Is a big down day a buying opportunity?

On average the forward return is higher than average, and the odds of being up are lower than average. Both are true at once because big days arrive in clusters: the sample contains the bottoms, where the rebound was violent, and the early stages of long declines, where it was not. An average pulled up by the first group is not a description of the typical case.

Why are so many of the big days in the 1930s?

Because the index was genuinely more volatile then. Roughly half of all 3% sessions in the record fall in a single decade, which means any statistic conditioned on a 3% day is heavily a statistic about the 1930s. The decade breakdown on this page is there so that a reader can see how much of the sample they are actually quoting.

Do big up days behave like big down days?

More alike than most people expect. They occur at a similar frequency, they cluster with each other, and their forward returns are in the same region. Large sessions of either sign are a signature of a high-volatility regime rather than a directional signal, which is the same conclusion the volatility page reaches from the other end.

Plutux

Free tools

Market statistics

What the S&P 500 did after a 2%, 3% or 5% single-day move

The average year after a 3% fall beats the average year — and finishes positive less often than the average year. Both halves of that sentence are on this page, because quoting either one alone is how this statistic is normally misused.

24,781 sessions · data through 28 Aug 2026

1928–2026

Sessions that fell 2%+: 878; 3.54% of all sessions
Sessions that rose 2%+: 802; 3.24% of all sessions
3% falls with another within a week: 64%; big days arrive in clusters
Year after a 3% fall: +11.28%; baseline +8.20%

After a big fall. Each cell is the average forward return over that horizon, with the share of occurrences that finished higher beside it. The final row is every session in the record over the same horizons.

Condition	Occurrences	1 month	3 months	6 months	1 year
Fell 5% or more	870.35%	+3.03%64% up	+9.07%64% up	+11.30%74% up	+18.29%64% up
Fell 3% or more	3641.47%	+0.83%56% up	+3.87%56% up	+5.19%60% up	+11.28%62% up
Fell 2% or more	8783.54%	+1.25%60% up	+3.60%60% up	+4.64%62% up	+8.95%62% up
Any session (baseline)	24,781	+0.67%60% up	+2.01%64% up	+3.98%67% up	+8.20%70% up

After a big rise, on the same construction — the control group for everything in the table above.

Condition	Occurrences	1 month	3 months	6 months	1 year
Rose 2% or more	8023.24%	+1.25%58% up	+2.58%61% up	+4.10%60% up	+9.62%65% up
Rose 3% or more	2991.21%	+1.72%56% up	+2.85%54% up	+3.90%58% up	+12.13%63% up
Rose 5% or more	790.32%	+2.79%62% up	+5.26%51% up	+5.39%61% up	+16.37%65% up
Any session (baseline)	24,781	+0.67%60% up	+2.01%64% up	+3.98%67% up	+8.20%70% up

When the big sessions actually happened. A conditional statistic is a statistic about the decades its sample came from.

Condition	20s	30s	40s	50s	60s	70s	80s	90s	00s	10s	20s
Fell 5% or more	9	42	6	2	1	—	5	2	13	1	6
Fell 3% or more	21	179	32	6	3	3	13	9	52	19	27
Fell 2% or more	33	339	62	23	14	30	51	42	146	71	67
Rose 2% or more	20	308	47	19	16	47	65	47	129	48	56
Rose 3% or more	10	155	10	3	5	12	11	9	54	12	18
Rose 5% or more	5	49	1	—	—	1	2	2	12	—	7

Price returns from daily closing levels. Index price returns exclude dividends. Full provenance, method and a citation line are in Sources and method below.

Two true sentences that point in opposite directions

In the year after a session that fell 3% or more, the S&P 500 averaged +11.28%, against +8.20% for the average year in the same record. In that same year it finished higher 61.8% of the time, against 69.8% unconditionally. The average is better and the odds are worse, and both are computed from the same 364 events.

That combination is not a paradox, it is what a fatter distribution looks like. Big sessions arrive in clusters — 64% of 3% falls had another one within a week — and the clusters happen at both ends of a decline. Some of these events are bottoms, and the year afterwards was violent in the good direction. Others are the second week of a two-year bear market. The mean is pulled up by the first group; the hit rate is pulled down by the second.

The decade histogram is the caveat that matters most for anyone quoting these figures. 49% of every 3% fall in the record happened in the 1930s. A conditional statistic drawn mostly from one decade is a statistic about that decade, and the 1930s were a market with a different volatility regime, a different market structure and no circuit breakers.

The up-day table is there because it is the control. Sessions that rose 2% or more happen about as often as sessions that fell 2% or more, cluster with each other in the same way, and are followed by forward returns in the same region. A big day is a fact about volatility, not a fact about direction — which is the same conclusion the volatility page reaches by counting 1% days.

Questions people ask about this

Is a big down day a buying opportunity?: On average the forward return is higher than average, and the odds of being up are lower than average. Both are true at once because big days arrive in clusters: the sample contains the bottoms, where the rebound was violent, and the early stages of long declines, where it was not. An average pulled up by the first group is not a description of the typical case.
Why are so many of the big days in the 1930s?: Because the index was genuinely more volatile then. Roughly half of all 3% sessions in the record fall in a single decade, which means any statistic conditioned on a 3% day is heavily a statistic about the 1930s. The decade breakdown on this page is there so that a reader can see how much of the sample they are actually quoting.
Do big up days behave like big down days?: More alike than most people expect. They occur at a similar frequency, they cluster with each other, and their forward returns are in the same region. Large sessions of either sign are a signature of a high-volatility regime rather than a directional signal, which is the same conclusion the volatility page reaches from the other end.

Sources and method

Data: Financial Modeling Prep — Daily adjusted closing levels and quotes, retrieved through Plutux's own data service.
S&P Dow Jones Indices — Publisher and methodology owner of the S&P 500 index itself.
How it was calculated: An event is a close-to-close change at or beyond the stated threshold. Forward returns are measured in trading sessions — 21 for a month, 63 for a quarter, 126 for half a year, 252 for a year — and each conditional row is shown beside the same horizon measured from every session in the series, so the comparison is drawn from the same window rather than from a quoted long-run average. The clustering column is the share of events with another event within five sessions. Price returns, dividends excluded.
How often it changes: Regenerated from the full daily history about once a year; the date it runs through is at the top of the page.
Citing this page: Free to quote — please link rather than copy the table.
Plutux. "What the S&P 500 did after a 2%, 3% or 5% single-day move." Data through 28 Aug 2026. https://plutux.ai/es/resources/tools/after-a-big-down-day

Historical figures for information only — not investment advice, and not a forecast.

Sobre nosotrosPolítica de privacidadTérminos de servicioAviso legalDirectorio de accionesETFÍndicesCarteras 13FPrecios objetivoHerramientas gratuitasMapa tecnológicoÚnete a nuestro Slack

Plutux no es un asesor de inversiones. Los datos de mercado y el análisis generado por IA son solo informativos y educativos, no asesoramiento de inversión. Aviso legal

1928–2026

Two true sentences that point in opposite directions

Questions people ask about this

Sources and method

Related tools