Trend Following

Trend Following

A trend-following strategy bets that a move already underway will keep going. It does not try to predict where the market will turn -- it waits for the market to start moving in one direction, joins that move, and stays in until the move shows signs of ending. The entry is deliberately late: by the time a trend signal fires, the first part of the move is already over. What the strategy gives up at the start, it aims to recover by holding through the middle, which is where most of a large move's profit lives.

This article covers the idea behind trend following, the three classic ways to build it, where it struggles, and how to test one honestly in a backtest.

The Core Idea: Cut Losses, Let Winners Run

Every trend-following system is built around one asymmetry. The exit rule closes a trade quickly when the expected move fails to develop, and it keeps the trade open for as long as the move continues. Losing trades get cut early; winning trades are given room to grow. The result is a very particular shape of performance: lots of small losses, punctuated by occasional large wins.

Concrete numbers make the shape easier to see. Imagine 10 trades where 7 lose $100 each and 3 win $500 each. The losers cost $700 in total, the winners bring in $1,500, and the strategy nets +$800 -- while being wrong 70% of the time. That is the trend-following trade-off in miniature: the strategy does not need to be right often, it needs its wins to be much larger than its losses.

Three Classic Trend-Following Approaches

Moving Average Crossovers

The most widely known trend signal. Two moving averages with different lookback periods track the same price -- a fast one that responds quickly and a slow one that filters out noise. When the fast average crosses above the slow one, recent prices are rising faster than the longer-term baseline, which is the definition of an emerging uptrend. Cross under, and the uptrend's premise is gone.

In Simula this is two conditions: enter when SMA(50) crosses over SMA(200), exit when it crosses under. The daily-chart version of that pair is famous enough to have a name -- the golden cross (and its bearish mirror, the death cross). The Your First Backtest walkthrough builds exactly this strategy step by step. Swapping the SMAs for EMAs makes both lines respond faster, which moves entries earlier and increases the number of false starts -- the same trade-off in a different place.

Breakout Systems

A breakout system enters when price escapes a range it has been stuck in. The logic: while price oscillates inside a band, buyers and sellers are balanced; when price pushes decisively outside it, one side has taken over, and the move that follows often has fuel behind it. Some of the best-known trend systems in trading history entered when price broke above the highest high of the previous few weeks.

In Simula, breakout logic is expressed through indicators that define the range. Price crossing over the upper Bollinger Band is a volatility breakout -- price has moved more than two standard deviations above its recent average. Price crossing above the Ichimoku cloud signals an escape from the zone the cloud defines as equilibrium. A Parabolic SAR flip marks the same kind of regime change from the trailing-stop side.

Momentum

Momentum strategies skip the geometry of crossings and bands and ask directly: is price higher than it was? A positive ROC (rate of change) means price is up over the lookback window. Price above a long moving average means the same thing in levels rather than percentages. MACD refines the idea by comparing two EMAs and watching the gap between them widen or shrink. These conditions stay true for as long as the state holds, which makes them natural filters as well as entries -- more on that in Combining Indicators.

A useful companion here is ADX, which measures trend strength without caring about direction. Requiring ADX above a threshold before entering filters out the weak, drifting moves that fail most often.

Where Trend Following Struggles

Trend following has one well-defined enemy: a market that goes nowhere. When price oscillates around its average instead of trending, the fast average crosses the slow one again and again -- and each cross is the start of a move that never comes. The strategy enters near local tops (the fast average only catches up after price has risen), exits near local bottoms, and pays spread and commission on every round trip. This is the whipsaw, and it is not a flaw in your settings -- it is the cost of the strategy's core bet.

The practical consequence is an equity curve with a specific personality: long, grinding stretches of small losses while the market ranges, interrupted by strong runs when a trend arrives. The losing stretches produce real drawdowns and they can last months. A trader who has not internalized this shape abandons the system in the middle of a normal losing stretch -- usually right before the trend that would have paid for it.

Exits Decide the Outcome

Two trend systems with identical entries and different exits can produce opposite results, because the exit is where the "let winners run" half of the asymmetry is implemented -- or destroyed.

The opposite crossover is the classic exit: it keeps you in until the trend measurably bends. Its cost is giving back profit between the trend's actual peak and the crossing that confirms it ended. A trailing stop expresses the same idea in price terms -- it follows the position up at a fixed distance and closes it on the first pullback deeper than that distance, which caps how much of the open profit can be surrendered.

Testing a Trend Strategy in Simula

Build the entry and exit as conditions and read the results with the strategy's personality in mind. Expect the win rate to be low and the average holding period to be long -- a trend system that exits within a few candles is not riding anything. Expect drawdowns from whipsaw clusters. The numbers that should decide the verdict are expectancy, profit factor, and max drawdown -- not the win percentage.

Watch the trade count. On a daily chart, a slow crossover pair like 50/200 may trade a handful of times per decade on one instrument, which is far too few trades to draw conclusions from. Testing on a longer date range or a shorter timeframe raises the sample size.

Common Mistakes

Judging the system by its win rate. The design produces frequent small losses on purpose. Evaluate the payoff structure, not the hit rate.

Testing only a market that trended. A backtest window that is one long bull market cannot tell you what the whipsaw cost will be. Include ranging years, or compare results across several windows.

Adding a tight take-profit. It caps the exact trades that pay for everything else.

Abandoning the rules mid-drawdown. If the backtest shows 8-month losing stretches, a live losing stretch of that length is the system working as tested, not failing. Decide before you start whether you can sit through the drawdown the backtest already showed you.

Where to Go Next

Mean Reversion -- the opposite bet: many small wins, occasional large losses.

Combining Indicators -- using a trend filter to gate other signals, and why indicator categories matter.

Your First Backtest -- builds and runs the golden cross strategy from this article, click by click.

Entry & Exit Conditions -- crossover vs. continuous conditions and how signals become trades.

Win Rate & Expectancy -- the math behind profitable low-win-rate systems.

Stop-Loss, Take-Profit & Trailing Stops -- the exit toolkit, including the trailing stops trend systems lean on.