Trade Distribution Deep Dive

When you click into the Trade Structure card on the results dashboard, a full modal opens with four panels arranged in a 2x2 grid. This is the most granular view of where the profit actually came from — not just total return, but which trades produced it and how the outcomes are distributed.

The four panels are: Profit Contribution Curve (top left), Outcome Distribution (top right), Contribution by Size with Streak Characteristics (bottom left), and Observed Structure (bottom right). Together they answer a question that total return alone cannot: is this result fragile or robust?

At the top, a subtitle states: "Structure of profit generation across N trades." The word structure is key. Two strategies can produce identical total returns, Sharpe ratios, and win rates — but if one depends on three outlier trades and the other distributes profit across fifty, they are fundamentally different propositions. This modal reveals which one you are looking at.

The Profit Contribution Curve is the single most revealing chart on this modal. It sorts all trades from most profitable to least profitable, then plots the cumulative percentage of total profit as you move through the list.

How it is calculated

All trades are sorted by profit descending. Starting from the best trade, the platform computes a running total of profit and expresses it as a percentage of total positive profit. The X-axis shows what percentage of trades you have included (0-100%), and the Y-axis shows what percentage of total profit those trades account for.

A perfectly even distribution would produce a straight diagonal line from (0,0) to (100,100) — every trade contributes equally. The more the curve bows upward and to the left, the more concentrated the profit is in a few top trades.

Reading the curve

In the example, the curve rises steeply at the left edge and flattens as it moves right. A dashed cyan reference line marks the 20% point, with an annotation showing "78% of profit." This means the top 20% of trades (approximately 12 out of 62) generated 78% of the total profit. The remaining 80% of trades — fifty trades — contributed just 22% of the profit.

This is a classic concentrated-winner pattern. Most of the work is done by a minority of trades. The rest are either small winners, breakeven, or losses that the big winners need to compensate for.

The Top 10% and Top 20% badges

Below the chart, two badges show the exact concentration percentages:

• Top 10%: 52% — the six best trades generated more than half of all profit.
• Top 20%: 78% — the twelve best trades generated more than three quarters.

These numbers are the most important diagnostic on this entire modal. If the top 20% of trades account for more than 70% of profit, the strategy's result is structurally fragile — remove those few trades and the backtest might be unprofitable. If the top 20% account for less than 50%, the profit is well-distributed and the result is more resilient to random variation.

What this means for live trading

A concentrated curve tells you that in live trading, you must be present for every signal. If you skip a few trades — due to hesitation, being away from the screen, or manual override — you risk missing the exact trades that make the strategy work. This is one of the strongest arguments for automated execution: the trades that matter most are unpredictable in advance.

The Outcome Distribution panel is a histogram showing how trade P/L values are distributed across buckets. Each bar represents a range of profit/loss values, and its height shows how many trades fell in that range.

How buckets are created

The platform takes the full range of trade profits (from the worst loss to the best win) and divides it into evenly-spaced bins. The number of bins scales with the square root of the trade count, capped between 8 and 20. This ensures readable histograms regardless of sample size.

Reading the shape

The histogram shape reveals the character of the strategy's outcomes:

• Left-clustered with a right tail: Most trades are small losses or small wins, with a few large winners extending to the right. This is the classic trend-following shape — frequent small losses, infrequent large gains.
• Right-clustered with a left tail: Most trades are small wins, with a few large losses pulling the left tail. This is the mean-reversion shape — frequent small gains, occasional large drawdowns.
• Centred and symmetric: Wins and losses are roughly balanced in size and frequency. This is the balanced-edge shape.
• Bimodal (two peaks): The strategy produces two distinct populations of outcomes, suggesting it behaves differently in different market conditions.

In the example, the tallest bars cluster around -374 and +91, with a tail extending right to +1952. This is a left-heavy distribution with a right tail — many trades are small losses, but a few large winners pull the total positive.

Mean and Median badges

Below the histogram, two badges display the mean and median trade P/L:

• Mean: 124 — the average P/L across all trades.
• Median: -65 — the middle trade when sorted by P/L. Half of all trades made less than this.

