The EMA crossover is the first “system” most Indian retail traders touch. It’s seductive — two lines, clean signals, easy to code. It’s also the strategy that quietly drains thousands of accounts because on its own, EMA crossover is a coin flip.
Here’s how to make it actually work.
What an EMA crossover is
You plot two Exponential Moving Averages — a faster one and a slower one — on the chart.
- When the faster line crosses above the slower → bullish crossover.
- When the faster crosses below the slower → bearish crossover.
The Exponential weighting gives more importance to recent prices than SMA does, so it reacts faster.
Why “EMA(9) crossover EMA(21)” is overrated
The 9/21 pair is the most popular intraday EMA combo on Indian forums. The problem: on a 5-min chart of any Nifty 100 stock, you’ll get 6–10 crossovers per day. Most are noise.
Naked 9/21 crossovers on Indian intraday have a win rate around 35–45% with negative expectancy after costs. That’s not a strategy. That’s a fee generator for your broker.
Periods that actually work
For intraday on NSE:
| Pair | Use case | Notes |
|---|---|---|
| 9 / 21 | Scalping | High signal count, needs strong filters |
| 20 / 50 | Standard intraday | Best balance; default |
| 50 / 200 | Trend confirmation | Almost never crosses intraday; use as filter |
| 8 / 34 | Bank Nifty scalp | Aggressive, intraday only |
For most retail traders, 20 / 50 on 5-min is the sane default.
The five filters that make EMA crossover viable
A naked crossover is noise. With filters, it becomes a workable system.
Filter 1 — Trend regime via higher timeframe. Take the 5-min 20/50 bullish cross only if the 15-min 50-EMA is rising. Filters out countertrend signals immediately.
Filter 2 — RSI sanity. Take the bullish cross only if RSI(14) on the same timeframe is above 50. See how to read RSI for intraday.
Filter 3 — Above VWAP. Long crossovers only when price is above session VWAP. Read the VWAP guide.
Filter 4 — Volume confirmation. Crossover candle volume > 1.3× 20-bar average.
Filter 5 — Time of day. 9:30 – 11:00 only. Crossover signals in the 11:00 – 13:30 chop zone have an essentially negative expectancy.
Five filters, taken together, will cut a noisy 8-signal day to maybe 1–2 high-quality signals.
Entry, stop, target
Entry: close of the candle on which the crossover occurs and all filters pass.
Stop: the most recent swing low (for longs), with a 0.2 × ATR buffer.
Target / trail:
- Take partial profit at 1R.
- Trail the rest with the slower EMA itself (price closes below 50-EMA → exit).
- Or trail with SuperTrend(10, 3) if you prefer. See SuperTrend guide.
EMA crossover vs MACD vs SuperTrend
These are all “trend confirmation” indicators built on moving averages.
- EMA crossover: binary signal at the cross. Lots of signals, lots of noise.
- MACD: the histogram of an EMA difference. Smoother, fewer false signals. See MACD explained.
- SuperTrend: ATR-based trailing line. Best for trailing, fewer signals.
A reasonable rule: MACD for confirmation, EMA crossover for trigger timing, SuperTrend for the trailing stop.
Why “golden cross” / “death cross” matters less on intraday
The “golden cross” (50 crosses above 200) and “death cross” (50 below 200) are daily-chart macro signals. On intraday they are noise — they cross multiple times per week.
If you see a strategy that uses 50/200 EMA crossovers for intraday signals, ignore it. That’s a daily-timeframe concept.
Worked example
Reliance, 5-min chart, 11 March 2026.
- 9:35: 20-EMA crosses above 50-EMA. Filters: 15-min 50-EMA rising ✓, RSI = 57 ✓, above VWAP ✓, volume 1.5× ✓, time 9:35 ✓.
- Entry: ₹2,520 (close of crossover bar).
- Swing low: ₹2,508. Stop: ₹2,505 (with buffer).
- 1R target: ₹2,535. Hit at 10:10. Take half. Trail rest with 50-EMA.
- Exit on 50-EMA close below at 10:50: ₹2,541.
Common EMA mistakes
- Using too many EMAs. Three EMAs is fine; five is decoration.
- Trading every cross. Without filters, you’ll lose to costs alone.
- Picking exotic periods like 7/19. No data supports them; you’re just being different for its own sake.
- Ignoring the higher timeframe. Intraday signals against the 15-min/hourly trend lose.
FAQs
Is EMA better than SMA for intraday? Slightly. EMA reacts to recent prices faster, which matters intraday. The difference is small at 20+ periods.
Should the fast EMA always be 50% of the slow? No rule. 9/21, 20/50, 50/200 are all popular. What matters more is consistency, not the ratio.
Can I automate EMA crossover signals? Yes — Chartink and TradingView both support alerts. Just don’t auto-execute without filters; you’ll bleed.
For the start-to-finish framework, see the intraday for beginners guide.