"Banks are leading" is a sentence that feels like a decision. It is usually a caption written after the close. In replay, you do not get the caption. You get two (or three) charts that are allowed to disagree. The practice is noticing disagreement before you know which camp "won the week."
Sector rotation, as a live narrative, is a graveyard of after-the-fact slides. Sector rotation, as a practice object, is narrower: relative behaviour as context, never as a buy list, never as a reason to abandon a written setup mid-session.
Nifty 50, a bank index, an IT index, a metal name — illustrations of baskets and members, not recommendations.
What you are allowed to observe
Strip it to prints:
- While the benchmark (you pick one: Nifty 50, or a sector index) makes a higher high, your practice name does not
- While the benchmark sells off, your name holds a prior low
- Both move together — no story, no "rotation"
That is the same family as relative strength practice, aimed at buckets rather than a two-name beauty contest. The extra Indian wrinkle is how often the public conversation is already a sector story. Your job in replay is to see the tapes without that conversation.
You are not forecasting that money "will rotate into X." You are asking: given my setup on this name, does the basket tape argue for skip, smaller size, or unchanged rules?
The protocol (fifteen sessions, one benchmark)
Write the card once. Do not edit it after a green day.
Sandbox
- One liquid cash name or one sector-index chart as the decision chart
- One benchmark chart you will glance at only at written times (not every tick)
- Hidden dates, randomised start, candle-by-candle on the decision chart
- Same setup rule as your current block so rotation does not become a new strategy
If you cannot run two charts, run the decision chart first, pause at the glance times, then check the benchmark's last completed bars only. Peeking the rest of the benchmark session is cheating.
Glance times (example — freeze your own)
- 09:25 IST — after the open has printed something
- 12:00 IST — midday
- 14:45 IST — only to review, not to enter, if you are cash-style
At each glance, log one of: RS-hold | RS-lag | RS-together | RS-unclear. No essay.
Decision rule (pick one, freeze it)
Version C1 — context only: RS never changes entry. It only tags the log for review.
Version C2 — skip veto: if RS-lag at 09:25 and your setup is a long breakout, skip. If RS-hold and your setup is a short breakdown, skip. Opposite of your direction is the veto. Neutral (RS-together) is not a veto.
Do not run C2 until C1 has ten sessions. Otherwise you will "see" lag whenever you wanted a skip.
Forbidden
- Building a rotation system ("buy the strong sector on Mondays")
- Switching names because the other bucket "looks stronger"
- Using Bank Nifty leadership as a Nifty directional call (index charts without tip culture)
- Sizing up because RS-hold "confirms"
What the chart is really teaching
A sector can look like it is "rotating" because one heavyweight reported, because the index is pin-heavy on expiry, or because your name is illiquid and the sector index is not. Relative prints mix those causes. Practice does not unmix them. It only stops you treating the mix as a signal.
Index versus single-stock practice already split basket tape from name tape (index vs stock). Rotation is the same split, one layer up: basket versus neighbouring basket. If you cannot yet keep index-week and stock-week ledgers separate, you are not ready for a rotation veto. Stay on C1.
Randomised historical starts matter here. Favourite periods — the weeks "everyone knows" IT led or banks led — are slideshows. Randomising the start is how you meet disagreement without the magazine cover.
Review after fifteen tagged sessions
Sit with the log, not with a heatmap.
- How often was
RS-unclear? If that is most sessions, your glance is too early or the pair is too correlated. Change the pair next block, not mid-block. - Under C1, did
RS-lagcluster with losing trades and with winning trades? If both, it was not a filter. It was a mood. - Under C2, did the veto fire on sessions you still chase-entered? Process failure, not a rotation failure.
- Did hidden dates stop you from saying "this was the PSU week"? Good. If you still said it, you peeked.
Fifteen sessions is a habit test. It is not evidence that sector context improves expectancy. Anyone who tells you it does from a replay blog is selling a slide.
Limits
You will not measure "rotation" as fund flows. You will measure two pictures. NSE sector indices change constituents; your historical chart is whatever series you are given. Do not pretend you reconstructed the live index as it was.
Do not practise rotation on a zoo of illiquid members. The strong-looking small-cap in a hot sector is often a liquidity problem wearing a narrative.
Replay Trader can give you the charts and hide the dates. It will not overlay a "sector leadership" badge, which is a mercy. AI review may note that your name diverged from the benchmark; treat that as a prompt to check your own tags, not as a rotation call, and not as an accuracy claim.
A ₹999 thirty-day block is enough to finish C1 and, if process scores hold, try C2. Ten decisions a day × 30 ≈ 300 includes a lot of ordinary setups. Rotation context should remain a minority column in that log — a comment on the decision, not the decision.
Glance, tag, leave the setup rule alone unless you froze a veto in writing. Join the waitlist.
Related reading
- Relative Strength Practice Without Stock Tips
- Practising on Index Charts vs Individual Stocks
- Practising Index Ideas Without Turning Into Tip Culture
- Why Hidden Dates Make Trading Practice More Honest
Educational note
Educational practice content only. Not investment advice. No guaranteed returns. Practice ≠ live profitability.