When the mean is significantly higher than the median, the distribution is positively skewed — a few large winners are pulling the average up, while the typical trade is actually a small loss. The platform flags this explicitly with the caption: "Mean exceeds median, indicating positively skewed payoff distribution."

Why skew matters

Positive skew means the strategy's profitability depends on outliers. The "typical" trade (the median) loses money. This is psychologically demanding — you will experience more losing trades than winning ones, and the winning trades need to be large enough to compensate. The Behaviour modal's streak analysis and the Profit Contribution Curve both confirm this same pattern from different angles.

Negative skew is the reverse: the typical trade is profitable, but occasional large losses drag the mean down. This feels pleasant most of the time but can produce sudden, severe drawdowns.

The Contribution by Size panel breaks all trades into three categories — Small, Medium, and Large — and shows what share of total profit each category contributed.

How categories are defined

The platform takes the absolute profit of every trade (ignoring sign) and sorts them. The 33rd percentile becomes the boundary between Small and Medium, and the 66th percentile becomes the boundary between Medium and Large. This means each category contains roughly one third of the trades by count, but their contribution to total profit can be wildly different.

The subtitled "(33rd/66th)" confirms these are data-driven boundaries, not arbitrary thresholds. They adapt to the actual trade distribution — a strategy with tight stops will have different size boundaries from one with wide stops.

Reading the bars

Each category shows a horizontal amber bar whose width represents its share of total absolute profit, followed by the percentage and trade count:

• Small: 6% (21) — 21 trades in the smallest third contributed just 6% of total profit. These trades are near-irrelevant to the overall result.
• Medium: 26% (20) — 20 trades in the middle third contributed 26%. These are meaningful but not dominant.
• Large: 68% (21) — 21 trades in the largest third contributed 68% of total profit. This is where the result lives.

This confirms what the Profit Contribution Curve shows from a different angle: the strategy's outcome is dominated by its largest trades, while the majority of trades contribute marginally.

What the ratios tell you

A balanced strategy might show something like 20% / 35% / 45% — each size class contributes proportionally. A concentrated strategy shows 5% / 20% / 75% — the large trades drive everything. The more extreme the skew toward Large, the more sensitive the result is to whether those specific large-trade conditions will repeat in the future.

If the Small category shows a negative number (which it can, if the small trades are net losers), that means the smallest trades are actually a drag on performance — the strategy would have been more profitable if it had skipped them entirely. This can inform position sizing or signal filtering in future iterations.

Below the size bars, a bordered section displays four streak-related values in a compact 2x2 grid. These overlap with the Behaviour modal's streak analysis but are presented here in the context of trade distribution.

Max Win Streak and Max Loss Streak

The longest consecutive run of wins and losses. In the example, 3 and 9 respectively. Combined with the distribution data above, this tells a complete story: the strategy wins infrequently (44% win rate), loses in long sequences (9-trade max loss streak), but the wins are large enough (concentrated in the top 20%) to more than compensate.

Win/Loss

The raw count of winning and losing trades: 27 wins, 35 losses. This is a quick sanity check against the win rate and confirms the sample size for each category.

Positive Trades

The percentage of trades that closed in profit: 44%. This is the same win rate shown in the Performance Metrics modal, but presented here alongside the distribution data to contextualise it. A 44% win rate with 78% of profit in the top 20% is a very specific kind of strategy — one that depends entirely on asymmetry rather than accuracy.

Why streaks appear here

Streak data is included in the Trade Distribution modal because it completes the picture of how trades contribute to the result. A strategy with 68% of profit in Large trades and a 9-trade losing streak tells you that between those important large trades, you are sitting through extended periods of losses. The distribution says "the big trades matter most." The streaks say "you will wait a long time between them."

The Observed Structure panel interprets the data from the other three panels and produces two text summaries: Distribution Pattern and Implications.

