Analyze Your Past Trades and Improve Your Trading Results: Using AI to Review Past Trades
Upload your tradebook and see what your trading actually did — win rate, expectancy per trade, drawdown, and where the money went.
💡 How to Read This
- Expectancy is the number that matters most — it is the average rupee outcome per trade. A positive expectancy with a low win rate is a perfectly workable system.
- A win rate on its own says very little. Winning 30% of the time with large winners beats winning 70% of the time with one huge loser.
- Profit factor below 1.0 means the losses outweigh the wins over the period examined.
- Look at the breakdowns for patterns you did not intend — a single symbol or one day of the week often accounts for a disproportionate share of the damage.
⚠️ Read the Numbers Carefully
- A tradebook does not contain brokerage, STT, GST or stamp duty. Unless you enter a charges estimate, every figure here is gross of costs.
- A small sample proves little. Twenty trades can look excellent purely by chance.
- Corporate actions such as splits and bonuses are not adjusted for and can distort individual trades.
- This describes what already happened. It is not a prediction, and a profitable past period does not establish an edge.
🎯 Getting Something Useful Out of This
A record like this is worth having for one reason: it stops you fixing the wrong thing. Five questions it can settle that a profit-and-loss figure cannot.
- Is the problem how often you win, or how much you win? A 35% win rate is perfectly workable if winners are large enough. Compare your average win against your average loss, then against the win rate that pairing would need to break even. Only one of those two numbers is usually the thing to change.
- Is the damage spread out, or concentrated? A loss spread evenly across hundreds of trades points at costs and per-trade edge. The same loss arriving from a handful of trades points at position size and exits. The per-trade column beside each breakdown separates the two.
- Does the record depend on one or two outcomes? Remove the single best trade and see what remains. If the picture changes completely, you are looking at one result rather than a pattern.
- What did the bad stretches actually look like? The longest losing run and the maximum drawdown describe what you had to sit through. Position sizing that felt fine on paper often did not survive the run that actually happened.
- What is missing from the numbers? Charges are absent from a tradebook, and expired contracts carry an assumption until you confirm them. Both can be large enough to change the conclusion, and both are under your control here.
What a record like this cannot do is tell you whether a strategy works. A few hundred trades over a few months is a short stretch, and a period that ends down is no more proof of a broken method than a period that ends up is proof of a sound one. Use it to find what to examine, not to decide what to believe.
📋 Disclaimer
This tool is provided for educational and informational purposes only. It analyses a file you supply and does not constitute financial, investment, trading or tax advice, nor any recommendation to buy, sell or hold. Figures are computed from your own data and may differ from your broker's official statements, which remain the authoritative record for accounting and tax purposes. Trading carries a substantial risk of loss. Consult a qualified adviser before making any trading or investment decision.
📐 How Each Metric Is Calculated
1. How trades are reconstructed
2. Options, futures and expiry
3. Win rate and averages
4. Expectancy
5. Profit factor
6. Maximum drawdown
What this analysis tells you that a P&L figure cannot
Your broker already shows you a profit or loss number. What it does not show is why that number came out the way it did — and the answer usually decides what is worth changing.
Whether the problem is how often you win, or how much you win
A 35% win rate sounds poor and is perfectly workable if the winners are big enough. What matters is the relationship between the two. If your average win is 1.3× your average loss, you need to win about 43% of the time simply to break even. Land at 37% and you lose money steadily while doing nothing obviously wrong. This tool computes that break-even win rate from your own trades and puts your actual rate beside it, so the gap becomes a number instead of a feeling.
Whether the damage is spread out or concentrated
A loss of ₹60,000 arriving as ₹100 across six hundred trades is a costs-and-edge problem. The same ₹60,000 arriving from nine trades is a position-sizing and exit problem. They look identical on a P&L statement and call for opposite responses. Every breakdown here shows the total and the per-trade figure beside it.
How much of the record rests on one or two outcomes
Remove your single best trade and see what remains. If the picture changes completely, you are looking at one result rather than a pattern — a distinction worth making before drawing any conclusion about a method.
What the bad stretches actually looked like
The longest run of consecutive losses and the maximum drawdown describe what you had to sit through. Position sizing that seemed reasonable in advance often did not survive the run that actually happened.
