Reports only mean something once the first three steps are in place: trades journalled, tagged to strategies and checked against rules. Read them before that and you are measuring noise.
This step is about reading a small set of numbers at fixed times and leaving with one decision. Watching P&L tiles during the session is not analysis, and it tends to make the next trade worse.
Read in this order
- Rule adherence and rule breaks. If breaks are high, stop here: the fix is process, not entries.
- Expectancy and profit factor for each strategy tag, after charges.
- Drawdown against the limit you wrote down in advance.
- AI or mentor notes, read as questions to check against your trade notes.
- New ideas to test, and only if the first four look clean.
Jumping straight to new ideas after a red day is how traders end up collecting indicators instead of data.
What to open, and when
| When | What to read | Decision |
|---|---|---|
| During the session | Realised loss against your daily cap | Stop if the cap is reached |
| Friday, 15 min | Adherence, net P&L, worst trade, P&L by day and time | One process change |
| First weekend of the month, 45 min | Expectancy per tag, drawdown, equity curve | Keep, resize or pause each tag |
| Once a quarter | Each live tag across different market conditions | Rewrite or retire tags |

Look for patterns by day and time
For intraday traders, one of the most useful Friday views is P&L split by time of day. Averages across the whole session hide the fact that a setup can work at the open and lose money at lunch.
In that illustrative month the account made +₹23k net, but the two midday windows lost −₹13k between them while the first 105 minutes made +₹29k. One candidate change is no new trades from 11:00 to 14:30. Run it for a month before you make it permanent.
AI as a second opinion
LYSA, the AI in TradeLyser, answers questions about your journal in plain language, such as which strategy made money last month. Treat each answer as something to check: open the trades it refers to and read your notes before you change a rule. It sees what is in the journal, not the reasons you never wrote down.
An illustrative case: a summary points out that Tuesday's Bank Nifty size was double your usual size. The trades show no setup tag, no checklist and a mood note of "revenge". The action is a rule for that situation, not switching the strategy off.
Green weeks and red weeks
Green week, rules broken
Red week, rules kept
Drawdown limits decide when to slow down
Write drawdown limits in rupees before you need them, so the decision is made while you are calm. An illustrative set for a ₹5,00,000 account with a monthly drawdown limit of ₹25,000:
Event weeks such as the Union Budget or election results can push you into the caution band with nothing wrong in a setup. Note the event in the journal, so the monthly pass does not mistake one unusual week for a broken strategy.
Insights mistakes
- Checking P&L every fifteen minutes during the session.
- Trusting an AI summary without opening the trades behind it.
- Changing a strategy after one red week when it has 40 trades of positive history.
- Reading gross P&L on scalping tags. Charges and STT can turn a thin edge into a loss.
- Reading any report before untagged trades are fixed.
Copy the Friday insights note
Week of: ___ Rule breaks: ___ Most common: ___ Best tag by expectancy: ___ Weakest: ___ Drawdown vs limit: ₹___ of ₹___ Day or time that stood out: ___ AI or mentor question to check: ___ What the trades showed: ___ Decision: ___
Questions traders ask
How often should I open reports?
Weekly for adherence, net P&L and patterns by day and time, and monthly for expectancy per strategy and drawdown. During the session, the only number worth watching is realised loss against your daily cap.
Can LYSA AI replace my review?
No. LYSA can answer questions about your journal and point at patterns, but you still open the trades and decide what to change. It only knows what is in the journal, not the reasons you never wrote down.
What is the first metric to check?
Rule breaks for the week. If there were none, move on to expectancy per strategy tag. If there were many, fix the process before you change any entries.