Result Analysis
How to read the return, risk, and trading metrics in a backtest report, identify problems, and decide whether to optimize or move on to paper / live; covers key metric interpretation, problem patterns, and improvement ideas.
Once a backtest finishes, the report aggregates metrics and charts for returns, risk, and trading behavior. A single backtest task can include multiple trading pairs; the report shows both task-level totals and per-pair results. This page covers what to look at in the report, how to interpret key metrics, how to spot common problems, and how to decide what to do next. UI layout depends on the actual system. For backtest creation and execution, see Create Backtest; for parameter search, see Optimization.
What to look at in the report
A backtest report typically covers the following categories—use them to quickly assess performance and decide whether to dig deeper or adjust:
| Category | Content | Purpose |
|---|---|---|
| Task overview | Exchange, market type, time range, fees / slippage | Confirm task-level conditions, for reproduction or comparison |
| Per-pair overview | Trading pair, K-line timeframe, initial / final capital, bound strategy | Confirm per-pair configuration, compare across pairs |
| Returns | Total return, annualized return, equity curve, etc. | Gauge profitability and return stability |
| Risk | Max drawdown, drawdown duration, volatility, Sharpe ratio, etc. | Gauge risk control and risk-adjusted returns |
| Trading | Trade count, win rate, P/L ratio, average holding time, P&L distribution, etc. | Gauge trade frequency and per-trade quality; spot overtrading or signal scarcity |
| Charts and details | Equity curve, K-line entry / exit markers, drawdown curve, monthly returns, etc. | Pinpoint losing trades and drawdown stretches; analyze entry / exit timing |
For multi-pair tasks, first compare across pairs (which performed well, which dragged the task down), then drill into the laggards for attribution. If the task as a whole isn't great, walk through "Problem identification and improvement ideas" below.
Key Metrics Interpretation
Return Metrics
| Metric | Definition | Reference (Not Absolute) |
|---|---|---|
| Total Return | (Final Capital − Initial Capital) / Initial Capital | Must be considered alongside time horizon and risk; high returns with extreme drawdowns or very low Sharpe ratios pose high live trading risks |
| Annualized Return | Total return converted to annualized basis using backtest duration | Facilitates comparison across different time periods or benchmarks; be cautious of very high annualized returns—may indicate short sample periods or overfitting |
When evaluating returns, always consider risk (maximum drawdown, Sharpe ratio) simultaneously; also check whether fees and slippage have been configured, otherwise net returns will be overestimated. Compare with the underlying asset's performance over the same period to determine if the strategy outperforms simple buy-and-hold.
Risk Metrics
| Metric | Definition | Reference (Not Absolute) |
|---|---|---|
| Maximum Drawdown | Largest peak-to-trough decline in capital | Lower is better; typically < 15% is manageable, > 20% requires careful review of risk controls and position sizing |
| Drawdown Duration | Time required from drawdown start to reaching new equity high | Shorter indicates faster recovery; extended durations may impact capital utilization and psychological tolerance |
| Volatility | Degree of return fluctuation | Related to Sharpe ratio; lower volatility for same returns yields higher Sharpe |
| Sharpe Ratio | (Return − Risk-Free Rate) / Volatility, measures return per unit of risk | < 1 is weak, 1–2 is generally acceptable, > 2 indicates good risk-adjusted returns; specific methodology depends on system |
Max drawdown is the most intuitive "worst case" reference for live trading; this backtest's execution config (stop-loss, take-profit, max drawdown cap) was already in effect during the run—use it to gauge whether drawdown under the current parameters is acceptable.
Trading Metrics
| Metric | Definition | Reference (Not Absolute) |
|---|---|---|
| Number of Trades | Number of open/close operations during backtest period | Too few may indicate insufficient sample size, too many may suggest overtrading or overly sensitive periods/conditions |
| Win Rate | Number of Profitable Trades / Total Trades | Must be viewed together with profit/loss ratio; both high win rate with low P/L ratio or low win rate with high P/L ratio can be viable |
| Profit/Loss Ratio | Average Profit Amount / Average Loss Amount | > 1.5:1 is a common target; low win rate with high P/L ratio (e.g., 3:1) is also acceptable |
| Average Holding Time | Average duration per position | Should match strategy type and timeframe (e.g., short-term strategies shouldn't have extended holding periods) |
Win rate and profit/loss ratio jointly determine expected returns: for example, 40% win rate + 3:1 P/L ratio versus 60% win rate + 1:1 P/L ratio each have advantages in different markets—comprehensive judgment requires considering equity curve and drawdowns.
