Mean Reversion vs Trend Following
Mean reversion bets that price returns toward an average, producing frequent small wins with occasional large losses. Trend following bets that a move continues, producing frequent small losses with occasional large wins. The two result shapes are exact mirror images of each other.
Two opposite bets on the same chart. One says a move has gone too far and will come back; the other says a move that has started will continue. Both work in the right conditions, and their result shapes are mirror images of each other.
What each one is
Mean reversion bets price returns toward an average. It buys weakness and sells strength, wins a high proportion of the time, and loses when a move keeps going. Mean reversion covers it.
Trend following bets a move continues. It buys strength, loses a high proportion of the time in small amounts, and makes its money on the few that run. Trend following covers that side.
Both are conditional on the market, not on skill. Each works when its condition holds and neither can tell you in advance which one you are in.
Where they differ
The shape of the results. Reversion produces a long run of small wins punctuated by a large loss. Trend following produces a long run of small losses punctuated by a large win.
Which one can ruin you. Only reversion, and quickly. Its losses are the large ones, and they arrive in exactly the conditions where its own logic says the position has become better value.
What is hard about running them. Trend following is hard because it is boring and mostly wrong. Reversion is hard because the discipline required is to exit a position the method itself argues for adding to.
Where they make money. Reversion needs a range. Trend following needs a move. On this site’s shared series direction runs average 2.01 bars with a longest of 11 — short runs are the norm, which favours reversion far more often than it favours trend.
Where they agree
Both need a written invalidation. Neither can be sized without one, and the arithmetic — risk figure divided by distance — is identical.
Both are judged on expectancy, not win rate. Comparing a reversion method’s high proportion of winners to a trend method’s low one is comparing halves of two different sums.
Both cost a round trip per trade — about 2% of the median bar range of 0.493 here — and reversion pays it far more often because it trades more.
And both fail in the other’s conditions. Neither is broken when it loses; it is being run in the market the other one needed.
Which one to use
Run trend following when you can sit through long flat stretches and a low proportion of winners. That temperament is rarer than it sounds, and the method is abandoned during the flat stretch far more often than it fails on the numbers.
Run mean reversion when you need frequent feedback and can enforce a stop mechanically. Frequent trades produce a reviewable record quickly, which is genuinely valuable — provided the exit is not left to judgement.
Run reversion only with an externally enforced exit. The method’s logic says a losing position is better value, so a stop that depends on your agreement at that moment will not have it.
And when you cannot decide, run trend following. Its failure mode is a slow bleed you can see and stop; the other’s is a single loss larger than the run of wins that preceded it.
Why only one of them can end an account
Because the losses sit on opposite ends. A trend method’s worst trade is a small one by design; a reversion method’s worst trade is its largest, and it is the one the approach was structurally least prepared for.
And because the run of wins beforehand is persuasive. Nine successful reversions build confidence and, frequently, position size — which is precisely the state the tenth one arrives into.
The original data
Of the 24,971 unique videos in research/search-study-corpus.jsonl, 2 compare the two directly in the
title, at a median of 8,943 views. Separately, trend following appears in 116 titles at a median of
3,608 across 91 channels, and mean reversion in far fewer. The counts come from site/rank_compare.py
and site/corpus_count.py.
116 videos on trend following against a handful on mean reversion. The approach that suits fewer temperaments has the coverage, while the one most retail methods actually implement — buying dips, fading extremes — is barely named as a strategy type at all.
The answer to the question on that chart is that a run of wins is what the method produces by design. It says nothing about whether the next one is the large loss — and sizing up after a winning run is how a survivable method becomes an account-ending one.
When it fails
The failure is a reversion method with a discretionary exit, and every individual decision is defensible. Price moves against the position, which by the method’s own logic makes it better value. Adding is consistent with the reasoning. The stop gets widened because the level was arbitrary anyway. Each step follows from the approach, and the combined result is a position several times the intended size in the one market condition the method cannot survive.
The second failure is running trend following without patience. It is abandoned flat, not losing.
A third is comparing win rates. The shapes are mirror images by design.
A fourth is switching after a losing run. You arrive as the other condition ends.
A fifth is running both without separating the records. Neither can then be judged.
And a sixth is treating either as broken when it loses. It is in the other’s market.
Related
Mean reversion covers the fade-the-move side. Trend following covers the ride-the-move side. And the expectancy calculator is what makes the two comparable at all.
The asymmetry that matters is which one can end you. A trend method’s losses are small and frequent, so a bad run is uncomfortable and survivable. A reversion method’s losses are the large ones, and they arrive exactly when the position has been added to because it looked better.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.