Quantum BiasHow Much Does the S&P 500 Move in a Week? Typical Weekly Ranges
Before any bias is useful, you need one number: how big a normal move is.
In this article
Over the last five years, a typical week took the S&P 500 about 2.7% from its high to its low, the Nasdaq-100 about 3.8%, Apple about 4.6% and Tesla about 10.3%. About one week in ten was twice as wide. A reference level inside the normal range will probably get touched anyway. One far outside it needs a reason.
| Question | Data | Answer |
|---|---|---|
| How wide is a normal week, from high to low? | Weekly bars for SPY, QQQ and seven mega-caps, Oct 2021 – Oct 2026 (and full history) | Median: 2.7% (SPY), 3.8% (QQQ), 4.6% to 10.3% for the stocks |
Why range comes before direction
Every bias, ours included, is a call on direction. But a direction is only useful once you know how far the market usually travels. A bullish read with a reference level 1% away means one thing on the S&P 500 and something very different on Tesla. On the index, 1% is a big chunk of a normal week. On Tesla, it's background noise.
Range also drives risk. Put a stop inside a market's normal range and regular movement will take it out, no matter which way the market ends up going.
So before you ask "up or down?", ask "by how much, usually?". We measured that on the two main U.S. indexes and on the seven biggest stocks.
How we measured it
- Weekly range = (the week's high − the week's low) ÷ the prior week's close.
- Median = the middle week: half of all weeks were wider, half narrower.
- 90th percentile = the width only one week in ten went past.
- We report the last five years (October 2021 to October 2026) and the full history since 1999, or since the stock's listing.
- Prices are split-adjusted, and we left out the weeks around stock splits.
What we found
| Market | Median weekly range, last 5 years | 1 week in 10 wider than | Median, full history | Annualized volatility, full history |
|---|---|---|---|---|
| S&P 500 (SPY) | 2.7% | 5.3% | 2.6% | 17.6% |
| Nasdaq-100 (QQQ) | 3.8% | 6.5% | 3.6% | — |
| Microsoft | 4.8% | 8.5% | 4.4% | 28.0% |
| Apple | 4.6% | 9.0% | 5.6% | 36.3% |
| Alphabet | 5.6% | 10.6% | 4.8% | 30.4% |
| Amazon | 6.0% | 11.1% | 6.2% | 47.5% |
| Meta | 6.8% | 13.2% | 6.0% | 38.2% |
| Nvidia | 8.7% | 15.2% | 8.8% | 59.6% |
| Tesla | 10.3% | 18.2% | 9.5% | 55.3% |
Exact values from table 2 in this article. Bar lengths share a zero baseline; no new calculation or simulated price path. Read the table and its stated sample and limits in this article.
Single stocks moved 1.7 to almost 4 times as much as the index. A median Tesla week covered nearly four times the distance of a median S&P 500 week. Even the calmest two of the seven over the last five years, Apple and Microsoft, moved about 1.7 times as much.
Wide weeks were close to double the normal ones. In every market, the 90th percentile was between 1.7 and 2 times the median. That gives you an easy rule of thumb: plan for the normal week, but size for one about twice as wide.
Ranges were also pretty stable over time. For the indexes, the last five years look a lot like the full history. Apple has calmed down since the 2000s, while Meta and Alphabet got a little wilder.
A quick estimate from volatility
If you know a market's annualized volatility, divide it by about 7 and you get a rough idea of its typical weekly range. For the S&P 500, 17.6% ÷ 7 ≈ 2.5%, close to the 2.6% we measured. For Nvidia, 59.6% ÷ 7 ≈ 8.5%, against 8.8% measured. It's a rough shortcut, and it understates the range of the jumpiest stocks, but it's a good sanity check when you look at a market you don't know. (The volatility figures here leave out split weeks, so they can be a tenth of a point off from the ones in our other articles.)
