Short answer: momentum trading is buying instruments that are already rising and selling those that are already falling, on the premise that recent relative strength tends to persist for a period before it reverses. It is a continuation approach rather than a value approach, and its results depend far more on exit discipline than on entry selection.
The idea behind the approach
Momentum rests on a simple observation: price moves are not perfectly random over short and medium horizons. Instruments that have outperformed over the past few months have historically tended to keep outperforming for a while, and the reverse holds for laggards.
Explanations differ. Some point to information spreading slowly through a market, so a genuine change in prospects is priced in over weeks rather than instantly. Others point to behaviour, with traders anchoring on old prices and reacting late, then over-reacting once the move is obvious. Institutional flows contribute too, since large positions are built gradually and each tranche of buying pushes price further.
None of these explanations tells you when a move ends, which is the central problem. Momentum works until it stops working, and the transition is usually abrupt.
Momentum versus trend following
The two are close relatives and frequently confused.
Trend following asks whether price is above or below its own past, and holds while that condition persists. It is absolute and applied to a single instrument.
Momentum, in its stricter sense, ranks instruments against each other and holds the strongest. It is relative and requires a universe to compare.
In practice most retail momentum trading blends the two: pick instruments showing the strongest recent moves, then use a trend rule to time entries and exits. Both share the same core weakness, which is a run of false starts during choppy conditions.
Indicators that measure momentum
| Indicator | What it measures | Bounded | Typical use |
|---|---|---|---|
| Rate of change | Percentage move over N periods | No | Raw ranking of strength |
| RSI | Speed of gains against losses | 0 to 100 | Extremes and divergence |
| MACD | Distance between two moving averages | No | Confirmation and turning points |
| ADX | Strength of a directional move | 0 to 100 | Whether a trend exists at all |
| Relative strength versus an index | Performance against a benchmark | No | Selecting between instruments |
The important distinction is between measuring speed and measuring direction. RSI at 75 tells you the recent advance has been fast, not that it will continue. ADX at 35 tells you a directional move exists, not which way. Combining one of each is usually more informative than stacking two of the same type.
Building a momentum strategy
- Define the universe. Decide which instruments you will rank. A currency trader might use the major pairs, an index trader a handful of benchmarks.
- Choose a lookback. Three to twelve months is common for position trading, and hours to days for intraday work. Shorter lookbacks react faster and produce more false signals.
- Rank and select. Take the strongest instruments by your chosen measure, not the ones with the best story.
- Time the entry with structure. Buy pullbacks within the strong instrument rather than chasing extended candles.
- Set the invalidation before entry. A recent swing low, or a volatility multiple, defines where the thesis fails.
- Define an exit rule. Either a trailing rule, a fixed holding period, or a rank rule where you exit when the instrument leaves the top group.
- Rebalance on a schedule. Weekly or monthly for swing horizons. Doing it ad hoc guarantees emotional decisions.
- Record every trade. Momentum strategies live or die on the distribution of outcomes, and you cannot see that distribution without a log.
Entry and exit in practice
The entry problem is chasing. By definition a momentum candidate has already moved, so buying at the extreme means a wide stop or a tight one that gets hit by ordinary noise. The usual solution is to wait for a shallow pullback into a moving average or a prior breakout level, which improves the risk-to-reward ratio without abandoning the premise.
The exit problem is the one that actually determines results. Momentum returns are typically driven by a small number of large winners, so exits that cut those winners early destroy the strategy. A trailing exit that gives the position room, combined with a hard invalidation level, is the standard compromise.
Scheduled events cut across both. Economic releases can end a move regardless of momentum readings, and the calendars are public: the Federal Reserve FOMC schedule and the Bureau of Labor Statistics employment release are the two most consequential for many markets.
Risks and common mistakes
- Crash risk. Momentum strategies tend to lose heavily and quickly at turning points, because they are by construction positioned in whatever has been working.
- Whipsaw in ranges. Sideways markets generate repeated entries and exits, each costing spread and commission charges.
- Chasing extension. Entering after a vertical move leaves no sensible stop placement.