Distribution Pattern

The pattern text is determined by the top-20% concentration metric:

• Above 70%: "Profit outcomes are highly concentrated in a small number of large trades. The majority of trades contribute marginally or negatively." — This is the concentrated-winner pattern. The strategy's edge depends on catching a few big moves.
• Between 50% and 70%: "Profit contribution is moderately concentrated. Returns come from both larger and smaller trades." — A mixed pattern where both outliers and routine trades contribute.
• Below 50%: "Profit is evenly distributed across trades. The strategy does not rely heavily on outlier trades." — The most resilient pattern. No single trade or small group of trades dominates the result.

Implications

Each concentration level produces a corresponding implication:

• Highly concentrated (>70%): "Performance outcomes are driven by capturing large moves rather than frequent wins. Variance in outcomes is expected." — High variance is the natural consequence of concentration. Forward performance will swing widely depending on whether the big trades materialise.
• Moderately concentrated (50-70%): "Performance outcomes depend on both consistent gains and occasional larger wins." — The strategy has diversified sources of return, reducing dependence on outliers.
• Evenly distributed (<50%): "Performance outcomes are driven by consistent execution across many trades." — The strategy's edge is expressed through volume, not magnitude. This is typically the most psychologically comfortable pattern to trade.

Why automated interpretation matters

The pattern and implications text saves you from a common mistake: reading each panel in isolation. A concentrated profit curve, a positively skewed histogram, and 68% of profit in large trades all tell the same story — but you might not connect them without the summary. The Observed Structure panel makes the conclusion explicit: this strategy's result depends on a few big trades, and you should expect high variance going forward.

The Trade Distribution modal is structured as a progression from visual data (charts) to numerical breakdown (size categories) to interpretation (observed structure). Each panel adds context to the others.

A practical reading order

1. Check the Top 20% badge first. If it is above 70%, the remaining panels will confirm a concentrated pattern. If it is below 50%, the strategy is well-distributed. This one number sets the frame for everything else.
2. Look at the Outcome Distribution shape. Is it symmetric, left-heavy, or right-tailed? Check the Mean vs Median badges — if the mean is much higher than the median, the typical trade loses money and the average is pulled up by outliers.
3. Scan the Contribution by Size bars. Does "Large" dominate? If so, the profit contribution curve and histogram are telling a consistent story. If "Medium" and "Small" contribute significantly, the strategy may be more resilient than the curve alone suggests.
4. Read the Observed Structure last. It synthesises everything above. If the pattern says "highly concentrated," cross-reference with the Behaviour modal's streak data to understand how long you will wait between the trades that matter.

How distribution connects to other cards

• Performance Metrics: A high Profit Factor can come from either many small wins or a few large ones. The distribution modal tells you which. A PF of 1.5 from concentrated winners is more fragile than a PF of 1.5 from distributed outcomes.
• Behaviour Analysis: A payoff-driven edge type should pair with a concentrated curve and positive skew. If the Behaviour Profile says "Balanced" but the distribution is highly concentrated, there is a mismatch worth investigating.
• Drawdown Analysis: Concentrated profits mean the drawdown between big winners will be sustained. The drawdown card will show long recovery periods that correspond to the flat stretches visible in the Behaviour modal's trade sequencing chart.
• Monte Carlo (Drawdown Envelope): Concentrated distributions amplify the impact of trade ordering. The Monte Carlo simulation will show a wider drawdown range because shuffling the order of a few dominant trades produces large differences in the equity path.

Common traps

Confusing distribution with quality: A concentrated curve is not inherently bad, and an even distribution is not inherently good. Trend-following strategies are supposed to be concentrated — that is how they work. The danger is when concentration is unexpected or when the concentrated trades are unrepeatable.

Ignoring the median: The mean P/L looks positive, so the strategy "works." But if the median is negative, the typical trade loses money. You will experience those typical trades far more often than the mean suggests. The distribution modal forces you to confront this gap.

Assuming small trades do not matter: If the Small category is net negative, those trades are reducing total profit. This does not mean the strategy is bad — but it does mean there might be a filtering improvement available that eliminates the smallest signals and improves overall efficiency.