How often the prior week's extremes get taken out
The same data answers a question every bias trader asks: how often does a week trade past the previous week's high or low?
| Market | Takes out prior high | Takes out prior low | Both | Neither (inside week) |
|---|---|---|---|---|
| S&P 500 (SPY) | 61.8% | 41.9% | 12.5% | 8.8% |
| Apple | 56.2% | 40.8% | 9.4% | 12.5% |
| Microsoft | 55.4% | 44.4% | 12.1% | 12.3% |
| Nvidia | 56.8% | 43.2% | 10.3% | 10.3% |
| Amazon | 54.1% | 44.8% | 10.5% | 11.6% |
| Alphabet | 58.3% | 42.3% | 11.7% | 11.0% |
| Meta | 55.7% | 42.9% | 10.5% | 11.9% |
| Tesla | 53.2% | 45.5% | 12.3% | 13.6% |
In close to nine weeks out of ten, at least one of the prior week's extremes got taken out. In every market, the prior high broke more often than the prior low, which reflects how much all of them went up over the period. Only about one week in ten stayed fully inside the previous range.
What this means for you as a trader
Check your levels against the normal range. If a reference level sits inside the typical range, the fact that it gets touched doesn't say much. It was probably going to happen anyway. If it sits past the 90th percentile, getting there takes an unusual week, and you should know what could cause one. Same lesson as our price-structure test: a level getting hit isn't, on its own, proof that a read was right.
Size by range, not by share count. To risk the same dollar amount on a normal week, a Tesla position over the last five years had to be about a quarter the size of an S&P 500 position (2.7% ÷ 10.3%). If you size every stock the same way, you're taking way more risk in some names than in others.
Use the index as your calm baseline. If your stock is moving three times its normal range while the index is quiet, the story is the stock. If both are wide, the story is the market.
Bring it over to futures. If you trade ES or NQ, these weekly ranges work as a ceiling for what to expect intraday, since one session usually covers only part of a week's range. Our session guide explains how the futures day is split up.
How to measure it for your own market
- Export weekly bars, adjusted for splits.
- Compute (high − low) ÷ prior close for each week.
- Take the median and the 90th percentile over the last few years.
- For an intraday version, do the same on daily bars or on the session you trade.
Update the numbers once a quarter. Ranges widen in stressed markets, like 2008, 2020 and 2022, and shrink in calm ones.
Limits
- Weekly bars. Daily and intraday ranges are smaller and behave differently, especially around the open and around data releases.
- Survivor stocks. We picked these seven because they're big today.
- Medians hide clusters. Wide weeks tend to come in groups, and a median tells you nothing about when the next one is coming.
- ETFs, not futures. SPY and QQQ track their indexes closely, but futures trade almost around the clock.
How we did it
- Weekly adjusted prices for SPY, AAPL, MSFT, NVDA, AMZN, GOOGL, META and TSLA from Alpha Vantage (
TIME_SERIES_WEEKLY_ADJUSTED), through October 9, 2026; QQQ fromTIME_SERIES_WEEKLY, split-adjusted by us. - Range = (high − low) ÷ prior close, with highs and lows scaled by the split-adjustment factor and split weeks excluded.
- Volatility = standard deviation of weekly returns × √52.
Sources
- Alpha Vantage. TIME_SERIES_WEEKLY_ADJUSTED and TIME_SERIES_WEEKLY. Data retrieved October 2026.
FAQ
What's the average weekly range of the S&P 500? Over the last five years, SPY's median weekly high-to-low range was about 2.7%, and one week in ten went past 5.3%.
How much does Tesla move in a week? A median week covered about 10.3% from high to low over the last five years, and one week in ten went past 18%.
Is Nvidia more volatile than Tesla? Over the full history, Nvidia's annualized volatility (59.6%) was slightly higher than Tesla's (55.3%). Over the last five years, Tesla's weekly ranges were wider.
How often does the market break last week's high? For SPY, 61.8% of weeks traded above the prior week's high and 41.9% below the prior week's low.
How should range affect position size? Scale your positions so a normal week means a similar dollar risk in each market. On this data, that means much smaller positions in Tesla or Nvidia than in an index.
Research and education only. Not financial advice. Past results do not predict future returns.
How this article was made: drafted with AI assistance from public data and the sources listed above; facts and figures checked by Gabriel Décarie and Karel Panneton.
The week just closed, summarized, and the economic calendar for the week ahead.