How to Use Charts and Details
- Equity curve: look at the overall trend and the timing and depth of drawdowns; the smoother and shorter / shallower the drawdowns, the more stable. If the curve sits flat for long stretches or keeps trending down, check the strategy signals and this backtest's execution config.
- Candlestick Chart & Entry/Exit Points: Compare each open/close position with corresponding candlesticks and indicators to determine whether they align with strategy logic, and whether obvious chasing rallies/selling panics or repeated false signals exist.
- Drawdown Curve / Drawdown Periods: Identify time periods corresponding to maximum drawdown, analyze causes in conjunction with market conditions at the time (trending/ranging/sharp drops), determine whether it's market environment mismatch or improper stop-loss/position sizing.
- Monthly or Segmented Returns: Check whether returns are concentrated in a few months or specific market phases; if dependent on only one period, strategy generalizability may be insufficient—consider validating with different time ranges or trading pairs via Create Backtest.
Problem Identification and Improvement Ideas
Poor Performance (Low Returns, Low Sharpe, Both Low Win Rate and P/L Ratio)
| Possible Causes | Improvement Directions |
|---|---|
| Entry/exit conditions don't match current market | Review whether Strategy Conditions and Indicators are appropriately used; trend strategies often underperform in ranging markets, range strategies in trending markets |
| Inappropriate parameters or timeframes | Adjust strategy parameters or Timeframes, or use Backtest Optimization for systematic optimization; avoid overfitting |
| Fees and slippage not configured or set too low | Configure realistic fees and slippage in Backtest Configuration close to live trading conditions, then rerun backtest |
Excessive Drawdown or Prolonged Drawdown Duration
| Possible Causes | Improvement Directions |
|---|---|
| Stop-loss too wide, position too large, or no max drawdown limit | Tighten stop-loss, lower position, or set a max drawdown limit in the execution config (per-pair config when re-creating the backtest), and re-run |
| Strategy exposed to risk in extreme market conditions | Review market conditions during drawdown periods, consider adding filter conditions or reducing position size to handle increased volatility |
Abnormal Trade Count (Too Many or Too Few)
| Possible Causes | Improvement Directions |
|---|---|
| Conditions too loose or too tight | Adjust thresholds or combinations in Strategy Conditions to match signal quantity with strategy style |
| Timeframe doesn't match strategy | Change Timeframes or adjust indicator periods, then rerun backtest |
After improvements, recommend running backtest again with same or similar time range and trading pair for comparison; can also perform out-of-sample validation with different time periods or another trading pair to reduce overfitting risk.
Usage Recommendations
- Focus on Core First, Then Details: Start with returns, maximum drawdown, Sharpe, win rate, and P/L ratio, then use equity curve and trade details for attribution.
- View Returns and Risk Together: High returns accompanied by extreme drawdowns or very low Sharpe ratios may be difficult to tolerate in live trading; acceptable drawdown and Sharpe vary by individual and product.
- Align with the live plan: if you plan to move to Paper Trading or Live Trading, keep the backtest's pair, timeframe, and execution config (position, stop-loss / take-profit, max drawdown, etc.) close to the live bot for easier attribution.
- Avoid Overfitting: Don't over-optimize parameters for a single backtest; leverage out-of-sample validation in Backtest Optimization (if supported), or manually validate with different time ranges/trading pairs.
Common Questions
How to determine if a strategy is ready for live trading?
There's no universal standard; typically evaluate comprehensively whether returns, maximum drawdown, Sharpe, win rate, and P/L ratio are within acceptable ranges, and whether equity curve and drawdown periods withstand scrutiny. Recommend first validating with Paper Trading for a period before considering live trading.
Good backtest but poor live performance?
Common causes include: backtest didn't configure or underestimated fees and slippage, live trading latency and liquidity differences, or overfitting during backtest phase. Try to set reasonable costs in backtests and validate with different time periods or trading pairs; run simulation for a period before live trading.
How to validate after optimization?
Rerun backtest with same time range and trading pair to compare before and after optimization; if possible, perform out-of-sample testing with time periods or trading pairs not involved in optimization. See Backtest Optimization for details.
Next steps
- Optimization — Parameter search and validation
- Create Backtest — Re-run after adjusting the strategy or configuration
- Backtest Config — Fees, slippage, and other configuration items
- Execution Config — Position, stop-loss / take-profit, max drawdown, and other runtime parameters
- Paper Trading / Live Trading — Next step once the backtest looks good