How to Spot a Reversal (Not a Pullback): The Two-Condition Rule on 33 Years of SPY
Michael Whitman
·
Trading since 2017, full-time since 2020
·
YouTube creator, 137K subscribers
·
Published 2026-09-05
·
Updated
·
Reviewed
A reversal is confirmed when an uptrend closes below its last higher low and then fails to make a new high, or the mirror of that in a downtrend. Before both of those have happened, what you are looking at is a pullback.
A reversal is confirmed when an uptrend closes below its last higher low and then makes a lower high, or the mirror of that in a downtrend. On SPY since 1993 that rule fired 75 times, and in 42 of the 74 with enough data after them, SPY closed back above the old high within 60 sessions.
Many losing trades in a trend start with someone deciding a pullback was a reversal. The difference is two events, in a fixed order, and both can be checked on a chart before anyone has an opinion.
How it works
An uptrend is a sequence of higher lows. Each swing low sits above the one before it.
The first event is a close below the most recent higher low. The sequence that defined the trend is broken. On its own that is one condition of two.
The second event is a lower high. Price rallies, fails to get back above the high the trend made, and turns down again.
Both, in that order. A market that breaks a low and then makes a new high has done nothing structural. A market that breaks a low and then cannot get back up has changed shape. A downtrend reverses the same way upside down: a close above the last lower high, then a higher low.
Where a pullback ends
A pullback and a reversal look the same for most of their length. Until price either closes above the old high or turns below it, nothing on the chart separates them.
Which is why the level has to be named in advance. “Below this low, the uptrend is broken” is a sentence you can write before anything happens. “That looks like a reversal” is a sentence you can only write afterward.
A worked example
SPY, August to October 2026, five-bar swing points on daily bars. On 20 August SPY made a swing low at $762.04, above the previous swing low of $729.10 on 29 July: a higher low. The trend’s high came on 28 August at $775.30.
Condition one. On 1 September SPY closed at $761.78, below $762.04. One condition of two.
Condition two. The rally that followed topped out on 22 September at $775.14, sixteen cents under the old high, and the five sessions after it were all lower. The lower high was confirmed at the close on 29 September, $764.20, which was 1.43% below the high and 21 sessions after it.
What happened next. On 6 October, five sessions later, SPY closed at $779.09, above the $775.30 high. The reversal had been confirmed by the rule and then undone, which is the most common outcome in the data below. The invalidation, a close back above the high, was the part that did its job.
The original data
Sample: SPY daily bars from Yahoo Finance, 29 Jan 1993 to 27 Oct 2026, 8,494 sessions. A swing low needs a low strictly below the five lows on each side, and it only counts from the session that confirms it, five sessions later. A higher low is a confirmed swing low above the one before it. Condition one is a close below it; each higher low is used once. The trend high is the highest high from that swing low to the break. Then whichever came first was recorded: a close above the trend high, or a confirmed five-bar swing high, formed after the break, below the trend high.
How often the second condition came. Condition one happened 119 times. A lower high came first in 75 (63.03%), confirming a reversal; in 44 (36.97%) SPY closed above the old high first, so the break had been a pullback. Those 44 took a median 12 sessions from the break to the new closing high.
What confirmation cost. At the break SPY’s close sat a median 3.84% below the trend high. By the close that confirmed the lower high it sat a median 4.80% below, a median 22 sessions after the high and 16 after the break.
The finding: a confirmed reversal was a change in structure, not a forecast. Of the 74 confirmed reversals with 60 sessions of data after them, 42 (56.76%) saw SPY close back above the old high within those 60 sessions. Sixty sessions after confirmation SPY was lower 21 times (28.38%). Across all 8,434 sessions with 60 sessions after them, SPY was lower 60 sessions later 27.88% of the time. On this fund, over this period, the rule picked out moments no more bearish than any other day. SPY rose over most of these 33 years, and a falling index would give a different count.
The check. Nasdaq’s own SPY history, 27 Oct 2016 to 27 Oct 2026, run through the same code (one flat row on 20 Apr 2026 with no volume removed), finds the same 34 breaks as Yahoo’s bars over those ten years, with the same dates and the same outcomes: 19 reversals and 15 resumed trends. The check covers the price data and the code, not a second market. Every event is in the SPY reversal file.
Across our study of 24,971 trading videos, 275 have reversal in the title, at a median of 6,966 views. The corpus carries description text for 126 of them, and 9 of those 126 mention invalidation, failure or what a bad read looks like.
Sharp and slow reversals
A sharp reversal turns and goes. The high, a fast break of the last higher low, a weak rally, and away. The two conditions arrive close together and the confirmation is expensive in price but quick in time.
A slow reversal rolls over. A marginally lower high, a drift, a break, a rally almost back to the high, another break. The conditions are met, then nearly undone, then met again, and each cycle looks like the trend resuming.
The rule handles both and feels much worse in the second. In a rolling top you will call the reversal, watch the rally, and call it again. Widening the timeframe helps more than loosening the rule: a rolling top on one chart is often a single clean sequence on the chart above it.
What confirmation costs
On SPY the median answer was 4.80% of price and 22 sessions. That gap is the price of the second condition: a rule that confirms sooner is acting on less evidence.
The honest framing: you are not trying to catch the top. You are trying to stop being wrong about the trend, which is a smaller goal. The Supertrend and moving average pages reach the same trade-off from a different direction.
When it fails
The old high comes back
This was the usual result on SPY. 42 of 74 confirmed reversals saw a close above the old high within 60 sessions, as in October 2026. A close back above the high is the invalidation, and naming it before the trade is what keeps a failed reversal small.
The lower high never comes
A market can break its last higher low and then grind sideways, making neither a new high nor a lower high for weeks. That is a trading range, and the two-condition rule has no answer for it, which is the honest limit.
You changed which low you were watching
If the level moved, you moved it. Picking a different swing low after price approached the first one is the same failure the trend lines page describes.
You called it after one condition
At the break, one condition of two is met. On SPY 44 of 119 breaks went on to a new closing high without a lower high ever forming. Every completed reversal looks obvious from the right-hand side of the chart, and so does every failed one.
Questions people ask
What is the difference between a reversal and a pullback?
A pullback is a dip inside a trend that ends with a new high; a reversal breaks the trend’s last higher low and then makes a lower high. On SPY since 1993, 44 of 119 breaks of a higher low turned out to be pullbacks.
How do you confirm a reversal?
Wait for both conditions: a close below the last higher low, then a confirmed lower high. On SPY’s daily chart that confirmation came a median 22 sessions after the high, 4.80% below it.
Is a lower low enough to call a reversal?
No. A lower low is condition one. In 36.97% of SPY’s cases the next decisive event was a close above the old high instead.
Does price fall after a confirmed reversal?
Not on this data. SPY was lower 60 sessions after 28.38% of confirmed reversals, about the same as the 27.88% for any day in the sample.
Related
Market structure is the sequence a reversal breaks, and where the two conditions come from.
Pullback is the thing a reversal is being told apart from, and 44 of SPY’s 119 breaks were one.
And swing highs and lows decides which turns count as levels, which is the judgment underneath both conditions.
This is the trade I am most careful with, because the error is always the same shape: calling it a reversal at the point where it is still a pullback. I refuse to use the word until both things have happened, which means I am late to every real one. I will take that, because the alternative is being early to every fake one, and that costs more.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.