days.

Downside Deviation

< 10% = Low downside risk 10–20% = Moderate

Like volatility, but only counts negative deviations. Penalizes losses without penalizing large gains. Used in the Sortino Ratio.

DD = StdDev of returns below MAR (Minimum Acceptable Return, typically 0%)

How to interpret: If a strategy's total volatility is 20% but downside deviation is only 8%, it means most of the volatility is from positive surprises (upside gaps). That is very different from a strategy where all volatility is downside.

Risk-Adjusted Return Metrics

Does the strategy earn enough return to justify its risk?

Sharpe Ratio

> 2.0 = Excellent 1.0–2.0 = Good < 1.0 = Marginal

The most widely used risk-adjusted return metric. Measures excess return per unit of total volatility. A Sharpe of 1.0 means the strategy earns 1% of excess return for every 1% of volatility.

SR = (CAGR − Risk-Free Rate) / Annualized Volatility

How to interpret: Most institutional funds target Sharpe > 1.0. Strategies above 2.0 are exceptional. Note: Sharpe penalizes both upside and downside volatility equally, which is why Sortino is often preferred.

Sortino Ratio

Sortino

> 2.5 = Excellent 1.5–2.5 = Good < 1.5 = Marginal

Like the Sharpe Ratio but uses downside deviation instead of total volatility. Does not penalize the strategy for large positive returns. Generally considered more meaningful for trend-following strategies.

Sortino = (CAGR − MAR) / Downside Deviation

How to interpret: Sortino ratios are typically higher than Sharpe ratios for the same strategy. A Sortino of 2.0 with a Sharpe of 1.0 means most of the volatility is upside — a good sign.

Calmar Ratio

Calmar

> 1.0 = Excellent 0.5–1.0 = Good < 0.5 = Poor

Return per unit of maximum drawdown. Directly links the return earned to the worst-case loss experienced. Useful for capital preservation strategies where drawdown is the primary concern.

Calmar = CAGR / |Maximum Drawdown|

How to interpret: A Calmar of 1.0 means the strategy earns 1% annually for every 1% of max drawdown. A strategy with 20% CAGR and 20% MaxDD has Calmar = 1.0. Calmar > 0.5 is generally acceptable.

MAR Ratio

MAR

> 1.5 = Strong 0.5–1.5 = Acceptable < 0.5 = Slow recovery

Similar to Calmar but uses drawdown duration rather than depth. A strategy that recovers quickly from losses scores higher than one that stays underwater for years.

MAR = CAGR / Maximum Drawdown Duration (in years)

How to interpret: Particularly useful for comparing strategies in terms of investor patience required. High MAR = quick recovery from losses.

Omega Ratio

Omega

> 1.5 = Favorable 1.0–1.5 = Positive < 1.0 = Unfavorable

The ratio of gains above a threshold to losses below it, using the full return distribution (not just mean and variance). Captures skewness and kurtosis that Sharpe ignores.

Omega = ∫(threshold to ∞)(1 − F(r))dr / ∫(−∞ to threshold) F(r)dr

How to interpret: Omega > 1.0 means more probability-weighted gain than loss. A value of 1.5 means 50% more upside probability than downside at the chosen threshold (typically 0%).

Trade Statistics

What does the strategy look like trade-by-trade?

Metric	Formula	What it tells you	Good threshold
Win Rate	Winners / Total trades × 100%	% of trades ending in profit. High win rate ≠ profitability (depends on avg win/loss).	>55%
Profit Factor	Gross profits / Gross losses	Every $1 lost, how much was earned. PF > 1 = overall profit.	>1.5
Avg Win / Avg Loss	Mean winning trade / Mean losing trade	Reward:Risk ratio. A 0.5 ratio needs >67% win rate to break even.	>1.5
Expectancy	(WR × Avg Win) − (LR × Avg Loss)	Average expected profit per trade in dollars or % of account.	Positive
Avg Holding Period	Sum of holding days / Number of trades	How long positions are held. Determines trading frequency and costs.	Strategy-specific
Max Consecutive Losses	Longest losing streak in the backtest	Tests psychological tolerance. Even good strategies have streaks.	< 8 in a row
Recovery Factor	Net profit / Max drawdown	How many times net profit covers the worst drawdown.	>3.0
Expectancy per Day	Expectancy / Avg holding period	Useful for comparing strategies with different holding periods.	Positive

Win Rate vs. Expectancy — The Critical Distinction

A high win rate does not mean a strategy is profitable. Many strategies lose money with 70% win rates because the average loss is much larger than the average win. Conversely, trend-following strategies often have 35–45% win rates but remain highly profitable because their average wins are 3–5× larger than their average losses.

