How often does the S&P 500 dip — and how deep?

Since 1950, the S&P 500 has fallen about 5% once a year, 10% once every three years, 20% once every seven years, and 30% once every thirteen years. If you hold index funds and keep some cash aside for dips, those base rates matter more than forecasts, because they tell you what counts as routine weakness and what deserves a plan.

Studies · Market historyPublished May 25, 2026DoubleTrends L.L.C.

The base rates

Since 1950, S&P 500 drawdowns have come along on a surprisingly regular schedule. Here's the short version for each depth: how often it shows up and what it should mean to a long term investor:

−5% every year on average

Background noise. They come around about once a year and are usually over in weeks. React to every one and you will never stay invested. Ignore.

−10% every 3 years on average

The most common discount that actually means something. More than two dozen since 1950, with full recovery in about a year on average. Start paying attention.

−20% every 7 years on average

The real ones, and scary while they last. Every single one in the post-war record fully recovered. This is when a pre-planned rule matters.

−30% every 13 years on average

Generational events — 1987, 2008, the COVID crash. A handful in 76 years. In hindsight they produced exceptional long-term entries; in real time they were the hardest to act on. Act from a plan, not panic.

What “every N years on average” actually means: across 76 years (1950 to today), we count each new peak-to-trough drop that crossed the listed depth before the index set a new all-time high. The frequency is the long-run rate — total years divided by that count — not a calendar. Some decades produce several; some produce none.

How deep, and how long

The numbers above told you how often each threshold gets crossed. The two below describe what actually happens once a decline crosses the −10% or −20% line: how far down the index typically reaches, and how long the round trip takes. The threshold gets you in the door. Depth and duration are what the trip looks like once you're on it.

A typical correction (≥10%)

Median depth around 20%. Roughly 8 months from peak to trough, then about a year back to a new high.

A bear market (≥20%)

Average depth around 35%. About 14 months down, and roughly two years to fully recover.

Every −30%+ decline since 1950 — six in seventy-six years:

2007–09 — Global Financial Crisis	−56.8%
2000–02 — Dot-com bust	−49.1%
1973–74 — Oil shock	−48.2%
1968–70 — Late-1960s bear	−36.1%
2020 — COVID crash	−33.9%
1987 — Black Monday	−33.5%

Each row is one episode: the peak-to-trough drop reached before the index set a new all-time high. Price-only, dividends excluded.

Drawdowns are the weather, not the storm

This is the part most investors get wrong. The market is not usually sitting at an all-time high. Since 1950, the S&P 500 has spent about 35% of all trading days at least 10% below its prior peak — and 16% of days at least 20% below.

In other words, about a third of the time the index is far enough below a prior peak to feel uncomfortable. Buying the dip is not mainly about catching a rare crash. It is about deciding in advance how you will behave during normal market weather, then having a rule that keeps you from improvising under stress.

Same data, different cut. The base rates above count new declines as they begin. This counts the share of trading days the index actually sat below a prior peak — time-under-water, not frequency of new dips.

But when is a dip the bottom?

You can't know the exact day in advance. Nobody can, and anyone who says otherwise is selling something. What you can do is stop guessing and use a rule. A good signal doesn't predict the future, it waits for the conditions where past selloffs have run out of steam and then flags it.

Since July 2016, the DoubleTrends™ signal has fired 7 times on the S&P 500 index. Every one of those fires happened in a panic or bear regime — the engine refuses to fire in calm or correction by design. The biggest landed at the COVID crash low (March 2020), the 2022 bear-market low (October 2022), the 2025 tariff selloff, and the March-April 2026 Iran-war drawdown.

One caveat: not every fire marks a generational entry. Some have been early; some happened on shallower stress that resolved without a deep test. The point is not that every alert is the perfect day to deploy capital — it is that the same disciplined rule was present at every moment the largest opportunities appeared, and absent the rest of the time.

See how the rule defines panic and bear →

Measuring “how cheap” against history

Once you accept drawdowns are routine, the next question is how to measure them so they compare cleanly across time. Two coordinates do most of the work: depth — how far below the trailing one-year high — and duration — how long the market has been bleeding.

A 10% drop in two weeks is structurally different from a 10% drop over six months, even though both look identical in the headline. The Maximum Drawdown lens is just the discipline of looking at both axes at once, instead of letting the depth number stand alone.

Read the Maximum Drawdown primer →