- Confusing volatility with momentum. A wildly swinging instrument is not trending, and volatility filters exist for this reason.
- Over-optimising the lookback. Testing a hundred lookback periods and selecting the best is curve fitting, not research.
- Ignoring correlation. Five strong instruments in the same sector or currency bloc are effectively one position.
- Position sizing by conviction. The strongest ranked instrument is not the one that deserves the biggest position. Volatility should set size.
Leverage compounds every item on that list. FINRA's overview of margin accounts explains the mechanics, and momentum's occasional sharp reversals are exactly the conditions in which leveraged accounts suffer most.
Position sizing and risk control
Momentum's return profile means most trades are small losses and a few are large gains. That distribution only survives if no single loss can end the programme.
Two rules do most of the work. First, risk a fixed small percentage of the account per position, calculated from the stop distance rather than from the position value. Second, cap total exposure across correlated instruments, since a momentum screen naturally selects positions that move together.
A third rule matters for anyone trading under external limits: know the daily and total loss thresholds before you size anything. Traders working a Pivex evaluation operate against a static, non-trailing total drawdown and a separate daily limit, and a momentum approach that ignores those numbers will breach them during the first sharp reversal.
How momentum behaves in different conditions
Sustained directional markets are where the approach earns its reputation. Signals persist, pullbacks are shallow, and the trailing exit stays out of the way.
Range-bound markets are where it bleeds. Each apparent breakout fails, and the strategy takes a series of small losses. This is not a flaw to be fixed with better indicators. It is the cost of the approach, and the correct response is either to reduce activity or to accept the drawdown as part of the distribution.
Transition periods are the most dangerous. A market that has trended strongly and then turns produces the largest single losses, because positioning is maximally aligned with the old direction at exactly the wrong moment.
Momentum across different markets
The premise travels well, but the practical experience varies a great deal by asset class.
Currencies trend in response to interest rate differentials and policy expectations, so momentum tends to appear in multi-week or multi-month legs rather than in sharp bursts. Central bank calendars matter more than anything on the chart, and a pair that has trended for weeks can reverse in minutes on a policy surprise.
Indices produce smoother momentum than individual instruments because they average away company-specific noise. That smoothness is why index momentum strategies tend to show fewer, longer trades. It also means the reversals, when they come, are broad rather than isolated.
Metals blend two drivers: real interest rates and risk sentiment. Momentum in gold in particular can persist for months and then reverse abruptly when the underlying driver changes. Position sizing matters more here than signal quality.
Cryptocurrencies exhibit the strongest and shortest momentum of the four. Moves extend further than most traders expect and unwind faster. Volatility-scaled position sizing is not optional in this market, it is the difference between a survivable strategy and a broken one.
Across all four, the same instruction applies: measure the instrument's own typical range before deciding how much to hold. A single position size applied uniformly across markets with different volatility profiles is a hidden bet on the noisiest one.
Testing a momentum idea before trading it
A momentum rule is easy to describe and easy to fool yourself about. Three checks separate a real edge from a curve-fitted one.
Check the number of trades. A rule that produced eleven trades over five years has told you almost nothing, regardless of how good the equity curve looks. Small samples produce impressive curves by chance.
Check the worst stretch, not the average. The relevant question is how long the strategy went without a new high and how deep the decline was. If that stretch is longer than your patience or larger than your account can absorb, the average return is irrelevant.
Check sensitivity. Change the lookback by a few periods in each direction. If results collapse, the rule was fitted to history rather than describing a real behaviour. A robust rule degrades gently.
Free long-horizon data makes this practical.A long daily history for that kind of test is free on FRED, the St. Louis Fed data service.stlouisfed.org/series/SP500) give enough history to run these checks in a spreadsheet before committing any capital.
Next step: pick five instruments you follow, rank them by three month performance, and track that ranking weekly for a quarter without trading it. Watching how often the leaders change will teach you more about the approach's real turnover and cost than any backtest you download.
Before participating, review Pivex's current program rules and decide whether a simulated proprietary trading evaluation fits your experience and risk limits. Participation does not guarantee a reward, income, or any particular trading result.