Trading Strategies
Momentum Trading: Buying Strength and Managing the Reversal
Buying what has already gone up is well documented and deeply uncomfortable. It also fails all at once, which is a sizing problem before it is an analysis problem.
Everything in your instinct says buy the thing that has fallen. Momentum says buy the thing that has already run. Hold it while it keeps running. Leave when it stops. It is among the most studied patterns in market history and among the hardest to act on, and the reason it is hard is worth understanding before you try: you will be buying at prices that felt too high a month ago and selling names you still like.
What the ranking measures
Momentum is a ranking on past returns measured over months. It says nothing about value, earnings quality, or how big the next move will be, and reading it as a forecast of a specific price is already a misuse of it.
Two versions are in circulation and they answer different questions.
Time series momentum compares a security to its own history, so above its level of twelve months ago, or above a long moving average, the signal is positive and nothing else in the market enters the calculation.
Cross sectional momentum ranks a universe against itself and buys the strongest portion. A stock down 5 percent on the year is a buy when the median name is down 20.
They agree in normal conditions. They split when everything falls together, because cross sectional momentum will hold the best of a bad list while time series momentum goes to cash. Much of the risk in this method lives inside that difference. The portfolio version of the argument is in factor investing explained.
Relative strength, and the high
The working version most traders use is relative strength: return over three, six or twelve months against a benchmark or a peer group. A common refinement skips the most recent month, because very short horizon returns tend to reverse. That is the whole subject of mean reversion trading.
Proximity to the 52 week high carries the same information in a different form. A stock within a few percent of its high has almost no overhead supply, because nearly every holder is in profit and nobody is waiting to get out at breakeven. Research into this measure has documented a tendency for such stocks to keep outperforming over subsequent months, on average and across large samples, which is a different claim from any one of your trades working.
| Signal | How it is measured | What it misses |
|---|---|---|
| 12 month return skipping the last month | Total return over the window | Ignores why the stock moved |
| Relative strength versus a benchmark | Stock return minus index return | Flatters the least bad name in a falling market |
| Distance from the 52 week high | Percent below the high | Treats a one day gap and a year long grind alike |
| Price above a long moving average | A yes or no filter | Slow to turn, whipsaws in ranges |
The screen comes first
This is a screening strategy before it is a trading strategy. The ranking is the entire idea. A workable screen carries four filters: a return ranking over six or twelve months, a liquidity floor so the spread does not eat the edge, a price filter that keeps the thinnest names out, and a trend filter so you are not buying something that already broke down from a high base.
The stock screener runs those together. What comes out is a list to work through. Ranking produces candidates. The entry still needs a level.
The exit is the strategy
Entry after a momentum screen is usually a pullback into rising structure or a break to a new high on volume. Either way you need a price at which the strength has ended. That price sets the size.
A $20,000 account risking 1 percent has $20,000 x 0.01 = $200 to spend. The stock is $50 and the most recent higher low sits at $48, so the distance is $2 and the size is $200 / $2 = 100 shares, a $5,000 position.
The winning side is where this method is made or lost. The return profile is lopsided. Many small losses, a few large gains. The large gains have to be large.
expectancy = (win rate x average win) - (loss rate x average loss)
Start with a 35 percent win rate. Average win $600 and average loss $200. Then (0.35 x $600) - (0.65 x $200) = $210 - $130 = $80 a trade. Now take profits sooner so the average win is $350: (0.35 x $350) - (0.65 x $200) = $122.50 - $130 = -$7.50 a trade. The win rate did not move. The losses did not move. The method is now negative.
The crash
Momentum does not bleed out slowly. It reverses sharply when a falling market turns, and both ends of the ranking move against you in the same sessions. The beaten down names rebound hardest. The previous winners stall. A full book of strong names takes correlated losses everywhere at once. A momentum book is one bet written out several times.
| Peak to trough loss | Gain needed to recover |
|---|---|
| 10% | 11.1% |
| 20% | 25.0% |
| 30% | 42.9% |
| 40% | 66.7% |
| 50% | 100.0% |
A 30 percent drawdown needs a 42.9 percent gain, which you can check: 0.70 x 1.429 = 1.00. Past 30 percent the curve steepens into territory most methods do not produce in a good year, so run your own figures through the drawdown recovery calculator before you decide how many correlated positions you can carry.
Sizing one bet written five times
Cap total portfolio heat, which is the sum of risk across open positions. Cap it below what a diversified method would tolerate. Cap single sector exposure too, because momentum screens fill up with whichever group is leading and a leading group is one headline from being last month’s leader.
Two habits follow from that. Cut size when the ranking is narrow. Narrow leadership reverses hardest. And judge the method on return relative to its volatility, which is what the Sharpe ratio calculator computes. Twenty percent earned with 35 percent volatility is a different animal from 15 percent earned with 12 percent, and only one of them is something you will still be running after a bad quarter.
The full sizing framework sits in risk management for traders. The risk management and position sizing quiz tests it under time pressure.
Turnover is a tax on the signal
Momentum turns over by construction. A monthly ranking replaces a meaningful slice of the book, and every replacement crosses the spread twice.
Two hundred round trips a year on 100 share positions in a stock quoted two cents wide costs 200 x $4 = $800, and $1 a side in commission adds 200 x 2 x $1 = $400. On a $20,000 account that $1,200 is 6 percent a year, taken out of the expectancy before you see a dollar of it, which is why the $80 expectancy above grosses $16,000 across 200 trades and nets $14,800. At the $7.50 loss, the costs simply deepen a hole.
Slower ranking windows and fewer, larger positions cut that bill and slow the signal down. There is no setting that avoids the trade off, and anyone selling you a momentum system that has never mentioned turnover has not run it against a commission schedule.
Push toward a longer hold and swing trading covers similar entries on a calmer clock. The account, plan and journal groundwork is in how to start trading stocks.
Frequently asked questions
What is momentum trading?
Buying securities that have risen relative to their peers or relative to their own past, on the observation that recent relative performance has tended to persist over a horizon of months. You hold while the strength continues and leave when it breaks. It takes the opposite stance to buying whatever has fallen the furthest.
What is the difference between time series and cross sectional momentum?
Time series momentum compares a security to its own history, so you own it when it is above where it stood twelve months ago regardless of what anything else did. Cross sectional momentum ranks a universe and buys the strongest part of the list. They usually agree. They disagree most sharply in a market where everything is falling together.
Does buying stocks at 52 week highs work?
It is one of the standard ways of expressing momentum, and research has documented a tendency for stocks near their 52 week highs to keep outperforming over the following months. That tendency is an average across a great many names and periods. Any single position can still fail badly, and the effect is uneven across market environments.
What is a momentum crash?
The sharp reversal that arrives when a falling market turns. The names that had been weakest rebound hardest, the names that had been strongest stall or fall, and a book positioned for continuation loses at both ends in the same few sessions. These turns are fast, which is why a momentum trader sizes for a bad week rather than an average one.
How do you set a stop on a momentum trade?
Use a price that says the strength has ended, such as a close below the most recent higher low or below a moving average the stock respected through the advance. Size the position so the distance from entry to that price equals your fixed dollar risk. A stop set at a round percentage instead of a structural level tends to sit exactly where ordinary pullbacks reach.
Is momentum trading the same as chasing?
They look identical from outside and differ in one respect. A momentum trade has an invalidation price and a size derived from it, so a failed entry costs a known amount. Chasing means buying after a large move with no exit decided, which leaves the entire decision to be made later with money already at risk.