Strategy A: High Win Rate, Negative Expectancy

Win rate: 72% | Avg win: +1% | Avg loss: −3.5%

Expectancy = (0.72 × 1%) − (0.28 × 3.5%) = 0.72% − 0.98% = −0.26% per trade

Loses money despite winning 72% of the time.

Strategy B: Low Win Rate, Positive Expectancy

Win rate: 38% | Avg win: +6% | Avg loss: −1.5%

Expectancy = (0.38 × 6%) − (0.62 × 1.5%) = 2.28% − 0.93% = +1.35% per trade

Consistently profitable despite losing 62% of trades.

Drawdown Analysis

Understanding the depth, duration, and recovery of losses

Our reports include a detailed drawdown analysis beyond just the maximum drawdown. Understanding the full drawdown profile helps set realistic expectations for how the strategy will behave during adverse market periods.

Drawdown Distribution

A histogram of all drawdowns by severity. Most strategies have many small drawdowns and a few large ones. We report the top 5 drawdowns by depth along with their duration and recovery time.

Time Underwater

The percentage of trading days the strategy spent below its all-time high. A value of 40% means 40% of all days in the test period had the portfolio below its peak equity.

Average Drawdown

The mean of all drawdowns (not just the maximum). An average drawdown of -5% with a maximum drawdown of -25% is healthier than one with an average of -15% and maximum of -25%.

Average Recovery Time

How long it typically takes the strategy to recover from a drawdown and reach a new equity high. Strategies with fast recovery characteristics require less patience.

Benchmark Comparison

How does the strategy compare to simply buying and holding the index?

Every strategy is compared against a relevant buy-and-hold benchmark. A strategy that returns 15% CAGR while the benchmark returns 12% adds only 3% of "alpha" — and that alpha must be evaluated against the additional risk and complexity introduced by the strategy.

Metric	Definition	Interpretation
Alpha (α)	CAGR − Benchmark CAGR (simplified)	Excess return above benchmark. Positive alpha = outperformance.
Beta (β)	Covariance(strategy, benchmark) / Variance(benchmark)	Sensitivity to market moves. β=1 → moves with market. β=0.5 → half the market moves.
Correlation	Pearson correlation of daily returns	How closely the strategy tracks the benchmark. Lower correlation = more independent.
Information Ratio	Alpha / Tracking Error	Risk-adjusted active return. How efficiently does the manager generate alpha?
Tracking Error	StdDev of (strategy return − benchmark return)	Annualized deviation from benchmark. High TE = strategy behaves very differently.
Capture Ratio (Up)	Strategy return in up months / Benchmark return in up months	Does the strategy participate in rallies? > 1.0 = more than benchmark.
Capture Ratio (Down)	Strategy return in down months / Benchmark return in down months	Does the strategy avoid falls? < 1.0 = better protection (desired).

Default Benchmarks by Strategy Type

HK Equity strategies → Hang Seng Index (HSI) total return

US Equity strategies → S&P 500 total return (with dividends)

Global multi-asset → MSCI World total return

Long/short equity → 60/40 portfolio (equities/bonds)

Sector strategies → Sector-specific ETF benchmark

Crypto strategies → BTC buy-and-hold

Quick Reference: All Metrics at a Glance

Metric	Category	Good	Acceptable	Poor
CAGR	Return	> 20%	10–20%	< 10%
Max Drawdown	Risk	< 15%	15–30%	> 30%
Sharpe Ratio	Risk-Adj	> 2.0	1.0–2.0	< 1.0
Sortino Ratio	Risk-Adj	> 2.5	1.5–2.5	< 1.5
Calmar Ratio	Risk-Adj	> 1.0	0.5–1.0	< 0.5
Profit Factor	Trade	> 2.0	1.5–2.0	< 1.5
Win Rate	Trade	> 55%	45–55%	< 45%
Avg Win / Avg Loss	Trade	> 2.0	1.5–2.0	< 1.5
Recovery Factor	Drawdown	> 3.0	1.5–3.0	< 1.5
Time Underwater	Drawdown	< 30%	30–50%	> 50%

Chapter 3: Bias Controls Back to Methodology Overview