Methodology
A base is a pause. A stock advances, then stops and trades sideways while buyers and sellers argue about the new price. The pivot is the price at the top of that argument. Clearing it means the sellers who were capping the stock have run out.
Everything on the screen is produced by the method below, computed once a night from split-adjusted daily prices. There is no discretion in it and no analyst opinion — but there are real limits to what the data can support, and those are in the last section rather than buried.
We mark every confirmed swing high — a day no day within five sessions either side traded higher. A swing is only confirmed five days after it happens, so the detector can never see a pivot on the day it forms.
A base must interrupt an advance, not a decline: we require a rise of at least 30% into that high over the prior six months. A stock drifting sideways after falling is not building a base, it is falling more slowly.
The pivot is the highest high of the base. No buffer is added — we would rather report the real level and let you decide how far above it you act.
A base runs from that swing high until price closes above the pivot, or until it has run 65 weeks. Bases shorter than three weeks are ignored (two for a recent listing, which has not had time to build one).
Contractions counts successive pullbacks that each go less deep than the one before. A pullback deeper than its predecessor resets the count — three contractions where the third is the deepest is not a tightening range, it is a widening one.
Coil is the average true range of the base's second half divided by its first. Volume dry-up is the same ratio for volume. Under 1 means the stock is winding up rather than swinging wider.
RS is a weighted return — 40% of the last quarter, 20% each of the three longer windows out to a year — ranked into a percentile from 1 to 99.
It is ranked across the market on that date, not against an index. That is deliberate: a stock is only ranked against the stocks that actually had a full year of history alongside it, which makes the reading honest for the date it describes rather than flattered by hindsight.
ETFs and index funds are excluded — a fund does not form a base in any meaningful sense. Where a company trades on both exchanges we use one series for its whole history, preferring NSE, because switching mid-series would corrupt every measurement that depends on continuity.
These are the limits of the underlying data. They are on this page because a method you cannot audit is not a method.
Our daily price history begins in 2020. A stock that peaked in 2017 and never regained that level will be shown as being at a high, when in truth there are owners above it still waiting to get out. We label the screen High since 2020 rather than blue sky for exactly this reason. For companies listed after 2020 the two mean the same thing.
We divide prices back through every split and bonus we hold a record of, which repairs about three quarters of the discontinuities we can detect. The rest — roughly 120 events across 115 companies since 2021 — remain. An unadjusted split looks like a 50% crash, which can fabricate a base or destroy a real one. Any setup whose window contains an unexplained move of more than 35% is flagged on the card, not silently dropped.
You will not see a hit rate, an expected return or a confidence score beside any individual setup, and that is deliberate. A percentage next to one stock would imply we can price its odds, and we cannot. Historical statistics describe a population of trades under one fixed set of rules; they say nothing about the specific base you are looking at today.
We do run a portfolio backtest of the method as a whole — with the full Indian cost stack, modelled slippage scaled to each stock's own turnover, and position sizes capped against its daily volume. Its assumptions and its limitations will be published here alongside any number it produces, never a number on its own.
What that backtest says so far is worth stating plainly: over 2021 to 2026 the method did not beat simply buying a midcap index fund, and it drew down further getting there. That window is one regime — a long advance in exactly the kind of stock this method trades — and it contains no sustained bear market, which is the only thing a stop-loss is really for. So this page will not lead with a return figure, and neither will the screen. What the tool offers is the base, the level, and the risk defined before you enter — not a promise about what follows.
Two reasonable implementations of this method will disagree about a meaningful share of bases — where one begins, whether a pullback counts as a contraction, whether 34% depth is still a base. Ours applies the rules above consistently, which is a different claim from being right. The measurements are all on the card so you can disagree with the label and still use the numbers.
Setups are computed from daily closes after the market shuts. A pivot crossed at 10:47am appears on this site that evening.
For information only, and not investment advice or a recommendation to buy or sell any security. Chart patterns describe what price has already done; they do not predict what it will do. Please consult a SEBI-registered adviser before investing.
