The 100-Trade Review: What Your Trading History Actually Tells You

Most traders can tell you exactly how their last five trades went. Far fewer can tell you what their last hundred actually looked like — which setup carried the month, where the losses clustered, whether execution has been holding up. That gap is the whole reason a 100-trade review is worth doing: it replaces "I think I've been trading fine" with a number you can check.

Here's how to run one, using a single trader's data to show what each step actually reveals.

Start With the Baseline

Take a trader — call her Maya — who pulls her last 100 trades and runs the fundamentals:

• Win rate: 45%
• Average win: 2R
• Average loss: 1R
• Total trades: 100 (45 wins, 55 losses)

On its own, 45% sounds mediocre. It isn't. Because her average winner is twice the size of her average loser, the math works in her favor even losing more often than she wins:

Expectancy = (Win Rate × Average Win) − (Loss Rate × Average Loss) (0.45 × 2R) − (0.55 × 1R) = 0.90R − 0.55R = +0.35R per trade

Across 100 trades, that's roughly +35R — a genuinely solid result that a bare win rate would never suggest. This is the first thing a 100-trade review does: it stops you from judging a strategy by the one stat (win rate) that's most likely to mislead you.

Break It Down by Setup

Maya's 100 trades aren't from one setup — she runs three. Aggregated, they produced +35R. Split apart, the picture changes:

Setup C is bleeding roughly −7.8R across those 25 trades — quietly offsetting a meaningful chunk of what A and B are earning. Blended into one account-level number, that's invisible. Broken out by setup, it's an obvious next question: fix Setup C's entries, trade it less, or drop it.

This is the single most useful thing tagging does. If you trade more than one setup and you're not separating them in your review, you don't actually know which one is working.

Check Whether the Trend Is Moving

A static 100-trade snapshot tells you where things stand. Splitting it in half tells you where things are heading.

Maya compares her first 50 trades to her last 50:

• First 50: 48% win rate, 1.9R avg win, 0.9R avg loss → +0.44R expectancy
• Last 50: 42% win rate, 2.1R avg win, 1.3R avg loss → +0.15R expectancy

Her expectancy is still positive, but it's dropped by roughly two-thirds, and the driver is clear: her average loss grew from 0.9R to 1.3R. That's a specific, answerable question — are stops being moved, is risk sizing drifting, are losing trades being held longer than planned? A single 100-trade average would have shown "+0.35R, all good" and buried a real decline. Splitting the sample in half is what surfaces a trend before it becomes a bigger problem.

Find Out Where the Losses Actually Cluster

Aggregate P&L tells you the total damage. It doesn't tell you the source. Sorting Maya's 100 trades by loss size, her 10 largest losses account for −22R — over 60% of her total losses despite being just 10% of her trades.

Look closer at those 10, and a pattern emerges: 7 of them came from Setup C, and 6 involved entries taken without her stated confirmation signal. That's not "bad luck spread evenly across the month." That's two specific, fixable behaviors accounting for most of the damage. A journal that only tracks outcome, not setup and rule adherence per trade, can't show you this.

Look for the Trades That Repeat a Mistake

Individual errors happen. The value of 100 trades is that repeated ones stop looking like bad luck and start looking like a pattern. In Maya's data:

• 14 of her losing trades were taken within an hour of a prior loss — more than double the rate of her winning trades entered in that window.
• Her average position size on trades taken after 2 consecutive losses was 40% larger than her plan calls for.

Neither of these facts exists in a P&L statement. Both come directly from tagging trades with timing and context, then looking at what her losers have in common that her winners don't.

Separate the Strategy From the Execution

Here's where a lot of "my strategy stopped working" conclusions get corrected. Setup A's rules, followed exactly, have historically produced roughly +0.55R expectancy in Maya's earlier backtesting. Her live results on Setup A this period: +0.375R (from the table above). That's a real gap — but it's not evidence the setup broke. Checking the trades shows late entries and early exits accounting for most of the difference. The fix isn't a new setup. It's tighter execution of the one that already works.

This distinction — strategy problem versus execution problem — is usually invisible without a record of what you actually did on each trade versus what the plan called for.

Build the Review Into a Playbook

The output of a 100-trade review isn't a feeling of "that was useful." It's a short, specific document you actually use. From the numbers above, Maya's next playbook update might read:

• Reduce Setup C frequency or rework its entry criteria — negative expectancy across 25 trades.
• No new entries within 60 minutes of a loss — this window accounts for a disproportionate share of losses.
• Flag any position sized more than 10% above plan for review — sizing drift is showing up specifically after losing streaks.

That's concrete and testable against the next 100 trades, which is the point. A playbook that just says "trade better" isn't one.

Where AI Actually Helps Here

Everything above is doable by hand at 100 trades. It gets considerably harder at 500 or 1,000 — cross-referencing setup, timing, position size, and rule adherence against outcome by hand doesn't scale. This is the realistic use case for AI in a trading journal: not predicting the next trade, but running the cross-tabulation — this setup underperforms after losses, this time window accounts for outsized damage — automatically as the dataset grows, instead of you rebuilding a spreadsheet pivot table every month.

TrackMyTrades.AI is built around that workflow: trades get tagged with setup, condition, and execution notes as you log them, and the platform surfaces the breakdowns — Setup C's negative expectancy, the post-loss sizing drift — that are easy to miss scrolling through a raw trade list but obvious once separated.

What This Won't Tell You

One honest limitation: none of this predicts Maya's next 100 trades. Markets shift, and a setup with +0.4R expectancy over the last quarter isn't guaranteed to hold that number going forward. What the review does give her is a diagnostic — where her process is strong, where it's drifting, and which specific behaviors are worth fixing before the next hundred trades repeat the same pattern.

Final Thoughts

A 100-trade review isn't about staring at a bigger P&L number. It's about splitting that number apart — by setup, by time period, by loss size, by rule adherence — until the things a single aggregate figure hides start showing up on their own. Maya's blended result looked like a solid, stable +0.35R system. Broken down, it was one losing setup, a fading execution trend, and a specific post-loss sizing habit, all sitting inside a number that looked fine from a distance.

That's the actual value of the exercise: not predicting what comes next, but finding out, in specific and fixable terms, what your last hundred trades were really made of.