Partial profit booking feels like adulthood. You took something. You left something. Whatever happens next, you can narrate it as wisdom. That is why partials are beloved, and why they wreck practice logs.
The problem is not booking a slice. The problem is regret math: scoring the trade by the story you could have had, then mixing two different R outcomes into one mood. If you cannot split the book, you are not practising partials. You are practising alibis.
Why partials feel good on NSE days
A long from a midday pullback on a liquid cash name reaches your first target before 14:00. You sell half. Price then either:
- stalls and comes back through entry — you congratulate the booked slice, or
- runs into the last hour — you resent the half you sold.
Both stories are available every time. The feeling is relief plus optionality. Relief is not a metric. Risk-reward practice asked you to pre-commit a design. Partials explode that design into two (or three) designs unless you write them as separate lines.
Intraday cash has a clock. Delivery does not in the same way. A partial on a swing hold "because I don't want to carry full size overnight" is a gap-risk decision. A partial on a 5-minute scalp "because I always book something" is often just fear of giving back a paper tick. Do not copy the ritual across products.
Practice fills ignore brokerage, taxes and slippage. Two exits cost more than one live. If your first slice is a tiny tick, live friction can eat the "win" you are so proud of in replay. Tag cost-blind on any partial closer than your written minimum.
The Split-Book Ledger (original rule)
Every scale is a child trade with its own R. Parent risk is 1R on full size. Children inherit quantity fractions, not vibes.
Worked example, hypothetical arithmetic, not a result:
- Full size risked ₹2,000 to a stop. That is 1R for the parent.
- Rule: sell half at +1R, remainder targets +2R or stops at breakeven after the first slice.
- Path A — both working: (0.5 \times 1R + 0.5 \times 2R = 1.5R). Not +2R. Logging +2R is a lie.
- Path B — first slice then remainder stopped at breakeven: (0.5 \times 1R + 0.5 \times 0 = 0.5R).
- Path C — stopped before any slice: −1R on full size.
- Path D — first slice, then remainder trailed to −0.4R: (0.5 \times 1R + 0.5 \times (-0.4R) = 0.3R).
The parent line in the journal is the sum of children, never the best child, never the unfilled fantasy target.
If you cannot write Path A–D before the session, you do not have a partial rule. You have a mood that will pick the flattering path after the fact.
Checklist: allowed to use partials this block?
Print this. All boxes must be yes, or you run full-size single exits for the next 20 decisions.
- Fractions are written (example: half / half, or one-third / two-thirds). Not "some".
- First-scale trigger is a price or a candle-close rule, not "when it feels enough".
- Remainder invalidation is written (breakeven, structure, or trail card — pick one).
- Each child will be logged as its own R line plus a parent sum.
- Minimum first-scale distance is written, so you do not book noise.
- Product is named: cash intraday / delivery / other. Overnight rules explicit if not intraday.
- You will not compare this block's win rate to last month's full-size block as if they were the same experiment. See win rate versus reward.
- You accept that expectancy thinking now needs more trades, because you split outcomes.
If any box is no, freeze partials. Full-size exits are easier to review and kinder to a small sample.
Failure modes (name them in the log)
Regret rewrite. You booked half at +1R, remainder stopped at breakeven, then wrote "should have held all to 2R" as if that path were known. Tag regret-rewrite. The missed 2R was not available as information at the decision. Practise tagging the urge, not obeying it in the journal comments.
Winner laundering. Remainder is stopped for −0.3R, first slice was +1R, parent is still positive, and you mark the trade a win in a binary column. Binary win/loss with partials is mostly propaganda. Use parent R.
Secret third slice. You add an unplanned exit because the last hour looks ugly. That is a new rule invented under time pressure. Tag unplanned-scale. Count it as a process fail even if the rupee column likes it.
Target inflation. You tell yourself the plan is "1R then 3R runner" but the 3R level was never a structure. It was a round number that made the story heroic. If the remainder target is not a level you would have written as a full-size target, it is fiction.
Intraday leftover. Cash intraday leftover size after 15:00 that your broker would have squared off — replay may still show candles. Write a hard flatten time in practice so the remainder is an actual decision, not a forgotten child.
Fee amnesia. Two exits, two lots of friction live. Replay will not scold you. You must.
A 20-decision protocol if the checklist passes
Randomised NSE sessions, hidden dates, one setup.
- Pre-commit the Split-Book paths (A–D or your written set).
- Enter only when full-size 1R still makes sense. Partials are not permission to take a sloppy stop "because I will book quick."
- Execute the first scale only if the trigger prints. If it does not, you are in a full-size trade. Do not invent a scale at a random tick.
- Log children before you look at the rest of the day.
- After 20 parents, tally: count of
regret-rewrite, count ofunplanned-scale, distribution of parent R (descriptive for your sample, not a published edge).
You may then compare process to a prior full-size block. You may not declare partials "better" because they reduced how often you felt stupid. Feeling stupid is not expectancy.
Indian specificity without folklore
Opening 9:15–10:00: first-scale targets that sit inside the opening range often fire from noise. If your plan uses partials, consider a skip window or a wider first-scale trigger for that block only — written, not improvised.
Lunch lull: partials become entertainment. If you are scaling because you are bored, that is overtrading wearing a professional hat.
Last-hour square-off pressure is real for many Indian intraday products. A remainder into 15:15 is a different trade than a remainder at 11:30. Session-stamp every child.
Replay Trader can force the candle-by-candle commitment so you cannot peek at whether the runner would have worked. AI review, if you use it after the session, is a second pair of eyes on whether you followed the Split-Book — not a score of whether partials are profitable.
If you scale out, log children. If you cannot, do not scale. Join the waitlist.
Related reading
- Risk-Reward Practice: Designing Trades Before Entry
- Trailing Stops: The Hardest Thing to Practise
- Win Rate vs Reward: The Tradeoff Nobody Practises
- Reviewing Practice Trades
Educational note
Educational practice content only. Not investment advice. No guaranteed returns. Practice ≠ live profitability.