H084
"Up 40% this month!" Did this stock really have a great month?
The headline is true. But where a chart starts decides what story it tells, and a window that starts at the bottom can turn a collapse into a comeback without a single false number.
One stock, three true numbers
A hypothetical stock's price, January to July
The shaded month is the only part the headline talks about.
Every trend depends on where you start measuring it. Start in June and this stock had a great month. Start in January and it lost almost three-quarters of its value. Both are true; the headline just picked the window that tells the better story.
The same trick works on anything that goes up and down: crime rates, test scores, a team's record, a company's sales, the temperature. Starting at an unusual low makes almost anything look like a rebound, and starting at an unusual high makes it look like a decline. When you see a trend, ask what the chart looks like if it starts earlier.
Try it somewhere else
A team won its last 5 games after losing its first 12. A headline says “Unbeatable!”
How would the story change if the chart started at the beginning of the season?
The same prices, measured from different starting months
| Starting month | Price then | Change to July's $28 | Possible headline |
|---|---|---|---|
| January | $100 | −72% | "Stock collapses" |
| February | $78 | −64% | "Stock collapses" |
| March | $58 | −52% | "Stock loses half its value" |
| April | $42 | −33% | "Stock down a third" |
| May | $28 | 0% | "Stock holds steady" |
| June | $20 | +40% | "Stock surges" |
Six true headlines about the same July price. Only the starting point changed.
Why a 40% gain barely dents an 80% loss
Percentages are measured from the starting price of each window. The 80% fall was 80% of $100, or $80. The 40% rise was 40% of just $20, or $8. To get from $28 back to $100, the stock would need to gain another 257%. Average return vs. what you actually end up with shows the same lopsided math.
Related exhibits
Sources: Prices are a hypothetical example built to show the effect; all percentages are calculated from the prices shown.