Technical analysis

Trend Trading: How to Identify a Trend and What Following It Looks Like

Trend trading buys rising markets and holds until a rule says the trend is over. We ran the simplest trend rule, a moving average, on three markets: it cut the deepest losses, lost often in sideways years, and made most of its gains on a handful of trades.

· Founder of PatternsWizard
Checked 11 min read
Three ways to read market direction: three panels. Left: a price line rising above a slower moving average that also rises, read as an uptrend. Middle: a zigzag whose highs and lows are each higher than the one before, marked by dots. Right: a price line swinging around a flat average, with a dot at each crossing, where a trend rule keeps changing its answer.
On this page
  1. How to identify a trend
  2. The rule we tested
  3. Results on three markets
  4. A few trades make the result
  5. When the trend is weak: five years of daily signals
  6. What the research says
  7. Managing risk in a trend trade
  8. Frequently asked questions

Trend trading means buying a market that is rising, or selling short one that is falling, and holding the position until a fixed rule says the trend is over. It does not try to buy the low or sell the high: it gives up the start and the end of every move to catch the middle. The price is many small losses when the market goes sideways, paid for by a few large gains when it trends. We ran the simplest trend rule on three markets to show both sides.

How to identify a trend

A trend is a direction that lasts. Two definitions are used, and both can be written as rules rather than judged by eye.

  • Swing points. An uptrend makes higher swing highs and higher swing lows; a downtrend makes lower highs and lower lows. This is the definition used in Dow theory, and the one trendlines are drawn on. It needs a rule for what counts as a swing, for example a high above the 5 highs on each side.
  • A moving average. Price above a long moving average is an uptrend, below it a downtrend. This is cruder, but it gives one answer on every bar and is easy to test. Faber (2007) used a 10-month average, which he chose as the closest monthly match to the 200-day average. Short averages are used the same way: the T-Line is an 8-day exponential average, held while price closes above it.

Sideways markets are where both definitions fail: the swings stop making progress, and price crosses the average back and forth. No rule can tell in advance whether a new crossing starts a trend or is one more back and forth.

Three ways to read market direction

Three ways to read market direction: three panels. Left: a price line rising above a slower moving average that also rises, read as an uptrend. Middle: a zigzag whose highs and lows are each higher than the one before, marked by dots. Right: a price line swinging around a flat average, with a dot at each crossing, where a trend rule keeps changing its answer.
Schematic, not market data.The left panel is the moving average rule tested on this page; the right panel is the sideways market where it pays for its mistakes.

The rule we tested

Long when the close is above its simple moving average, out of the market when it closes below. Each decision is taken at the close and the trade starts or ends at that close. No short selling, no trading costs, no taxes, and no interest earned while out of the market. We applied it to three series:

  • S&P 500 ETF, monthly closes, 10-month average, Faber's rule, from November 1993 to September 2026;
  • Bitcoin, weekly closes, 40-week average (a little over nine months of data per value), from May 2018 to October 2026;
  • S&P 500 ETF, daily closes, 200-day average, from October 2021 to October 2026.

S&P 500 ETF and its 10-month average, monthly closes

SMA(10) on S&P 500 ETF, dividend-adjusted prices, monthly, December 2005 to September 2026. Illustration of how the indicator reads, not a backtest.
  • Price, monthly candles (hollow up, filled down)
  • SMA(10) of the close (solid line)
SMA(10) on S&P 500 ETF, dividend-adjusted prices, monthly, December 2005 to September 2026. Illustration of how the indicator reads, not a backtest.

On the monthly chart, the rule closed its long position at the end of December 2007 at 103.88 (dividend-adjusted price) and did not buy again until the end of June 2009, at 67.76: it sat out the 2008 crash. It also shows the cost. In 2022 it bought at the end of March, sold at the end of April 8.78% lower, bought again in November and sold in December 5.76% lower.

Results on three markets

Total return: trend rule against buy and hold

Total return: trend rule against buy and hold. S&P 500 ETF, monthly, Nov 1993 to Sep 2026: Trend rule (long above the average) +1,890.6%, Buy and hold +2,852.4%; Bitcoin, weekly, May 2018 to Oct 2026: Trend rule (long above the average) +1,661.8%, Buy and hold +992.2%; S&P 500 ETF, daily, Oct 2021 to Oct 2026: Trend rule (long above the average) +60.0%, Buy and hold +84.6%. Our computation; the inputs and the method are in the text.
  • Trend rule (long above the average)
  • Buy and hold
Our computation; the inputs and the method are in the text.10-month, 40-week and 200-day simple averages; no costs, taxes or interest on cash; the open position is counted at the last close.

Deepest fall from a peak, on closes

Deepest fall from a peak, on closes. S&P 500 ETF, monthly, Nov 1993 to Sep 2026: Trend rule (long above the average) −23.0%, Buy and hold −50.8%; Bitcoin, weekly, May 2018 to Oct 2026: Trend rule (long above the average) −46.4%, Buy and hold −75.2%; S&P 500 ETF, daily, Oct 2021 to Oct 2026: Trend rule (long above the average) −19.8%, Buy and hold −24.5%. Our computation; the inputs and the method are in the text.
  • Trend rule (long above the average)
  • Buy and hold
Our computation; the inputs and the method are in the text.Same rules and windows as the chart above; the rule's fall is measured on its own equity, which stays flat while out of the market.

