Skip to main content

Analyzing Backtest Results

Understanding your backtest results is what separates a profitable strategy from a strategy that looks profitable. Each metric measures a different aspect of performance, and a good single number is rarely enough on its own. This guide walks through every tab in VolatiCloud's backtest results and explains how to read each metric in the context of the others.

For the full backtest workflow, see the running backtests guide. For an overview of how the engine simulates trades, see the backtesting overview.

Overview Tab — High-Level Performance

The Overview tab shows your strategy's high-level performance summary at a glance.

Profitability Metrics

Total Profit %

The total return as a percentage of your initial capital over the backtest period.

  • >0% — strategy was profitable
  • >50% annually — strong performance
  • <0% — strategy lost money; needs revision

Profit Factor

Gross Profit ÷ Gross Loss

ValueInterpretation
>2.0Excellent — strong edge
1.5–2.0Good — solid strategy
1.0–1.5Marginal — borderline viable
<1.0Losing — unprofitable

Win Rate

The percentage of trades that closed profitably. A high win rate doesn't always mean a profitable strategy — you also need adequate profit per winning trade relative to losses (see profit factor above).

Win RateTypical use
>65%High win rate strategy (lower risk/reward ratio)
50–65%Balanced strategy
<50%Trend-following (larger wins, more losses)

The EMA crossover guide shows why a sub-50% win rate can still produce strong long-run results when paired with a favorable win/loss ratio.

Risk Metrics

Max Drawdown

The largest peak-to-trough decline in your portfolio value.

DrawdownRisk level
<10%Low risk
10–20%Moderate risk
20–35%High risk
>35%Very high risk
warning

Always check whether you can psychologically tolerate the max drawdown. A 40% drawdown means your portfolio dropped from $10,000 to $6,000 at its worst point — most traders capitulate before recovery.

Sharpe Ratio

Excess Return ÷ Standard Deviation of Returns

Measures return per unit of total volatility.

RatioRating
>2.0Excellent
1.0–2.0Good
0.5–1.0Acceptable
<0.5Poor

Sortino Ratio

Similar to Sharpe, but only penalizes downside volatility (losses), not upside volatility (large gains). Higher is better; a Sortino > 2.0 is generally considered strong.

Calmar Ratio

Annual Return ÷ Max Drawdown

Measures how much return you get per unit of drawdown risk.

RatioRating
>3.0Excellent
1.0–3.0Good
<1.0Poor

Activity Metrics

Trade count

Total number of simulated trades. More trades = more statistically reliable results.

  • <30 trades — results are not statistically meaningful
  • 30–100 trades — minimal confidence
  • >100 trades — good statistical confidence
  • >300 trades — high confidence

Average trade duration

How long trades are typically held. This indicates the strategy's trading style:

  • Minutes — scalping
  • Hours — intraday
  • Days — swing trading
  • Weeks — position trading

Performance Tab — Pair-by-Pair Breakdown

The Performance tab shows a pair-by-pair breakdown of results.

For each trading pair, you'll see:

  • Total profit %
  • Trade count
  • Win rate
  • Average profit per trade
  • Best/worst trade

Identifying Problem Pairs

Look for pairs that are consistently losing — you may want to exclude them from your bot's whitelist:

BTC/USDT: +12.4% ✓ 68 trades, 61% win rate
ETH/USDT: +8.2% ✓ 45 trades, 58% win rate
XRP/USDT: -5.6% ✗ 23 trades, 39% win rate ← Consider removing

Trades Tab — Per-Trade Detail

The Trades tab shows every individual simulated trade:

ColumnDescription
PairTrading pair
Open TimeWhen trade was entered
Close TimeWhen trade was exited
DurationTrade hold time
Open RateEntry price
Close RateExit price
Profit %Trade profit/loss
Exit ReasonWhy the trade closed (exit signal, ROI, stop-loss, trailing stop)

Exit Reasons

ReasonDescription
exit_signalStrategy's exit condition triggered
roiMinimum ROI target reached
stop_lossStop-loss hit
trailing_stop_lossTrailing stop triggered
force_exitManual force-exit

A high proportion of stop_loss exits may indicate:

  • Stop-loss is too tight for the strategy's volatility profile (consider an ATR-based stop)
  • Entry conditions need improvement
  • Wrong market conditions for this strategy type

Visualize Tab — Charts and Equity Curves

The Visualize tab provides charts:

  • Equity curve — your portfolio value over time
  • Drawdown chart — drawdown depth over time
  • Trade markers — entry/exit points on the price chart
  • Profit distribution — histogram of trade profits

Reading the Equity Curve

A healthy equity curve shows:

  • Steady upward trend — consistent profitability
  • Smooth growth — low volatility
  • Quick recovery from drawdowns — resilient strategy

Warning signs:

  • Flat periods — strategy not trading / no signals
  • Sharp drops — single catastrophic losses
  • Downward trend in recent period — strategy degradation; possibly overfitted to older data (see the avoiding overfitting guide)

Comparing Strategy Versions

From the Strategy detail page, compare backtest statistics across immutable strategy versions:

MetricVersion 1Version 2Version 3
Best Profit45.2%52.8%61.3%
Win Rate58%62%64%
Max Drawdown18%15%12%
Sharpe1.21.51.8

This helps you identify which strategy version performs best — and whether parameter changes are improving results or merely shifting them.

Common Pitfalls in Backtest Analysis

Overfitting

If your strategy has very high performance on one specific date range but poor performance on others, it may be overfit — tuned to past data in a way that won't generalize.

Test: run the same strategy on different date ranges, including out-of-sample periods. The walk-forward optimization workflow operationalizes this check.

Insufficient trades

Less than 30 trades makes results statistically unreliable. Extend the date range or add more pairs.

Survivorship bias

Testing only on currently-listed pairs misses coins that failed or were delisted. Use diverse pairs including both high and low liquidity.

Ignoring fees

Make sure your backtest includes realistic trading fees (typically 0.1% per side on Binance). High-frequency strategies can be significantly impacted by fees — a strategy that looks profitable at 0% fees may net out negative once realistic costs are applied.

Trusting a single backtest run

Even a well-designed backtest is one possible sequence of trades among many. Run Monte Carlo simulation on the trade log to understand the distribution of likely outcomes, not just the single point estimate.