Backtesting
Analyzing Your Backtest Results Like a Professional
Choose the problem that reflects your current situation.
Analyzing Your Backtest Results Like a Professional
1. Interpreting Win Rate, Risk-Reward, and Expectancy
👉 My win rate looks strong, but the numbers don’t prove I have an edge
The Reality Check
Updated 2026
A high win rate can still be a weak system.
Traders often celebrate how often they are right and ignore whether the average win pays for the average loss. Totals look encouraging until you ask what one typical winner and one typical loser actually do to the account.
âť“ The Painful Question Traders Ask
“My win rate looks strong — so why don’t the numbers prove I have an edge?”
The Core Insight
Updated 2026
Win rate, risk-reward, and expectancy are one story, not three trophies.
Win rate is how often you are right. Risk-reward is how much you make when you are right versus how much you lose when you are wrong. Expectancy is what that combination is worth per trade. Read them together or you will trust a headline that hides a fragile edge.
Related Reflection Questions
- If I removed my largest winner, would the win rate still look useful?
- Do my losers cost more than my winners pay, even if I am right more often?
- Can I state my expectancy in R, or only as a feeling that the spreadsheet is “good”?
- Am I using win rate to feel confident, or to understand the system?
⚠️ The Brutal Consequences of Avoiding This
- You keep a high-win method that slowly drains the account
- One large loss wipes a week of small wins
- You cannot compare two strategies because you only tracked accuracy
- Overfitting hides behind a pretty win rate
- Live trading shocks you because expectancy was never calculated
âś… The Deep Solution
Continue to the Full Lesson
2. Measuring Maximum Drawdown and Recovery Periods
👉 The backtest made money, but I am afraid this system will wipe me out
The Reality Check
Updated 2026
A profitable line can still be untradeable.
If the deepest fall is larger than you can sit through, you will abandon the system at the worst moment. Drawdown is not a footnote under net profit. It is the cost of using the method in real time.
âť“ The Painful Question Traders Ask
“The backtest made money — so why am I afraid this system will wipe me out?”
The Core Insight
Updated 2026
Maximum drawdown shows the worst peak-to-trough damage in the sample. Recovery period shows how long it took to get back to a previous high.
Together they tell you the pain and the wait. A system you cannot survive psychologically is not an edge you can use, even if the ending balance looks fine.
Related Reflection Questions
- What is the largest drop from a peak in this test, in percent and in R?
- How many days or trades did the longest recovery take?
- Would I still follow the rules after that drop in live trading?
- Is my live size small enough that this drawdown would not change my behaviour?
⚠️ The Brutal Consequences of Avoiding This
- You size as if the equity curve is smooth
- You quit during a normal drawdown and never see the recovery
- You confuse a survivable system with a comfortable one
- Prop or personal limits get hit because the test pain was ignored
- You keep switching methods every time the account falls
âś… The Deep Solution
Continue to the Full Lesson
3. Identifying Your Strategy’s Strengths and Weaknesses
👉 I have a full backtest, but I still don’t know what to keep and what to drop
The Reality Check
Updated 2026
An average result can hide a sharp edge and a leaking hole in the same spreadsheet.
If you only look at the total, you will keep the weak setups that drag the strong ones down. Professionals do not ask only whether the test made money. They ask which parts made money and which parts should be cut.
âť“ The Painful Question Traders Ask
“I have a full backtest — but I still don’t know what to keep and what to drop.”
The Core Insight
Updated 2026
Strengths and weaknesses live in segments, not in the headline.
Compare setups, sessions, market types, and rule-follows versus rule-breaks. The goal is not to defend the whole method. It is to see which pieces have a repeatable role and which pieces only survived because they were averaged in.
Related Reflection Questions
- Which setup type produced most of the profit — and which produced most of the pain?
- Do my best trades share a condition my worst trades lack?
- Am I judging the strategy by trades I actually followed, or by a mixed bag of mistakes?
- If I dropped the weakest segment, would the remaining system still have enough trades?
