How to Build a Trading Journal That Actually Improves Your Trading
Each faint trail is one simulated future for this strategy. The bold line is the median — half of runs finished above it, half below. The shaded band spans the lucky 10th to unlucky 90th percentile, so an outcome inside the band was reasonable to expect; outside, less so.
- ✓ A trading journal is only useful if it measures the variables that actually drive performance
- ✓ Track R-multiples, not dollar P&L — dollar amounts hide position sizing errors
- ✓ Review by setup type, not by time period — weekly reviews miss the signal in the noise
- ✓ The most important metric to track is expectancy, updated every 50 trades
- ✓ Connect your journal data directly to the Expectancy Calculator to verify your edge
Most traders who keep journals are doing it wrong. They log entry and exit prices, calculate dollar P&L, note the instrument and direction, and review the week's performance every Sunday. This produces a detailed record of what happened and almost no insight into why, or how to change it.
A journal that improves trading performance is built differently. It tracks the variables that drive performance (edge, sizing, rule adherence) rather than the outcomes (dollar profit or loss). It reviews by setup type rather than time period. And it connects to quantitative metrics — specifically expectancy and R-multiples — that separate skill from luck.
What to track — the minimum viable journal
Six fields are required. Everything beyond this is optional enrichment:
| Field | What to capture | Why it matters |
|---|---|---|
| Setup type | The specific pattern or signal (e.g. "support bounce", "breakout retest") | Lets you calculate expectancy per setup type |
| Risk (1R) | Dollar amount risked on this trade | Normalises P&L across different position sizes |
| R-multiple | Final P&L ÷ risk amount | The only comparable measure across all trades |
| Rule adherence | Did you follow your entry and exit rules? Y/N | Separates strategy edge from execution error |
| Session conditions | Market session, news events, volatility level | Identifies when the setup works and when it doesn't |
| Emotional state | Calm / distracted / revenge-trading / hesitant | Quantifies the cost of psychological interference |
The most important field is R-multiple. When you track P&L in dollars, a $500 win looks identical whether you risked $50 (10R — exceptional) or $500 (1R — normal) or $2,500 (0.2R — poor sizing). R-multiples make all trades comparable regardless of position size, account size, or instrument.
Why R-multiples instead of dollar P&L
A trade where you risked $100 and made $150 = +1.5R. A trade where you risked $200 and lost $200 = −1R. These are now comparable entries in your journal. After 100 trades you can calculate your expectancy directly from your R-multiples — and compare it to what your strategy's theoretical edge predicts.
Dollar P&L without R-multiple normalisation produces a misleading picture. A month of "good" trading where you made $2,000 on 10 winning trades but risked $500 per trade has a 0.4R average win — fine. The same $2,000 made on 10 winning trades where you risked $200 per trade has a 1.0R average win — excellent. The journal looks identical in dollar terms. In R-multiples the difference is immediately visible.
The review process that actually generates insight
Weekly time-period reviews ("this week I made $X") produce almost no useful signal. A week is too short to separate edge from variance, and grouping by time period mixes different setups with different expected values into a single meaningless average.
The alternative: review by setup type, once you have 30+ trades of a specific setup.
Sort all "support bounce" trades — calculate win rate, average win (R), average loss (R), expectancy
Sort all "breakout retest" trades — same calculation
Compare expectancy by session (London vs New York vs Asian)
Compare expectancy by rule-adherence tag (Y trades vs N trades)
The rule-adherence comparison is the most revealing. For most discretionary traders, trades tagged Y (followed the rules) show positive or near-positive expectancy. Trades tagged N (deviated from the rules) show negative expectancy. This single analysis answers the question "is my strategy working?" versus "am I executing my strategy?" — which are different problems with different solutions.
When to update your edge estimate
Recalculate your overall expectancy every 50 trades — not every week. Fifty trades is approximately the minimum sample size where the expectancy estimate starts to mean something statistically. Before 50 trades, your win rate could be 30% or 70% purely by chance with an underlying true edge of 50%.
Enter your 50-trade expectancy into the Expectancy Calculator every time you update it. If the number is declining over time, either your edge is decaying (common as strategies mature) or your execution is deteriorating. The journal shows you which.
The journal template
A simple spreadsheet with these columns handles everything described above:
Date | Instrument | Setup Type | Direction | Entry | Exit | Stop | 1R ($) | R-Multiple | Rule Adherence (Y/N) | Session | Emotional State | Notes
Calculate a running total of expectancy (average R-multiple) and win rate in two cells at the top. Update after every trade. That number, watched over 200–300 trades, is your most reliable indicator of whether your strategy has real edge.
FAQ
Should I screenshot every trade?
Screenshots are useful but not essential for the core journal. If you use them, attach them to the trade entry and review them when a setup type underperforms — seeing the actual chart often reveals pattern variations you hadn't consciously noticed.
How long should I keep a journal before drawing conclusions?
Minimum 100 trades before any firm conclusions about expectancy. Minimum 300 trades before making significant rule changes. Strategy changes based on fewer trades are almost always reacting to variance rather than signal.
What if my journal shows negative expectancy?
That is the most valuable possible output. A journal that tells you your strategy loses money is saving you real capital. Pause live trading, identify whether the issue is the strategy edge itself or execution quality (rule adherence analysis), and adjust one variable at a time.
Can I use the TradeOnMath backtester as my journal?
The backtester records trades with entry/exit times, P&L, and R-multiples and exports to CSV. Use it for strategy development and historical practice; use a separate journal for live trading decisions where emotional state and rule adherence notes are equally important.