On the monthly ETF the rule closed 23 trades, 14 winners averaging +30.29% and 9 losers averaging −4.82%, and was in the market 77.7% of the time. On weekly Bitcoin: 10 trades, 6 winners averaging +87.89% and 4 losers averaging −4.87%, 61.3% of the time in the market. On the daily ETF: 15 trades, 3 winners averaging +17.99% and 12 losers averaging −1.23%, 78.3% of the time in the market. The rule results include the position still open at the end of each series. Largest drop is the deepest fall from a peak, measured on closes. Buy and hold starts at the first close where the average exists.

The rule never had the larger gain on the S&P 500 ETF; what it did was cut the deepest fall in half on the monthly series, from −50.8% to −23.0%. On Bitcoin it beat buying and holding on both counts, with one caveat: 10 trades are too few to call that a result, and most of it came from two trades.

A few trades make the result

Trend trading lives on its outliers. On Bitcoin, the long position from April 27, 2020 to June 21, 2021 gained 290.15% and the one from January 16, 2023 to August 26, 2024 gained 152.34%. Without its three best trades, the rule would have gained 34.9% instead of 1,661.8%. On the monthly S&P 500 ETF, without its three best trades, 272.6% instead of 1,890.6%. Missing one big trend, by skipping a signal after a run of losers, costs more than all the small losses together.

Every trade of the 40-week rule on Bitcoin

Every trade of the 40-week rule on Bitcoin. Apr 2019 to Nov 2019: +31.3%; Jan 2020 to Feb 2020: −8.6%; Apr 2020 to Jun 2021: +290.2%; Jul 2021 to Sep 2021: +8.3%; Sep 2021 to Dec 2021: −3.2%; Dec 2021, one week: −6.9%; Mar 2022, one week: −0.9%; Jan 2023 to Aug 2024: +152.3%; Sep 2024 to Mar 2025: +32.6%; Apr 2025 to Nov 2025: +12.5%; Aug 2026, still open at Oct 19, 2026: +3.1%. Our computation; the inputs and the method are in the text.
Our computation; the inputs and the method are in the text.Weekly closes; each bar is one trade, from the week of the buy to the week of the sell. Two trades out of eleven made most of the rule's +1,661.8%.

Bitcoin and its 40-week average, weekly closes

SMA(40) on Bitcoin, weekly, January 10, 2022 to October 19, 2026. Illustration of how the indicator reads, not a backtest.
  • Price, weekly candles (hollow up, filled down)
  • SMA(40) of the close (solid line)
SMA(40) on Bitcoin, weekly, January 10, 2022 to October 19, 2026. Illustration of how the indicator reads, not a backtest.

The weekly chart shows the shape of a trend trade: the rule bought at 22,707.88 on the week of January 16, 2023 and stayed in through every pullback that held above the average until the week of August 26, 2024, when it sold at 57,301.86. Since the week of August 17, 2026, it has been long again from 77,734; on the last week of the series it was up 3.11%, an open trade whose result is not known yet.

When the trend is weak: five years of daily signals

The daily rule shows the other side. Between October 2021 and March 2023, the S&P 500 ETF crossed its 200-day average again and again and the rule lost on 11 trades in a row, each small. Then three trades made the whole result. Every trade is listed below, with dividend-adjusted prices.

BoughtSold, sessions heldReturn
Oct 18, 2021
418.26
Jan 21, 2022
411.08, 66 sessions
−1.72%
Jan 24, 2022
412.83
Jan 25, 2022
407.79, 1 session
−1.22%
Jan 28, 2022
414.81
Feb 11, 2022
413.41, 10 sessions
−0.34%
Feb 15, 2022
418.70
Feb 17, 2022
410.22, 2 sessions
−2.03%
Mar 18, 2022
418.52
Apr 11, 2022
414.19, 16 sessions
−1.03%
Aug 16, 2022
406.31
Aug 17, 2022
403.43, 1 session
−0.71%
Nov 30, 2022
387.07
Dec 5, 2022
379.39, 3 sessions
−1.98%
Dec 13, 2022
381.65
Dec 14, 2022
379.21, 1 session
−0.64%
Jan 11, 2023
377.25
Jan 18, 2023
373.41, 4 sessions
−1.02%
Jan 20, 2023
377.60
Mar 10, 2023
368.09, 34 sessions
−2.52%
Mar 14, 2023
373.64
Mar 15, 2023
371.30, 1 session
−0.63%
Mar 16, 2023
377.82
Oct 25, 2023
402.69, 154 sessions
+6.58%
Nov 1, 2023
407.61
Mar 10, 2025
549.78, 337 sessions
+34.88%
Mar 24, 2025
564.71
Mar 26, 2025
559.31, 2 sessions
−0.96%
May 12, 2025
573.48
Mar 20, 2026
645.30, 215 sessions
+12.52%
Apr 8, 2026
672.60
Still open, 141 sessions so far+14.82% at the Oct 28, 2026 close

Without its three best trades, counting the open one, the daily rule would have lost 8.2%. Fifteen closed trades on one market are not enough to judge the 200-day rule; they are enough to show what following it feels like.

