The stop-loss is the single most important line you draw on a chart. Most retail intraday accounts in India don’t blow up because of bad entries — they blow up because of bad exits.
This guide is a quick taxonomy of the five stop-loss methods that actually hold up on NSE intraday, with the trade-off for each.
Why a fixed % stop is the worst default
“Set a 0.5% stop on every trade” is what most YouTube tutorials teach. It is the laziest possible rule and ignores volatility.
A 0.5% stop on Bank Nifty intraday is a sneeze. A 0.5% stop on a low-beta FMCG name is two hours of price action. Same number, completely different risk.
If you take only one thing from this article: stop-loss distance must scale with volatility, not be a fixed percent.
Method 1 — ATR-based stop (the default that should be the default)
Use Average True Range as your volatility unit.
- 5-min chart, ATR(14).
- Stop distance =
1.5 × ATRfrom entry. - Long trade entry at ₹500, ATR = ₹3 → stop at ₹500 − ₹4.5 = ₹495.5.
This auto-tightens on quiet stocks and loosens on volatile ones. The same rule trades a calm Asian Paints and a hyper Bank Nifty without modification.
Method 2 — Structure-based stop
Place your stop just beyond the last meaningful structure (swing low for longs, swing high for shorts), with a small buffer.
- Identify the last clear pivot on the 5-min chart.
- Stop = pivot − 0.1 × ATR (small buffer to avoid wick hunting).
Pros: respects what the chart is telling you. Cons: subjective. You’ll disagree with yourself about what counts as a swing.
Method 3 — Indicator stop (SuperTrend / EMA20)
Let an indicator carry the stop for you. The cleanest version: use SuperTrend(10,3) as the stop line. It auto-trails.
Variant: stop = below the 20-EMA for longs. Looser, fewer whipsaws, larger drawdowns.
Method 4 — Time-based stop
Often ignored, often the most powerful.
- “If the trade hasn’t moved 1R in my favor within 30 minutes, I exit at break-even.”
The logic: if a thesis is right, it usually works fast. Trades that stagnate tend to fail. Combining a time stop with an ATR stop gives you both a price floor and a patience cap.
This pairs naturally with the best time-of-day windows — you exit before the chop zone, regardless of P&L.
Method 5 — Chandelier exit (trailing)
Designed for trend continuation:
- Stop =
Highest high of last N bars − k × ATR. - Typical: N = 22, k = 3 for swing; N = 10, k = 2.5 for intraday.
Excellent for riding clean trend days. Useless on choppy days — it gives back gains.
Which to use when
| Setup | Best stop |
|---|---|
| Opening-range breakout | Structure (below range low) |
| Momentum continuation (post-9:30) | ATR or SuperTrend |
| Mean-reversion at VWAP | ATR with tighter multiplier |
| Trend day, full ride | Chandelier |
| Scalping (1m on Bank Nifty) | ATR (1.0×) + hard time stop |
The sizing rule that ties it all together
Stop-loss distance is meaningless without position sizing.
- Define max risk per trade as 0.5%–1% of account.
Position size = (Account × Risk%) / Stop distance per share.- Risk on every trade is normalized; only winners scale.
A ₹2,00,000 account, 1% risk, ₹4.5 stop distance → ~444 shares. Not a round number, not a “feel” number — math.
For the wider framework, revisit intraday for beginners.
Things that destroy stop discipline
- Moving the stop “just a little”. Once is forever.
- Mental stops. Brokers don’t enforce hopes.
- No stop at all on “high conviction” trades. Conviction is the leading indicator of a margin call.
- Stops inside the spread. Use limit orders for entry, SL-M for exit.
FAQs
Should I use SL or SL-M for intraday? SL-M (stop-loss market) for exits guarantees execution but not price. SL guarantees price but may not execute on a gap. For intraday, SL-M is usually the right default.
What about cover orders (CO) and bracket orders (BO)? COs hard-wire a stop at entry — useful for discipline. BOs have been restricted on many brokers post-margin changes; check your broker. See intraday margin rules.
Is averaging down a substitute for a stop? No. It’s a stop-loss in reverse, and statistically it ruins more accounts than any other single behavior.