A 52-week high breakout is a close at the highest level in 252 sessions after at least 60 sessions without one. Across $SPY, $QQQ, $IWM and $DIA since inception, 90 such breakouts were followed by an average 20-session price gain of +1.3%, positive 77.8% of the time, versus +0.7% and 62.0% for all days. The edge showed up clearly in $SPY and $QQQ, thinly in $DIA, and not at all in $IWM. The twist: buying every new 52-week high did worse than buying a random day.
What a 52-week high breakout is
A 52-week high is the highest price of roughly one trading year. This test uses closing prices and a 252-session window, so a new high means the close beats every close of the prior 251 sessions. An intraday spike that fades by the bell does not count.
A breakout from a base adds one filter: the previous new 52-week closing high came 60 or more sessions earlier. That is about three months of going sideways, or of repairing a drop, before price clears the old ceiling. Repeat highs inside a running trend are not breakouts under this rule. Only the first one after a pause is.
The idea has academic backing. Thomas J. George and Chuan-Yang Hwang, in "The 52-Week High and Momentum Investing" (Journal of Finance, October 2004), found that the 52-week high, coupled with the current price, "explains a large portion of the profits from momentum investing." Their explanation: traders use the 52-week high as a reference point and are reluctant to bid a stock higher near it, so good news gets priced in late. Their study ranked individual US stocks on CRSP against each other from July 1963 to December 2001. This article asks a simpler question about four index ETFs over time.
How the test works
- Data: daily closes from Yahoo Finance historical data for $SPY from January 29, 1993, $DIA from January 20, 1998, $QQQ from March 10, 1999 and $IWM from May 26, 2000, through the November 18, 2026 close.
- Prices: the price close, split-adjusted but not dividend-adjusted, so every return below excludes dividends.
- Signal: judged at the close. Each ETF needs 251 prior sessions, so signals start one year after inception.
- Outcome: the close-to-close change 20, 60 and 120 sessions later, roughly one, three and six months. Signals too recent for a full window are left out of that horizon.
- Benchmarks: every new-high day, and every day in the same span.
For the last ten years the Yahoo closes were cross-checked against Nasdaq.com, and the November 18, 2026 closes match exactly. The test ignores costs, taxes and slippage, fills at the signal close, and uses four ETFs that survived, so it describes history rather than a ready-made system.
Results: breakouts vs. every new high vs. every day
Each cell shows the average / median / % positive of the forward price change.
| ETF | Signal | Events (20-session) | Next 20 sessions | Next 60 sessions | Next 120 sessions |
|---|---|---|---|---|---|
| $SPY | Base breakout | 24 | +1.6% / +2.0% / 91.7% | +3.7% / +4.6% / 79.2% | +7.1% / +7.1% / 83.3% |
| $SPY | All new highs | 950 | +0.2% / +0.9% / 63.1% | +1.6% / +2.7% / 72.1% | +4.5% / +5.2% / 75.8% |
| $SPY | All days | 8,239 | +0.8% / +1.3% / 63.9% | +2.3% / +3.3% / 69.9% | +4.8% / +5.9% / 73.8% |
| $QQQ | Base breakout | 22 | +2.5% / +3.0% / 86.4% | +5.4% / +6.0% / 81.0% | +7.5% / +8.7% / 71.4% |
| $QQQ | All new highs | 756 | +0.3% / +1.0% / 60.2% | +1.7% / +2.3% / 64.5% | +5.8% / +7.1% / 80.0% |
| $QQQ | All days | 6,697 | +0.8% / +1.5% / 61.6% | +2.5% / +3.8% / 67.0% | +5.1% / +7.8% / 71.9% |
| $IWM | Base breakout | 21 | +0.4% / +0.1% / 52.4% | +1.6% / +3.1% / 71.4% | +1.4% / +4.5% / 66.7% |
| $IWM | All new highs | 444 | +0.0% / +0.2% / 51.1% | +0.6% / +1.8% / 62.2% | +1.4% / +1.2% / 54.3% |
| $IWM | All days | 6,389 | +0.7% / +1.2% / 59.4% | +2.2% / +2.9% / 63.0% | +4.6% / +5.0% / 65.4% |
| $DIA | Base breakout | 23 | +0.7% / +1.3% / 78.3% | +1.6% / +2.9% / 69.6% | +4.7% / +6.6% / 78.3% |
| $DIA | All new highs | 664 | +0.2% / +0.7% / 62.0% | +1.1% / +1.7% / 63.9% | +2.7% / +3.5% / 71.9% |
| $DIA | All days | 6,983 | +0.6% / +1.1% / 62.6% | +1.7% / +2.4% / 65.2% | +3.4% / +4.0% / 69.3% |
| All four | Base breakout | 90 | +1.3% / +1.7% / 77.8% | +3.1% / +3.7% / 75.3% | +5.2% / +6.8% / 75.3% |
| All four | All new highs | 2,814 | +0.2% / +0.8% / 60.2% | +1.4% / +2.2% / 66.5% | +3.9% / +4.9% / 72.7% |
| All four | All days | 28,308 | +0.7% / +1.3% / 62.0% | +2.2% / +3.1% / 66.5% | +4.5% / +5.6% / 70.4% |
Event counts are for the 20-session window. $QQQ has 21 breakouts at 60 and 120 sessions because its September 22, 2026 signal is too recent. Three patterns stand out.
