12 min readJorgAI TeamNov 15, 2026

Buy the Dip: What Every S&P 500 Dip Since 1950 Shows, Measured

A country road dipping down and rising toward a mountain under a blue sky, cover for the buy the dip study

"Buy the dip" means buying after a price has fallen from a recent high, on the bet that the fall is temporary. It is the most repeated piece of advice in retail investing and one of the least measured. So we measured it on the S&P 500, every trading day from January 1950 to November 13, 2026, 19,340 sessions. The index closed at least 5% below its record high 75 separate times in that span, the latest on Nov 5, 2026, which is where this question lives right now. The short version: buying on the first close 5% below the record was followed by a median gain of +2.4% over the next month, positive 80% of the time, against +1.1% and 62% for a random day, and the index was back at a record within a year 84% of the time. The catch is the other 16%: 16 of those 75 dips fell at least another 10% after the buy, 9 fell another 20% or more, and the simplest dip-buying plan, holding cash until a dip and then buying, has trailed the boring plan of buying every month, in 67% of all ten-year stretches since 1950. The rest of this page is the measurement, including what buying at 10% and 20% down has looked like, what happens the day after a 1% drop, and where the index sits now.

On Nov 13, 2026 the S&P 500 closed at 7,561.67, 3.3% below its record close of 7,818.93 on Oct 6, 2026. It had closed more than 5% below that record on Nov 5, 2026, which opened the 75th episode in this record, and has since climbed back to within 3.3%. A 5% dip from the record is 7,428, a 10% dip 7,037, a 20% dip 6,255. None of this predicts which of those prints; it is the record of what followed the 74 finished dips before this one, from Yahoo Finance's daily closes, method at the end.

What buying the dip means, and what people actually do

The phrase hides three different plans. The first is a reaction: a stock or the market falls, you buy because it is cheaper than yesterday. The second is a rule: you hold cash and buy only when the index is some fixed amount, 5% or 10%, below its high. The third is a feeling: BTFD, the internet's cruder version of the phrase, which means buying every dip on the belief that the market always comes back. The first two can be measured, and this page measures both. The third is a belief about the long run, and the long run has mostly rewarded it, as every record high since 1950 shows; the trouble is that "mostly" has included two stretches of more than six years below the old high, and a plan that cannot survive those is not a plan.

What the S&P 500 did after a 5%, 10% and 20% dip

Buying on the first close this far below the recordDips since 1950Next 1 monthUpNext 3 monthsUpNext 6 monthsUpNext 12 monthsUp
5% below75+2.4%80%+5.7%74%+7.0%77%+8.7%72%
10% below26+2.3%65%+2.5%62%+5.7%73%+13.6%69%
15% below15+2.8%60%+4.8%73%-0.7%40%+14.7%67%
20% below11+2.0%82%+5.2%82%+6.4%64%+22.9%73%
Any trading day19,088+1.1%62%+2.7%67%+5.2%71%+10.7%75%
S&P 500 close-to-close moves after the first close at least the stated distance below the running record close, one entry per episode (the next entry needs a new record first). Months are 21, 63, 126 and 252 sessions. Medians, with the share of cases that were higher. The 5% row's latest entry, November 5, 2026, has no forward data yet.

The first month after a 5% dip has been the best month in this table: +2.4% with the index up 80% of the time, against +1.1% and 62% for any day. Three and six months out the edge persists. By a year it is gone: +8.7% against +10.7% for any day, and a lower hit rate, because the dips that kept falling are in the sample too. The deeper rows show the same shape stretched out. Buying 10% down was followed by a weak quarter and a strong year (+13.6%, up 69%); buying 20% down, which has happened 11 times and is the definition of buying into a bear market, was followed by the best year of all, +22.9%, up 73%, with a median first month of +2.0%. The dip has rewarded the buyer on average at every depth. The averages are not the problem.

The problem: how much further it fell after you bought

Buying on the first close this far below the recordDipsMedian further fall before the lowFell at least 10% moreFell at least 20% moreBack at a record within 3 monthsWithin 12 monthsMedian sessions to a new recordLongest wait (sessions)
5% below75-2.2%16946 of 7462 of 74431,880
10% below26-9.9%1265 of 2617 of 262051,825
20% below11-8.2%430 of 114 of 114021,677
The same entries: the lowest close between the buy and the next record, as a move from the buy price; and how long the index took to set a new record. The November 5, 2026 entry has not recovered yet and is excluded from the recovery columns.