Options and futures are handled properly
Most generic trade analysers break on Indian F&O data. This one decodes contract names, so NIFTY26APR22500PE is read as a NIFTY put option, strike 22500, expiring in April 2026. That makes two things possible:
- Grouping by underlying. A hundred strikes on the same index collapse into one NIFTY line instead of a hundred rows each showing a single trade.
- Catching expired positions. An option that expires has no closing trade in the tradebook — the contract simply ceases to exist. Left alone, the position looks permanently open and its result never reaches your statistics at all. The tool books such positions at settlement, and lets you type the actual settlement price where a contract finished in the money.
That second point is worth dwelling on: it is entirely possible for a completed, settled result worth tens of thousands of rupees to be invisible in your own analysis. It is one of the most common gaps in do-it-yourself trade records.
How it works
Built for Zerodha Console tradebook exports, and works with most other broker CSVs — columns are detected automatically and you can correct the mapping before analysing.
- Export your tradebook. In Zerodha, that is Console → Reports → Tradebook, downloaded as CSV. Other brokers offer an equivalent trade or transaction report.
- Drop the file in. It is read on your own device using the browser File API. No upload, no storage, no account.
- Check the column mapping. Symbol, date, buy/sell, quantity and price are detected automatically. If a column is not recognised, pick it from the dropdown.
- Add a charges estimate. A tradebook contains no brokerage, STT, GST or stamp duty. Entering a flat per-trade figure gives a far more honest result.
- Read the analysis — and download any panel as an image if you want to keep or share it.
What it does not do
This is a description of trades that already happened. It does not predict anything, does not generate buy or sell signals, and cannot tell you whether a strategy will work in future. A few hundred trades over a few months is a short stretch: a period that ends down is no more proof of a broken method than a period that ends up is proof of a sound one. Use it to work out what to examine, not to decide what to believe.
Frequently asked questions
Is my trading data uploaded anywhere?
No. The CSV is read on your own device using the browser File API and analysed inside the page. Nothing is sent to a server, nothing is stored, and no account is needed. Closing the tab clears everything. You can confirm this by disconnecting from the internet after the page loads — the analysis still runs.
Which brokers does it work with?
It is built and tested against Zerodha Console tradebook exports, and works with most other broker CSVs because it detects columns by name rather than by position. If a column is not recognised automatically, you select it from a dropdown before analysing — so an unusual export format is a small inconvenience rather than a blocker.
What is expectancy, and why does it matter more than win rate?
Expectancy is the average rupee result per trade — your net profit or loss divided by the number of completed trades. It matters more than win rate because it combines how often you win with how much you win. A 30% win rate with large winners can produce a positive expectancy, while a 70% win rate with one oversized loser can produce a negative one. Win rate alone tells you almost nothing without the size relationship beside it.
Does it include brokerage and taxes?
Only if you tell it to. A tradebook export contains executions, not charges, so brokerage, STT, GST and stamp duty are absent from the raw data. There is a charges-per-trade box where you can enter a flat estimate, which is then subtracted from every completed trade. On a thin per-trade result this can change the conclusion entirely, so it is worth filling in.
How are option positions that expired handled?
An expired contract has no closing trade, so it would otherwise sit as a permanently open position and never enter your statistics. The tool detects contracts whose expiry has passed and books them as settled — at ₹0 by default, which is exact when the contract expired worthless. Where one finished in the money you can enter the real settlement price. Futures are never closed automatically, because they settle against the spot price rather than at zero.
Why do my figures differ slightly from my broker’s P&L statement?
Several reasons are possible: charges are excluded unless you enter an estimate, corporate actions such as splits and bonuses are not adjusted for, and any expired contract carries a settlement assumption until you replace it with the actual figure. Your broker’s statement remains the authoritative record for accounting and tax. This tool is for understanding patterns, not for filing returns.
Educational and informational content only. EquityTimer is not a SEBI-registered investment adviser and does not provide buy, sell or hold recommendations, trading signals, or portfolio management. This tool analyses a file you supply and describes past results, which are not indicative of future performance. Trading and investing carry a substantial risk of loss. Consult a qualified financial adviser before making any trading or investment decision.