What the research says

The broadest evidence comes from futures. Moskowitz, Ooi and Pedersen (2012) studied 58 liquid futures contracts, from equity indexes to currencies, commodities and bonds, with data from as early as January 1965 to December 2009. They found that the past 12-month excess return of each contract was a positive predictor of its future return, that the effect lasted about a year and then partly reversed, and that 12-month trend profits were positive for every one of the 58 contracts. Their strongest result comes from a portfolio of trend rules across all asset classes at once, not from any single market.

Faber tested his 10-month rule on the S&P 500 from 1901 to 2012: it compounded at 10.18% a year against 9.32% for buy and hold, with cash earning Treasury bill interest.

Faber's test on the S&P 500, 1901 to 2012

Faber's test on the S&P 500, 1901 to 2012. Average return a year: 10-month timing rule 11.22%, Buy and hold 11.26%; Compounded return a year: 10-month timing rule 10.18%, Buy and hold 9.32%. Values as published by the source below.
  • 10-month timing rule
  • Buy and hold
Values as published by the source below. Source: Faber, A Quantitative Approach to Tactical Asset Allocation, 2013 update.Total return series, cash in 90-day Treasury bills, before taxes, commissions and slippage.

He also reports its cost: the rule underperformed the index in roughly half of all years since 1901, while being invested about 70% of the time and making less than one round trip a year. Faber gives lower volatility as one of the reasons for the gain: the two average yearly returns are almost equal, but buy and hold lost nearly 2 points between its average and its compounded return, and the rule about 1.

Our monthly test, from 1993 and with no interest on cash, gained less than buy and hold but cut the deepest fall in half. For how professionals use trend tools, see do professional traders use technical analysis.

Managing risk in a trend trade

  • Keep each loss small. The rule above loses often; the losses only stay harmless if each one is a small part of the account. Size the position from the distance to the exit, as shown in position size.
  • Let the exit trail. A moving average exit moves with the price instead of staying at the entry level, so a long trend carries the exit up with it. A trailing stop does the same with a fixed distance under the price.
  • Spread the bets. One market can chop around its average for more than a year, as the daily ETF did from late 2021 to March 2023, with eleven small losing trades in a row. The research results above come from many markets traded at once.

Trend continuation patterns, such as the flag pattern, are measured on our bench with the method described in our methodology. More guides are in technical analysis. A two-candle continuation signal, the separating lines candlestick pattern, is measured on the bench the same way.

Every pattern's results on one sheet

The only pattern cheat sheet with tested stats: 152 patterns, what the textbook says and what really happened on 6 markets. Free PDF, sent to your inbox.

The PDF by email, plus our newsletter: new tests and results that change. Unsubscribe anytime. Privacy.

Frequently asked questions

What is a trend trading strategy?

A set of rules that buys when a market is in an uptrend, sells short or stays out when it is in a downtrend, and exits when the trend ends. The simplest: long while the close is above a long moving average, out when it closes below.

Is trend trading profitable?

Not always. On the S&P 500 ETF our 10-month rule gained less than buying and holding over 33 years, but with half the deepest fall; on Bitcoin it gained more, on only 10 closed trades. On the S&P 500 from 1901 to 2012, Faber's 10-month rule compounded at 10.18% a year against 9.32% for buy and hold, with interest on cash, and still trailed the index in roughly half of the years. Across 58 futures, 12-month trend profits were positive in every contract from 1965 to 2009 (Moskowitz, Ooi and Pedersen 2012).

What is the difference between trend trading and trend following?

Both names describe the same idea: holding a position in the direction of the move until a rule says it has ended. Neither has a fixed technical definition; what changes the results is the rule, the timeframe and the number of markets.

What is the best timeframe for trend trading?

There is no best one, only a trade off: longer averages give fewer false signals and later exits. On the same ETF, our monthly rule closed 23 trades in 33 years, the daily 200-day rule 15 trades in five years, 12 of them losers.

Is trend trading good for beginners?

The rule is simple; following it is not. It asks you to accept long strings of small losses, 11 in a row on the daily ETF example, and to stay in a winning trade through pullbacks.

Sources

  1. Faber, M. T. (2007, updated 2013), A Quantitative Approach to Tactical Asset Allocation, The Journal of Wealth Management
  2. Moskowitz, T. J., Ooi, Y. H. and Pedersen, L. H. (2012), Time series momentum, Journal of Financial Economics 104
, Founder of PatternsWizard

Passionate about finance, quantitative trading and data analysis. Amaury built the PatternsWizard test bench, which detects each pattern on full price history and measures what happened next against random entries.

PatternsWizard tests chart and candlestick patterns on full market history, against random entries. How we test patterns.

179 patterns testedRandom-walk self-test passed

Related guides

All technical analysis articles