This is the table the slogan leaves out. After buying 5% down, the median dip went another 2.2% lower before it turned, which is nothing, and the index was back at a record within 43 sessions, about two months, in the median case. But 16 of the 75 kept falling by 10% or more after the buy and 9 by 20% or more, and the longest wait for a new record was 1,880 sessions, from the 5% dip of Feb 7, 1973 to the record of July 1980. Buying 10% down was the same bet with worse odds in the middle: a median further fall of 9.9%, 12 of 26 fell at least 10% more, and only 5 of the 26 were back at a record within three months. The Nov 26, 2007 ten-percent dip fell another 51.9% before the March 2009 low; the Apr 14, 2000 dip fell another 42.7% and took 1,788 sessions, seven years, to see a new record.

The 20% row is the bear-market row, and its two halves could not be more different. Oct 19, 1987, the crash day itself, was a 33.2% dip at the close; buying it meant a further fall of only 0.4% and a record 447 sessions later. Mar 12, 2020 was a 26.7% dip; it fell 9.8% more and set a record 110 sessions later. The other side: the 20% dip of Nov 27, 1973 fell another 34.9% and waited 1,677 sessions for a record, Mar 12, 2001 fell another 34.2%, and Jul 9, 2008 fell another 45.6% before the low and took 1,188 sessions to recover. Three of the 11 bear-market buys were followed by a second bear market's worth of decline. The buyer could not tell, on the day, which three.

Every 10% and 20% dip since 1950

First close 10% below the recordDepth that dayNext 12 monthsFurther fall before the lowSessions to a new record
Jun 29, 1950-10.1%+21.7%-4.4%59
Jun 9, 1953-11.5%+20.1%-3.8%189
Oct 11, 1955-10.6%+14.8%0.0%23
Nov 21, 1956-10.0%-8.5%-12.7%463
Mar 4, 1960-10.1%+17.4%-4.3%227
Apr 30, 1962-10.2%+7.0%-19.8%339
May 16, 1966-10.3%+10.3%-13.3%244
Mar 5, 1968-10.1%+15.7%0.0%37
Jun 19, 1969-10.3%-21.3%-28.7%686
Apr 27, 1973-10.8%-15.9%-41.9%1,825
Aug 24, 1981-10.7%-7.5%-18.4%303
Feb 13, 1984-10.3%+16.5%-4.6%237
Oct 15, 1987-11.5%-7.7%-24.9%449
Jan 30, 1990-10.2%+4.0%0.0%82
Aug 17, 1990-11.1%+17.6%-9.9%124
Oct 27, 1997-10.8%+21.5%0.0%28
Aug 14, 1998-10.4%+25.2%-9.9%70
Sep 29, 1999-10.6%+12.5%-1.7%34
Apr 14, 2000-11.2%-12.1%-42.7%1,788
Nov 26, 2007-10.1%-39.5%-51.9%1,343
Aug 24, 2015-11.2%+15.5%-3.4%221
Feb 8, 2018-10.2%+5.0%0.0%137
Nov 23, 2018-10.2%+19.0%-10.7%101
Feb 27, 2020-12.0%+27.9%-24.9%120
Feb 22, 2022-10.3%-6.8%-16.9%479
Mar 13, 2025-10.1%+21.3%-9.8%73
All 26 corrections and bear markets measured from the first close at least 10% below the record, which is the same 26 episodes as our correction and drawdown pages. A further fall of 0.0% means the entry day was the low.

Read the last two columns together. The fast ones, Oct 27, 1997, Aug 14, 1998, Feb 8, 2018, Mar 13, 2025, were back at a record inside a quarter or so, and Mar 13, 2025 was followed by +21.3% over the next year after a further 9.8% fall in the tariff selloff. The slow ones are the bear markets, and Feb 22, 2022 is the most recent lesson: the index fell another 16.9% after the 10% mark, was -6.8% a year later, and needed 479 sessions to set a record. Every 10% decline since 1950 sorts these into corrections and bears, and the drawdown page has all 682 dips from a record, down to the smallest.