⚠️ The Brutal Consequences of Avoiding This
- You keep trading the setups that quietly destroy expectancy
- You abandon a strong core because the average looks mediocre
- You cannot improve because you do not know what to improve
- You add filters that fix the wrong problem
- Live results surprise you because the weakness was never named
âś… The Deep Solution
Continue to the Full Lesson
4. Segmenting Results by Market Condition, Timeframe, and Setup
👉 I have hundreds of trades, but I can’t tell which conditions my strategy is actually built for
The Reality Check
Updated 2026
One blended number can make a fragile mix look professional.
A setup that works in trend and fails in range will still print a “decent” total if you never split the sample. Segmentation is how you stop averaging yourself into a false edge.
âť“ The Painful Question Traders Ask
“I have hundreds of trades — so why can’t I tell which conditions my strategy is actually built for?”
The Core Insight
Updated 2026
Condition, timeframe, and setup are three different cuts of the same data.
A method can be strong on one timeframe and weak on another, useful in one market type and expensive in another, or valid for one setup and noise for the rest. Until you segment, you are judging a mixture, not a process.
Related Reflection Questions
- Does this strategy earn more in trend, range, or news-heavy conditions?
- Which timeframe produced the cleanest expectancy — and which only added noise?
- If I kept only one setup, would the remaining sample still be tradeable?
- Am I mixing live timeframes in one bucket because it is easier than tagging them?
⚠️ The Brutal Consequences of Avoiding This
- You take the method into the wrong market regime
- You keep a dead timeframe because the blend still looks okay
- You cannot write a simple “when I trade / when I stand down” rule
- Optimization hides inside mixed samples
- Live results diverge because the live mix is not the test mix
âś… The Deep Solution
Continue to the Full Lesson
5. Finding Patterns in Missed Trades and Mistakes
👉 I reviewed my results, but the same misses and mistakes keep showing up
The Reality Check
Updated 2026
The spreadsheet of taken trades is only half the record.
Missed valid setups and repeated execution errors change the real expectancy. If you only review what you clicked, you will keep the same skips and the same rule breaks and call the method inconsistent.
âť“ The Painful Question Traders Ask
“I reviewed my results — so why do the same misses and mistakes keep showing up?”
The Core Insight
Updated 2026
Missed trades and mistakes are patterns, not one-off bad days.
A skip after a loss, a chase after a miss, or a size change after a win is a process leak. Tag those events the same way you tag setups. The method cannot be judged until you know which results came from following it and which came from breaking it.
Related Reflection Questions
- Which valid setups did I skip, and what did those missed trades do after I passed?
- Do my errors cluster after losses, after wins, or at a certain time of day?
- If I removed rule-breaking trades, would expectancy rise or fall?
- Is “I missed it” a rare event, or a weekly habit I have not named?
⚠️ The Brutal Consequences of Avoiding This
- You blame the strategy for errors that belong to execution
- You keep skipping A+ trades after a losing morning
- You cannot tell whether the edge is real or luck mixed with mistakes
- Coaching and journaling stay vague because the pattern was never tagged
- Live results stay noisy even after “more backtesting”
âś… The Deep Solution
Continue to the Full Lesson
6. Creating a Trade Tagging System to Group Outcomes
👉 I write notes on every trade, but I can’t group my results in a way I can actually use
The Reality Check
Updated 2026
Without tags, a journal is a list of events, not a research tool.
You cannot group what you never named. Professionals do not wait until they “feel” a pattern. They tag each trade so later they can sort, count, and decide.
âť“ The Painful Question Traders Ask
“I write notes on every trade — so why can’t I group my results in a way I can actually use?”
The Core Insight
Updated 2026
A tagging system is a short, repeatable set of labels — not a novel after each trade.
Use the same fields every time: setup, condition, timeframe, execution quality, and outcome group. If the labels change every week, you cannot compare months. Simple tags you will fill beat a perfect system you abandon.
Related Reflection Questions
- Can I tag a trade in under a minute, or is the system too heavy to keep?
- Do my tags describe the setup and the behaviour, or only whether I made money?
- If I filtered by one tag, would I get a clean group or a mix of unrelated trades?
- Which three tags would change my next-week rules if I actually counted them?
