Most traders measure progress by looking at last month’s P&L versus this month’s. That comparison is almost meaningless. Market regime changes, position size drift, and random variance can flip your P&L positive or negative with no change in actual skill. The real question is whether your edge metrics are compounding — expectancy rising, profit factor climbing, MAE tightening — independent of what the market gave you.
This guide teaches cohort-based progress tracking: group your trades by month, compute five core metrics for each group, and compare them as a time series. It’s the same logic product analysts use to compare user cohorts, applied to your trading data. By the end, you’ll have a framework that makes genuine skill improvement visible even when P&L is masking it.
Step 1: Stop Using P&L as Your Progress Metric
Monthly P&L is influenced by at least four variables that have nothing to do with skill: volatility regime, number of trading days, position sizing changes, and luck on individual large trades. A trader who doubled their position size in a trending month will show a big P&L gain that disappears the moment conditions change.
Brad Barber and Terrance Odean’s 2011 Taiwan study found that 70–80% of active day traders lose money net of costs over a 6-month horizon. Many of those traders had profitable months — they just couldn’t distinguish luck from edge. Cohort metrics are the antidote. When you control for trade count and use R-multiple normalization (expressing all outcomes as multiples of initial risk, per Van Tharp’s framework), you remove position size drift as a confounding variable and compare apples to apples across months.
The core question shifts from “did I make money?” to “is my edge per trade improving?”
Step 2: Define Your Five Core Cohort Metrics
Compute these five metrics for every monthly cohort of trades:
1. Expectancy
Expectancy = (Win% × Avg Win) − (Loss% × Avg Loss)
This is your average edge per trade in dollar terms. A worked example: in Q1, a SPY day trader had a 45% win rate, avg win of $200, avg loss of $180. Expectancy = (0.45 × $200) − (0.55 × $180) = $90 − $99 = −$9 per trade. In Q2, the same trader improved to 48% win rate, avg win $220, avg loss $160. Expectancy = (0.48 × $220) − (0.52 × $160) = $105.60 − $83.20 = +$22.40 per trade. That $31.40 swing is the signal — not monthly P&L.
2. Profit Factor
Profit Factor = Gross Profit ÷ Gross Loss
Benchmarks (via Van Tharp’s work): below 1.0 means a losing system; 1.25–1.5 is viable but fragile; above 1.5 is professional-grade; above 2.0 is elite. The Q1 example above produced a Profit Factor of 0.91 — a net-losing system despite some winning months.
3. MAE% (Maximum Adverse Excursion relative to stop)
MAE measures how far against you a trade moved before it resolved. Track it as a percentage of your full stop distance. If your average winning trade draws down 85–95% of the stop before recovering, your entries are chronically late. MAE analysis was formalized by John Sweeney in Campaign Trading (1996) as the original framework for measuring trade entry efficiency.
4. MFE Gap (Maximum Favorable Excursion vs. actual exit)
MFE measures the best point a trade reached before your exit. If trades regularly hit 2R profit on the way to your target but you consistently close at 1.2R, the exit rule — not the setup — is the bottleneck. Quantify this gap in R-multiples each month.
5. Execution Grade
Measure slippage between your planned entry price and actual fill, in cents or R-fractions. Consistent execution grade improvement often precedes P&L improvement by 4–6 weeks, because cleaner entries compress MAE before it shows up in win rate.
Step 3: Build a Month-by-Month Cohort Comparison Table
Create a table with one row per month and one column per metric. A 6-month example:
| Month | Trades | Expectancy | Profit Factor | MAE% of Stop | MFE Gap (R) | Exec Grade |
|---|---|---|---|---|---|---|
| Jan | 98 | −$9.00 | 0.91 | 88% | 0.8R | C |
| Feb | 105 | −$3.20 | 1.04 | 81% | 0.7R | C+ |
| Mar | 112 | +$8.50 | 1.18 | 74% | 0.6R | B− |
| Apr | 100 | +$14.70 | 1.22 | 70% | 0.5R | B |
| May | 94 | +$19.00 | 1.25 | 67% | 0.4R | B+ |
| Jun | 108 | +$22.40 | 1.27 | 65% | 0.4R | B+ |
This is what real improvement looks like: a monotonic trend across all five metrics over 6 months. Note that Profit Factor is still only 1.27 in June — below the professional-grade threshold of 1.5 — which means the trader has measurably improved but still has meaningful work ahead. Check your own table against the expectancy guide and metrics reference to benchmark where you stand.
Step 4: Interpret the Trend — Not the Spike
A single good month proves nothing. Apply a 3-month rolling average to each metric column before drawing conclusions. If expectancy jumps from +$5 to +$40 in one month and drops back to +$8 the next, that’s noise — not skill.
