TradingView EMA crossover strategy

Fast and slow EMAs have different jobs

An EMA gives more weight to recent prices. A shorter period responds sooner to price changes; a longer period responds more slowly. Use the same price source when comparing two periods, and keep chart offsets at zero so the displayed lines match the signal timing. A fast EMA above a slow EMA can filter another setup, such as a pullback entry, without requiring a new crossover.

Specify the crossover event and exit rule

For a bullish cross, the fast EMA must be above the slow EMA now and at or below it on the previous bar. The bearish event is the reverse. Evaluate the rule at a confirmed close if intrabar changes are outside the strategy definition. The following long-only baseline is a test specification, not a recommendation or a measured result.

• Calculate both EMAs from closing prices on the same chart timeframe, with the fast period shorter than the slow period.
• When flat, submit one long market entry on a confirmed bullish crossover. Do not add entries merely because the fast EMA remains above the slow EMA.
• On a confirmed bearish crossover, close the long position and remain flat. This baseline has no short entry, separate stop, profit target, or end-of-session exit.
• Set a fixed order quantity before the test and keep it unchanged across parameter comparisons. An opposite-cross exit does not cap the possible loss.
• Test any short reversal, stop, or session exit as a separately defined variant. An opposite-direction strategy.entry order reverses a position by default; it is different from a close-only order.

Match the signal bar to the simulated fill

Under TradingView defaults, a strategy calculates at the bar close, and a market order can fill on the next available tick, typically the next bar open. A signal marker and an entry marker can therefore appear on different bars. Processing orders on the close or enabling intrabar recalculation changes the assumptions; record those settings with each result. A backtest order is simulated and does not itself execute a trade in a brokerage account.

Test a range filter without hiding its tradeoffs

Repeated crosses in a sideways market can close trades at a loss and incur another round of costs. A candidate filter could require the slow EMA to be higher than its value three bars earlier before accepting a bullish entry. That filter can also reject the start of a new trend. Compare the baseline and filtered version on identical data, and allow the bearish exit even when the entry filter fails.

• Keep the filter threshold fixed while comparing EMA periods; optimizing both at once expands the search.
• If a crossover fails the filter, decide whether to skip it entirely or allow a later entry. A later entry is a different rule.
• Include trending and ranging periods in the evaluation instead of selecting only charts where crossovers look clean.

Plan a parameter sensitivity map

A sensitivity map compares neighboring settings under identical assumptions. As an illustrative test grid, put fast periods 8, 10, and 12 on the rows and slow periods 20, 25, and 30 on the columns. These nine combinations are arbitrary candidates, not recommended settings. No backtest for this grid has been run for this page, so it contains no performance colors or claimed winner.

• For every cell, record net profit, maximum drawdown, and closed-trade count using the same symbol, timeframe, dates, sizing, and costs.
• Plot separate heatmaps for net profit and drawdown with labeled units and a shared scale across comparable runs. Leave missing runs blank instead of treating them as zero.
• Inspect neighboring cells around a candidate. One isolated peak is less persuasive than similar results across nearby settings; neither proves future profitability.

Freeze the rule before the holdout test

Reserve a later date range before comparing settings, and leave it out of parameter selection. Use the earlier period to choose a candidate, then freeze the periods, filter, sizing, and exits before evaluating the holdout. If you revise the rule after seeing the holdout, that interval has become development data. Record results across different market conditions and follow with a forward simulation.

• Use standard price bars and document symbol, exchange, session, timeframe, date range, and EMA warm-up history.
• Set commission and slippage assumptions. Document how spread is represented, and avoid counting the same cost twice.
• Repeat the comparison with higher costs. Frequent crossover systems may lose their apparent advantage when execution becomes more expensive.
• Review trade count, average trade after costs, drawdown, and exposure alongside net profit. A sparse sample leaves substantial uncertainty.