1. Breakouts beat random days in $SPY and $QQQ. $SPY's 24 breakouts were higher 20 sessions later 22 times (91.7%), with an average gain of +1.6% versus +0.8% for all days. $QQQ's 22 breakouts averaged +2.5% at 20 sessions and +5.4% at 60, against +0.8% and +2.5% for all days.
2. Small caps did not play along. $IWM's 21 breakouts were positive at 20 sessions only 52.4% of the time, under the 59.4% base rate, and their 120-session average of +1.4% trailed the +4.6% all-day average. $DIA landed in between: better hit rates, but a 60-session average of +1.6% that sat a touch below the +1.7% for all days.
3. Ordinary new highs lagged. Pooled across the four ETFs, all 2,814 new-high days averaged just +0.2% over the next 20 sessions, against +0.7% for all days, with a median of +0.8% versus +1.3%. New highs bunch up late in long advances. $SPY logged 71 new 52-week closing highs in 2021 and one in 2022. $QQQ logged 46 in 2007 and none in 2008. The 60-session base filter keeps only 90 of those 2,814 days, and that subset carried the edge.
Where breakouts failed
A high hit rate is not the same as low risk. The worst outcomes after a breakout signal:
- $QQQ's July 22, 2011 breakout fell 16.1% over the next 20 sessions.
- $IWM's April 16, 2002 breakout was down 20.7% after 60 sessions and 33.6% after 120.
- $IWM's November 2, 2021 breakout lost 19.9% over 120 sessions.
- $DIA's May 21, 2001 breakout was down 13.0% after 120 sessions.
The 2002 case shows a quirk of the rule. $IWM's $52.15 close that day was a 52-week high, yet still below its $54.50 close of July 17, 2000. A 52-week high is measured against the past year only, so it can appear in the middle of a bear market. That is one reason traders pair entries with a defined exit, such as an ATR trailing stop. A close cousin of this signal, the Donchian channel breakout, applies the same highest-high logic to a shorter window.
Where $SPY, $QQQ, $IWM and $DIA stand at the November 18, 2026 close
| ETF | Nov 18, 2026 close | 52-week closing high | High set on | Distance from high | Sessions since high |
|---|---|---|---|---|---|
| $SPY | $730.20 | $779.09 | October 6, 2026 | -6.3% | 31 |
| $QQQ | $659.51 | $759.66 | October 6, 2026 | -13.2% | 31 |
| $IWM | $270.12 | $305.09 | August 14, 2026 | -11.5% | 67 |
| $DIA | $501.73 | $542.81 | August 5, 2026 | -7.6% | 74 |
None of the four sits at a 52-week high. $QQQ is furthest away, 13.2% below its October 6, 2026 closing high of $759.66. Its most recent breakout signal, on September 22, 2026 at $747.46, was down 4.0% after 20 sessions and sits 11.8% lower at the November 18 close. Its 60- and 120-session windows are still open, so it is not in the averages for those horizons.
The rule also defines what would count next. $IWM and $DIA have gone 67 and 74 sessions without a new 52-week closing high, so a close above $305.09 or $542.81 would qualify as a base breakout under this definition. $SPY and $QQQ are 31 sessions past their last highs. A new high before the 60-session mark would be a trend continuation, not a base breakout. None of this says which way prices go next.
How traders use the 52-week high
- As a filter, not a trigger. Pooled across the four ETFs, the first high after a 60-session pause historically beat the average new-high day at every horizon.
- Index by index. The pattern held in $SPY and $QQQ but not in $IWM over these samples.
- With sample size in mind. Each ETF has only 21 to 24 signals, and many land on the same dates across funds: $SPY and $DIA both fired on September 26, 2006, August 16, 2012 and July 8, 2016.
- With an exit plan. Several single signals lost more than 10% within six months.
More data-driven trend studies are on Traders On Trend and in the Markets section.
Bull/Bear Verdict
Bull Case: Where it helps: in $SPY and $QQQ, the first new 52-week closing high after a 60-session base was higher 20 sessions later 91.7% and 86.4% of the time, against base rates of 63.9% and 61.6%, and the base filter screened out the weaker repeat highs.
Bear Case: Where it misleads: with only 21 to 24 signals per ETF, a few trades drive the averages, $IWM breakouts trailed all days on average at every horizon, and one signal still fell 33.6% within 120 sessions.
FAQ
What is a 52-week high breakout?
It is a close above every close of the prior 251 sessions. In this article it also requires that the previous such high came at least 60 sessions earlier, so the signal marks a move out of a base rather than another day in a running trend.
Do ETFs keep rising after a new 52-week high?
They rose, but by less than a random day in this data. All new-high days in the four ETFs gained +0.2% over 20 sessions on average, below the +0.7% for all days. The first high after a base did better, at +1.3%. George and Hwang's 2004 study found that nearness to the 52-week high helped forecast returns for individual US stocks, a different test from this one.
Should a 52-week high use closing or intraday prices?
Both versions exist. This test uses closes, so an intraday poke above the old high that fades by the close does not count. An intraday version can fire on days this one ignores.
How far is $SPY from its 52-week high?
At the November 18, 2026 close of $730.20, $SPY was 6.3% below its 52-week closing high of $779.09, set on October 6, 2026, 31 sessions earlier.