⚠️ The Brutal Consequences of Avoiding This
- Reviews stay anecdotal (“I think I do worse on Fridays”)
- You cannot prove which setup to drop
- Coaching has no shared language
- You rebuild the journal every few months and lose history
- Strategy changes are based on memory, not counts
âś… The Deep Solution
Continue to the Full Lesson
7. Visualizing Results: Equity Curves, Heatmaps, and Stats
👉 I have the stats, but I still do not see how this strategy actually behaves
The Reality Check
Updated 2026
A table of totals can hide a violent path.
If you never look at the equity curve, heatmap, or grouped stats, you will approve a method because the ending number is green. Pictures of the path show clustering, dead zones, and luck you would miss in a single average.
âť“ The Painful Question Traders Ask
“I have the stats — so why do I still not see how this strategy actually behaves?”
The Core Insight
Updated 2026
Visuals are not decoration. They are another way to read the same trades.
An equity curve shows sequence and drawdown. A heatmap shows where results cluster by time, setup, or condition. Core stats keep the picture honest. Use all three before you change a rule.
Related Reflection Questions
- Does the equity curve rise in a few spikes, or in a repeatable slope?
- Where does the heatmap go quiet or turn red?
- Would I still trade this if I had to sit through the ugly middle of the curve?
- Which stat am I ignoring because the picture looks exciting?
⚠️ The Brutal Consequences of Avoiding This
- You size for a smooth path that never existed
- You miss that profit came from one cluster of trades
- You keep hours or days that the heatmap says are dead
- You confuse a lucky spike with a usable edge
- Live trading feels nothing like the spreadsheet ending
âś… The Deep Solution
Continue to the Full Lesson
8. Detecting Red Flags: Are You Fooling Yourself?
👉 The backtest looks clean, but I still suspect I am fooling myself
The Reality Check
Updated 2026
A backtest can be a story you wanted to believe.
Skipped losers, curve-fit filters, tiny samples, and “I would have taken that” thinking all inflate the result. If you do not hunt for red flags, you will take a flattering sheet live and call the market unfair.
âť“ The Painful Question Traders Ask
“The backtest looks clean — so why do I still suspect I am fooling myself?”
The Core Insight
Updated 2026
Red flags are process errors, not bad luck in the sample.
Look for too few trades, too many rules added after seeing the result, missing costs, discretionary overrides, and profit that vanishes when you remove the best trades. Honesty in the test is the edge. A prettier sheet is not.
Related Reflection Questions
- Did I add a filter after I already saw it would help the numbers?
- If I remove the best 5% of trades, does the edge disappear?
- Did I include spread, slippage, and skipped fills?
- Would a stranger with only my rules get a similar result?
⚠️ The Brutal Consequences of Avoiding This
- You fund a curve-fit method and lose quickly live
- You keep adding rules until the past looks perfect
- You cannot tell skill from a lucky cluster
- You waste months defending a sample that was never honest
- Confidence collapses when live results fail to match the sheet
âś… The Deep Solution
Continue to the Full Lesson
9. Turning Data Into Actionable Strategy Adjustments
👉 I have the numbers, but I can’t turn them into a change I will actually follow
The Reality Check
Updated 2026
Data that does not change a rule is a report, not an analysis.
Traders collect stats, then keep trading the same mix. A professional adjustment is small, written, and retested. If the change cannot be stated in one sentence, it is not ready.
âť“ The Painful Question Traders Ask
“I have the numbers — so why can’t I turn them into a change I will actually follow?”
The Core Insight
Updated 2026
Actionable means one frozen change, then a new sample.
Do not rewrite the whole method from one review. Pick the highest-cost leak, write the new rule, and test only that. More than one change at a time makes the next backtest unreadable.
Related Reflection Questions
- What is the single leak that cost the most expectancy in this sample?
- Can I write the new rule so a stranger could follow it?
- Am I changing entries because I am uncomfortable, or because the segment data says so?
- What will I measure on the next sample to know if this change helped?
⚠️ The Brutal Consequences of Avoiding This
- You keep reviewing and never decide
- You change five things and cannot tell which one mattered
- You overfit the last month of pain
- You abandon a valid core because the adjustment process is chaotic
- Live rules drift until nothing matches the test
âś… The Deep Solution
Continue to the Full Lesson
Continue Learning
Next Module: Stress-Testing Your Strategy Across Market Conditions →