The simplest dip: the day after a down day

After a day the S&P 500 fellDays since 1950Next day, medianNext day upNext month, medianNext month up
1% or more1,9390.0%50%+1.5%62%
2% or more444+0.2%55%+2.4%65%
3% or more132+0.3%58%+1.7%61%
Any day19,339+0.1%53%+1.1%62%
Close-to-close moves after a single down day of the stated size. Overlapping days are counted separately.

The one-day dip is the version most people practice, and it has been close to a coin flip. After a 1% down day the next session was up 50% of the time with a median of 0.0%, a little worse than any day; after a 2% down day, 55% and +0.2%; after a 3% down day, 58%. A month out the 2% days have the best record, +2.4% with 65% up, which is a small edge over any day's +1.1%. Nothing here is large enough to trade on its own; how often the index drops 1%, 2% or 3% in a day has the counts by decade and what they mean for a stop. The reason the one-day dip disappoints is that big down days cluster: the day after a 3% drop is a day inside the same storm, and the failed rallies since 1950 are what buying the first bounce has often bought.

Waiting for the dip has cost more than the dip has paid

Plan, $100 on the first session of every month since January 1950Value on November 13, 2026Against buying every monthNumber of buys
Buy every month, whatever the price$5,919,310923
Hold cash, buy everything when the index is 3% below its high$5,914,066-0.1%740
Hold cash, buy everything when the index is 5% below its high$5,873,546-0.8%593
Hold cash, buy everything when the index is 10% below its high$5,697,325-3.8%390
Hold cash, buy everything when the index is 20% below its high$5,138,265-13.2%185
Both investors receive $100 on the first session of each month. The first buys at once. The dip waiter holds the cash until the S&P 500 closes the stated distance below its highest close since 1950, then buys with all of it, and keeps holding what it bought. Price index, no dividends, no interest on cash, no costs; both are measured at the same close.

Here is the rule tested against the alternative it competes with. Across 923 monthly contributions since 1950, the investor who bought every month ended with $5,919,310; the investor who waited for a 5% dip ended -0.8% behind, the one who waited for 10% -3.8% behind, and the one who held out for a 20% dip -13.2% behind, because the market spent years at a time setting new highs while that cash waited. Even the 3% rule, which fired 740 times, came out slightly behind. In rolling ten-year windows starting each January from 1950 to 2016, the monthly buyer beat the 5% dip waiter in 45 of 67 windows, by a median 0.4%, with the dip waiter's best window only +0.7% ahead and its worst -3.2% behind; against the 10% waiter the monthly buyer won 47 of 67 windows, and the waiter's worst decade was -36.7%, the 1990s, when the index did not close 10% below its high from 1990 until 1997 and the cash sat. Lump sum against dollar-cost averaging measures the other half of this question for someone with a lump of cash today.

Two honest caveats. The simulation pays no interest on the waiting cash; with the money in Treasury bills the gap narrows in high-rate decades and never closes in the long ones. And it ignores that a dip buyer who is already fully invested has nothing to buy with: the rule only works for new money, which is the money the monthly plan was investing anyway.

What the record says about buying this dip

The index closed 5% below its record on Nov 5, 2026 and has since recovered most of it. Of the 74 finished 5% dips before this one, 46 were back at a record within three months and 62 within a year; 16 fell at least 10% more first. The two most recent, Nov 20, 2025 and Mar 18, 2026, were the fast kind, back at records in 14 and 19 sessions with the next quarter +5.4% and +12.0%; Mar 4, 2025 fell another 13.8% into the April tariff low before returning to a record in 80 sessions and finishing the year +18.2%; Jan 19, 2022 was the slow kind, another 21.1% down, -12.4% a year later, 502 sessions to a record. Nothing in the depth of a 5% dip has told you which kind it is. Two things have helped at the margin: the golden cross and death cross record shows the trend signal has lagged every turn, and the deepest dips, the ones that printed after a fast fall, were followed by stronger years than the shallow grinds.

How a rules-based trader buys a dip

If the dip is the entry, the exit is the plan, because the record above says the entry is right most of the time and badly wrong some of the time, and you will not know which on the day. The rules that follow from the numbers:

Buy in pieces, not at once. The median 5% dip went 2.2% lower before turning and the worst went 45% lower; a position built in thirds over the decline is in the market for the common case and has room for the rare one.

Put the stop at the broker before the bounce, not after. A dip that turns into a 10% correction has taken a median 205 sessions to make a new record; the stop is what keeps a dip trade from becoming a year-long hold you did not plan.

