Day Trading vs Trend Following

Michael Whitman · Trading since 2017, full-time since 2020 · Published 2026-09-08 · Updated 2026-10-04 · Reviewed 2026-10-04

These get filed as competing styles, and they are not. One decides when you get out; the other decides what you get into. The useful question is whether the pair fits together, and the answer is in how long trends last on a real chart.

What each one is

Day trading is a holding period. Every position closes before the bell, whatever generated the entry. Day trading covers it.

Trend following is a logic. It buys strength, holds while the strength lasts, and accepts many small losses for a few long runs. Trend following covers it.

So you choose one of each. Every method has a clock and a logic; the comparison is about whether this particular clock leaves the logic room to work.

Where they differ

What each decides. The clock decides when you are out regardless of the chart. The logic decides what counts as a reason to be in, and says nothing about the time of day.

What each needs. Day trading needs your presence for the whole session. Trend following needs runs that last long enough to outweigh the false starts, which depends on the market, not on your schedule.

Where the payoff sits. For a day trader, inside one session. For a trend follower, in a handful of runs that last for months, as the SPY count below shows.

How each is judged. A clock is neither right nor wrong. A trend logic is expected to be wrong often, whereas the few long runs pay for the rest.

Which one to use: Day Trading or Trend Following?

Use trend following when your holding period can last months. On SPY the runs that carried the gains were longer than 60 sessions; a swing or position horizon can hold them, while a session clock cannot.

Use day trading with a different logic. Inside one session, methods built on reversion or on the first move of the day fit the time available better than one built on long runs. Mean reversion is the usual alternative.

Use neither combination when you cannot change the clock and the logic needs runs longer than the clock allows. The result will reflect the mismatch, not the entries.

And when you test a trend rule, test it on the holding period you will actually use. Results from a daily chart held for months say nothing about the same entries closed at 4:00 PM.

How long the runs lasted on SPY

The test. Every SPY close from 1993 to 2026 was compared with the average of the last 100 closes. A stretch begins with a close above the average and ends at the first close back at or below it. There were 195 stretches; 194 have ended, and one has been running since 9 April 2026.

Most were short. The median stretch lasted 6 sessions. 133 of the 194 finished (68.56%) were over within 20 sessions. Only 33 (17.01%) lasted more than 60 sessions, and the longest ran 313 sessions, from 13 January 1995 to 10 April 1996, with a rise of 34.80% from the first close above to the close that ended it.

The gains sat in the long ones. Chained together, the 33 stretches longer than 60 sessions came to +1,940.73%; all 194 together came to +292.07%. The short stretches gave most of that back, which is what a trend follower’s record is supposed to look like.

Why a session clock breaks the logic

Because it closes the long runs on their first evening. A 313-session stretch is 313 separate day trades under a session clock, each started fresh and each needing its own reason to enter.

And because the short stretches are not cut. A false start that lasts two sessions hurts a day trader as much as a trend follower. The clock removes the part of the distribution that pays and leaves the part that costs.

The logic that suits a session is one whose payoff arrives within hours, which is not the logic measured here.

The original data

The sample. SPY’s daily closing prices from Yahoo Finance, without dividends, from 29 January 1993 through 2 October 2026. The 100-session simple average exists from the 100th close; closes sat above it on 6,008 of the 8,378 sessions after that (71.71%).

How the combined change is built. Each stretch’s change runs from its first close above the average to the close that ended it. Chaining means multiplying (1 + change) across stretches, so -69.32% is what 94 back-to-back short stretches would have done to $1. Every stretch, with its dates and closes, is in the SPY stretches file.

What it does not show. This is not a tested strategy: there are no costs, dividends or slippage, and the closes that define each stretch are known only once the session ends. The 100-session length is one choice of many; a different length shifts every row. Nothing here says when a run will last.

When it fails

The failure is trend-following entries with a session exit, and the record looks like bad entries. The rule buys a breakout above the average at 10:00 AM and sells at 3:55 PM. Most days the move is small or reverses, as the short stretches above did, and the losses pile up exactly as a trend follower expects. The few days that begin a 100-session run are sold the same afternoon, so the part of the distribution that was supposed to pay never arrives. The entries were fine; the clock removed the payoff.

The second failure is treating the two as alternatives. You choose one of each.

A third is judging a trend rule on a few months. Most stretches are short and most of the gain came from a few.

A fourth is changing the average’s length after a loss. That fits the rule to the last mistake.

And a fifth is forgetting costs. The short stretches are the most frequent, and each pays a round trip.

Related

Day trading covers the holding period. Trend following covers the logic. And mean reversion covers the logic most intraday methods actually use.

This page is educational, not financial advice. Test every idea on your own charts before risking money.