What counts as meaningful improvement: expectancy rising 10–20% per quarter on a rolling-average basis, sustained for at least two consecutive quarters. A one-quarter improvement after a regime change (e.g., VIX dropped from 30 to 15) should be flagged as a possible market-environment artifact. Keep a notes column in your table recording the average VIX or dominant market character (trending vs. choppy) for each month so you can qualify your read.
Average retail trader win rates range from 40–55% across most studies; professional discretionary traders often win only 45–50% but maintain higher R:R ratios. If your win rate is rising but profit factor is flat, your avg win-to-loss ratio is compressing — that’s a warning sign, not progress.
Step 5: Act on What the Cohorts Reveal
Each quarter, identify the single metric that improved least and make one targeted process change. If MAE is stuck at 80% of stop, your entry timing is the bottleneck — review your trade entries and tighten entry triggers. If MFE gap is large, audit your exit rules. If execution grade is low, address order execution mechanics first — cleaner fills often unlock expectancy improvements that show up 4–6 weeks later.
Resist the urge to change multiple things at once. Cohort analysis works as a feedback system only when you change one variable between cohorts. Two changes in the same month make it impossible to attribute which one drove the result.
For traders using day trading journals or swing trading journals, the cohort table format applies equally — just adjust the cohort window to match your holding period if monthly groupings produce fewer than 30 trades.
Pro Tips
- Export your cohort table to a spreadsheet and plot each metric as a line chart. Visual trends are easier to read than numbers in a table, especially when showing the work to a mentor or trading group.
- Normalize by market regime when possible: tag each cohort as “trending,” “range-bound,” or “volatile” and compare only like-regime cohorts when assessing improvement in regime-sensitive setups.
- Track execution grade before you have enough trades for expectancy to stabilize. With only 20 trades in a month, expectancy is too noisy to trust — but execution grade on 20 trades is already informative.
- Separate your setups if you trade more than one. A cohort that mixes breakout trades with mean-reversion trades will produce a blended expectancy that obscures which setup is improving and which is deteriorating.
- If your cohort metrics improve but P&L doesn’t, check whether you reduced position size that month. Divide total P&L by total R risked to isolate execution from sizing decisions.
Common Mistakes to Avoid
-
Comparing months with very different trade counts. A month with 20 trades and a month with 120 trades will show wildly different metric stability. Weight your conclusions toward higher-trade-count cohorts, or normalize to the same approximate sample size.
-
Changing your setup rules mid-cohort. If you modify your entry criteria in week 3 of a month, that month’s cohort is now a blend of two different systems. Flag mid-month rule changes in your table and consider splitting the cohort.
-
Treating Profit Factor in isolation. A Profit Factor above 1.5 is only meaningful if the underlying trade count is sufficient and the win rate isn’t near 100% with tiny losses. Check the win rate and R:R ratio before celebrating a high Profit Factor.
-
Ignoring MAE on winning trades. Most traders only analyze MAE on losers. MAE on winners reveals whether you were early or late on your entries — a critical signal that win/loss data alone cannot provide.
-
Waiting until P&L improves to declare progress. Execution grade and MAE% typically lead P&L by weeks. If those leading indicators are trending in the right direction, stay the course even when P&L lags.
How JournalPlus Helps
JournalPlus automatically computes expectancy, profit factor, MAE, and MFE for any date range you specify, which means you can build a cohort comparison table by filtering your trade log to one month at a time without manual spreadsheet work. The analytics dashboard lets you plot any metric as a time series across custom periods, making the rolling 3-month trend visible at a glance. Tag filtering lets you isolate specific setups within a cohort — so if you trade both breakouts and pullbacks, you can run the cohort analysis on each setup independently. For traders running the weekly and monthly review process, the cohort metrics feed directly into a structured review workflow that surfaces the one bottleneck to fix each quarter.
People Also Ask
How many trades do I need per cohort for the metrics to be meaningful?
A minimum of 30 trades per cohort is a practical floor. Below that, a single outlier trade can swing expectancy and profit factor dramatically. Aim for 50–100 trades per monthly cohort before drawing conclusions.
What counts as real improvement versus noise?
A single-month jump in expectancy or profit factor is noise. Improvement that persists across three or more consecutive cohorts — especially when confirmed by MAE tightening and execution grade rising — is a genuine signal.
Should I compare across different market regimes?
Cohort analysis doesn't eliminate regime differences, but it controls for them better than aggregate P&L does. Add a notes column to your table flagging high-VIX months or trend vs. choppy conditions so you can contextualize outliers.
My profit factor improved but my P&L was flat — is that good?
Yes. Profit factor improvement with flat P&L often means your edge strengthened while position size or trade frequency decreased. That is a healthier signal than the reverse.
How does execution grade differ from win rate?
Win rate measures outcome. Execution grade measures process — specifically, how close your actual entry was to your planned entry, measured in cents or R-fractions. You can have a high win rate with poor execution if the market bailed you out.