Keep the monthly plan running. The simulation above is the whole argument: money that waits for a dip has lost to money that did not, in most decades, by a little, and in the long bull markets by a lot. Dip buying is a tactic for new money and for position sizing, not a reason to hold cash.

JorgAI runs the rules a member sets inside their own brokerage account: position caps, a daily spend cap, stops and targets placed as orders at the broker, a daily loss brake, and a log of every candidate it declined and why. It does not time dips and it does not hold cash for one; the caps are what size the entry and the orders at the broker are what end it. Every cap and exit is defined in the settings glossary; to see the rules run on a simulated account before connecting a broker, start on the free plan.

Quick answers

What does buy the dip mean?

Buying a stock or an index after it has fallen from a recent high, on the expectation that the fall is temporary. BTFD is the same idea in cruder internet form. It can mean buying after a single down day, buying when the market is a set percentage below its high, or simply buying every decline on the belief that the market recovers.

Does buying the dip work?

On the S&P 500 since 1950, buying the first close 5% below the record was followed by a median +2.4% over the next month, up 80% of the time, better than a random day, and the index was back at a record within a year 62 of 74 times. But 16 of 75 such dips fell at least 10% further first, and a plan that holds cash waiting for dips has trailed buying every month in 67% of ten-year windows. It works as an entry tactic, not as a reason to wait.

Should I buy the dip now?

The S&P 500 closed 5% below its record on Nov 5, 2026 and was 3.3% below it on Nov 13, 2026. History says a 5% dip has been followed by a recovery to a record within three months 46 of 74 times and by a further fall of 10% or more 16 of 75 times, and it does not say which this one is. Buy in pieces, with the exit placed at the broker, and keep any monthly plan running.

How often does the stock market dip 5%?

75 times since 1950 measured from the record close, about once a year, with 26 of them reaching 10% and 11 reaching 20%. The median 5% dip was back at a record in 43 sessions; the longest took 1,880, from February 1973 to July 1980.

Is it better to buy the dip or dollar-cost average?

Dollar-cost averaging, on this record. The investor who put $100 into the S&P 500 every month since 1950 finished ahead of the one who held each $100 until a 5% dip, by 0.8% over the full period and in 45 of 67 ten-year windows; against a 10% rule the gap was 3.8%, and against a 20% rule 13.2%.

What happens the day after a big down day?

Close to a coin flip. After a 1% drop the next session was up 50% of the time; after a 2% drop, 55%; after a 3% drop, 58%; any day, 53%. A month later the 2% days had the best record, up 65% with a median +2.4%.

How we put this together. Daily S&P 500 closes from Yahoo Finance (^GSPC), January 3, 1950 to November 13, 2026, 19,340 sessions, price index without dividends. A dip entry is the first close at least the stated percentage below the running record close; the episode ends when the index closes at a new record, and the next entry at that depth requires a new record first, so each episode counts once. Forward moves are close to close over 21, 63, 126 and 252 sessions, compared with the same horizons from every session. The further fall is the lowest close between the entry and the next record; sessions to a new record count from the entry. The down-day table takes every session with a close-to-close drop of the stated size. The simulation credits $100 on the first session of each month from January 1950, 923 contributions; the monthly buyer invests at that close, the dip waiter accumulates cash and invests all of it at the first close at least the stated distance below the highest close since the start, then holds; no dividends, no interest on cash, no costs, both valued at the November 13, 2026 close. The ten-year windows start on the first session of each January from 1950 to 2016 and end on the first session of January ten years later. The data and script are kept, and the page is refreshed when the current dip resolves.

The figures in this article describe what an index did after past declines and are not a prediction or a recommendation to buy or sell any security. JorgAI is software that carries out the rules you set in your own brokerage account; it is not an investment adviser. Past performance does not guarantee future results, and trading involves risk of loss.

Let the AI do the trading.

Set it up in minutes. 7 day free trial on Starter, Plus and Pro.

Get started free

Get the weekly market read and new tools by email

One email a week: what moved, what the data says, and the next free tool when it ships. No spam, unsubscribe in one click.

More from the blog

Every other post, newest first

247 more posts

Stops, targets and a daily loss limit, enforced for you.

JorgAI runs the rules you set inside your own brokerage account. Zero withdrawal permissions, disconnect anytime.

Get Started

Showing 18 of 247. Keep scrolling.