- How to classify the market environment in a few objective checks
- Why breadth matters more than the index level
- How a regime gate works, and why publishing nothing is sometimes the right answer
Individual setups do not fail randomly. They fail in clusters, because they share a common dependency: the market environment. A breakout strategy that produces 40% winners in a rising market can produce 10% in a falling one — with no change in setup quality.
Classifying the regime
- Index versus its 200-DMAConstructive: Above, and the average risingHostile: Below, and the average falling
- BreadthConstructive: A healthy share of stocks above their own 200-DMAHostile: A narrow market — index held up by a few large names
- New highs vs new lowsConstructive: Expanding new highsHostile: Expanding new lows even as the index holds
- VolatilityConstructive: Stable or fallingHostile: Spiking — correlations go to one and setups stop mattering
Breadth deserves particular emphasis. An index is a weighted average, so a handful of very large companies can hold it up while the majority of stocks are already in stage 4. If you trade individual stocks, you live in the breadth, not in the index — and the two can diverge for months.
Illustrative, but the direction and rough magnitude of the effect is the well-documented part.
- Constructive regime — win rate
- ≈ 42%
- Constructive regime — average win / loss
- +2.9R / −1.0R
- Hostile regime — win rate
- ≈ 17%
- Hostile regime — average win / loss
- +1.6R / −1.0R
Each scanner here runs a market-condition check before it publishes anything. If the gate fails, the screen shows the gate and its reason instead of showing setups.
That is a deliberate product decision and it follows directly from the arithmetic above: a breakout list published into a hostile tape is not a neutral act, it is a list of trades with negative expectancy. The engine publishing nothing is part of the edge, not a failure of the engine.
What changes with regime
- Strong uptrend, broadPosition sizing: Full sizeSetup types: Breakouts, pullbacks, momentumHolding period: Longer — let winners run
- Uptrend, narrowing breadthPosition sizing: ReducedSetup types: Only the strongest relative strengthHolding period: Shorter; tighter trails
- Range / choppyPosition sizing: Half size or lessSetup types: Range edges only; skip breakoutsHolding period: Short
- DowntrendPosition sizing: Minimal or none on the long sideSetup types: Cash is a positionHolding period: N/A
The hardest part of this lesson is that hostile regimes still produce beautiful-looking charts. A textbook base with perfect volume characteristics can form and break out into a falling market — and fail anyway, because a broad decline drags everything with it regardless of individual structure.
The correct response is not to look harder for exceptions. It is to accept that setup quality and regime are separate filters that must both pass. A 9-out-of-10 base in a hostile tape is a 9-out-of-10 base you do not take.
The Nifty 50 fell from about 12,430 in January 2020 to roughly 7,511 by 23–24 March 2020 — approximately 38% in around 45 trading sessions.
During that period, essentially every long setup failed regardless of its individual quality. Correlations converged: strong stocks, weak stocks, defensive sectors and cyclical ones fell together. Bases that had taken months to build broke down in days.
No amount of chart-reading skill would have made those individual setups work. The only decisions that mattered were made at the portfolio level — how much exposure was on, how concentrated it was, and how quickly it was reduced.
- ☐Is the index above a rising 200-day average?
- ☐What share of stocks are above their own 200-day average — is breadth confirming?
- ☐Are new highs expanding, or are new lows expanding beneath a flat index?
- ☐Is volatility stable or spiking?
- ☐Given the answers: full size, reduced size, or nothing?
- ☐If the regime is hostile, am I looking for exceptions? That question is itself the warning.
See it live · Every screen here states its regime gate and shows it when closed — including the reason — rather than publishing setups into an environment where they carry negative expectancy.
See the gates →In one screen
- 1Setups fail in clusters because they share a dependency on the market environment.
- 2Classify with four checks: index versus 200-DMA, breadth, new highs versus lows, and volatility.
- 3If you trade individual stocks you live in the breadth, not in the index.
- 4Regime can flip the same method between positive and negative expectancy.
- 5In a regime break, exposure and size decide the outcome — not chart quality.
Check yourself
Nothing is scored or stored — the explanation after each answer is the point.
The index is making new highs but the number of stocks above their own 200-DMA has been falling for three months. What does this describe?
Why does a scanner publishing nothing during a hostile regime count as a feature?