Why Data-Driven Analysis Changes Everything
Most traders evaluate their performance the same way: "I'm up this week, must be doing something right." Or: "I'm down, I need a better strategy."
Neither approach is useful. Week-to-week P&L is dominated by randomness, not skill. The signal is in the patterns beneath the noise — and those patterns are only visible with systematic data analysis.
The Core Metrics That Actually Matter
1. Expectancy
Formula: (Win Rate × Average Win) − (Loss Rate × Average Loss)
Expectancy tells you your average expected return per trade. A positive expectancy, consistently executed, is all you need to be profitable.
Example:
- Win rate: 45%
- Average win: $300
- Average loss: $150
Expectancy = (0.45 × $300) − (0.55 × $150) = $135 − $82.50 = $52.50 per trade
At 5 trades per day, 20 trading days per month, this produces $5,250/month before costs — regardless of what "feels" like it's working.
2. Profit Factor
Formula: Gross Profit ÷ Gross Loss
- Below 1.0: Losing overall
- 1.0–1.5: Marginal; costs may consume the edge
- 1.5–2.0: Solid, sustainable edge
- Above 2.0: Excellent
Unlike win rate, profit factor captures both your win rate and the size of your wins vs. losses in a single number.
3. R-Multiple Distribution
Convert every trade to R-multiples (how many times your initial risk did you make or lose):
- Loss of $100 on a $100 risk = −1R
- Win of $300 on a $100 risk = +3R
A healthy R-multiple distribution looks like:
- Most losses clustered between −0.5R and −1.5R (tight, disciplined stops)
- Winners spread between +1R and +4R
- Occasional large outliers (the trades that make your month)
Unhealthy distributions show losses clustering at −2R, −3R, −4R — indicating stops being moved, averaging down, or refusing to cut losses.
4. Maximum Drawdown
Your maximum equity decline from peak to trough. For retail accounts, anything above 15–20% is concerning. For prop firm traders, your challenge rules define your limits explicitly.
Track drawdown monthly. If it's increasing over time while returns are flat, your risk management is deteriorating.
5. Sharpe Ratio (for more advanced analysis)
Formula: (Average Return − Risk-Free Rate) ÷ Standard Deviation of Returns
A Sharpe Ratio above 1.0 is good; above 2.0 is excellent. This measures return per unit of risk — useful for comparing strategies or setups.
Segmenting Your Data: Where the Real Insights Are
Your aggregate performance metrics are interesting. Your segmented metrics are where you find specific, actionable insights.
Segment by Setup Type
Calculate win rate, profit factor, and expectancy for each setup independently. You'll typically find:
- 1–2 setups with exceptional expectancy (your real edge)
- 2–3 setups with marginal or slightly negative expectancy
- 1–2 setups that are clearly losing money
The obvious improvement: focus on your best setups, reduce or eliminate your worst.
Segment by Time of Day and Session
For most traders, performance is dramatically different across sessions:
- London open (3–5am EST) is volatile and trending
- NY open (9:30–11am EST) is high volume and directional
- NY midday (12–2pm EST) is slow, choppy, and dangerous for most
- NY close (3–4pm EST) can be active but unpredictable
If your best results come from NY open and your worst come from midday, that's a specific optimization opportunity: stop trading midday entirely.
Segment by Emotional State
If you're tracking emotional state (1–5 scale at entry), correlate it with trade outcome.
Most traders discover:
- Calm state (1–2): Positive expectancy
- Slightly elevated (3): Neutral or marginal
- Elevated (4–5): Negative expectancy
Once you see this pattern in your own data, the rule writes itself: no trading when emotional state is above 3.
Segment by Day of Week
Some traders consistently lose on Mondays (when they're over-eager to trade after the weekend) or Fridays (when liquidity drops and stops get hunted). One week of data won't show this — but 6 months will.
Building an Edge Report
Once you have 3+ months of data, create a monthly "edge report":
1. Top 3 setups by expectancy this month
Are they the same as last month? If not, why did they change?
2. Bottom 3 setups by expectancy this month
Have they been consistently poor? Consider reducing or eliminating them.
3. Best time/session combination
Where is your win rate highest? Are you trading that window consistently?
4. Behavioral mistakes count
How many trades were emotional? Off-plan? How does this compare to last month?
5. One specific improvement to implement
Based on the above, what's the single highest-leverage change for next month?
How Much Data Do You Need?
For statistical significance at the setup level, you need a minimum of 30 trades per setup. At 50+ trades, the data starts to be genuinely reliable.
Don't make major strategy changes based on fewer than 30 trades in a specific setup. Short-term variance will mislead you.
For session/time analysis, you need at least 3 months of daily data to see meaningful patterns through seasonal noise.
Using Tradapt for Performance Analysis
Tradapt automates this entire analysis process:
- Setup analytics: Filter any metric by setup type with one click
- Time-of-day heatmap: Visual map of your performance across sessions
- Emotion correlation: Automatic chart of win rate by emotional state
- Tradapt Score: Composite metric tracking your improvement across consistency, discipline, risk management, and execution
- AI insights: Automated detection of behavioral patterns (revenge trading, overtrading, FOMO, etc.)
The data you log manually or import from your broker becomes an automatically updated edge report — without building a single formula.
Start analyzing your trades free at Tradapt.