How far below the 52-week high: what the 52-week position says and does not say

One of the sentences heard most often when people choose stocks is this: it is down 30 percent from its 52-week high. Listeners usually take it to mean the stock is now priced low. Yet the sentence contains no information about the company at all. All it contains is the highest price tag the stock carried within the past year.

The high used as the reference is not a value the company set, nor one derived from earnings or assets. It is the most that market participants paid on any single day in the past year. If that day was overheated, a 30 percent fall is a return to normal; if it was an ordinary day, the same 30 percent means something entirely different. The rate of decline alone gives no way to tell the two apart.

People fall for this sentence easily because of a long-studied bias. When there is no good basis for judgment, people tend to take whatever number is in front of them as a reference point, and in stocks the 52-week high often plays that role. Depending on what the high happened to be, the same company can look priced low or priced high.

The sections below set out exactly what the 52-week position that Jupick shows on stock pages and market pages measures, how the Korean and overseas distributions differ, and which questions this number cannot answer.

Jupick's 52-week position maps where the current price sits between the low and the high of the past year onto a scale of 0 to 100. It divides the current price minus the low by the high minus the low. At the low it is 0 percent, at the high 100 percent, and exactly in the middle 50 percent.

It is a different number from the decline from the high. The decline uses only the upper end as its reference, while the position uses both ends. Two stocks can each be 30 percent below their highs, yet a stock that swung more than twofold over the year and one that moved in a narrow range end up with very different positions.

Jupick also set rules for when not to show it. If either the high or the low is missing, if the high is not greater than the low, or if the low or the current price is zero or below, the position is not calculated. The market page leaves that value as a hyphen, and the stock page does not draw the range bar at all. Even a value that looks computable becomes a wrong answer in the shape of a number if a collection accident is mixed into it.

One more thing should be clear. Position is not a judgment on whether a company is priced low or high. It only says where today sits within the past year's price record. If price relative to earnings is the question, there are separate metrics for it, PER (price to earnings ratio) and PBR, and those too cannot be compared as is across sectors.

Jupick lined up the 52-week positions of the major stocks it tracks, using values measured on August 31, 2026. The position was calculated for 198 Korean stocks and for every overseas stock. Why two Korean stocks were left out is covered separately later.

The Korean median was 37.7 percent, meaning more than half the stocks sat in the lower half of their one-year range. By quartile, the lower quartile boundary was 24.8 percent and the upper quartile boundary 54.0 percent. Thirty-two stocks had a position below 20 percent, and 12 were above 80 percent.

The overseas list leaned the other way. The median was 65.4 percent, the lower boundary 46.6 percent and the upper boundary 79.8 percent. Fifteen stocks were below 20 percent and 49 were above 80 percent. A position that ranks in the upper tier in Korea sits near the median overseas.

This contrast matters because the same number, a 60 percent position, means different things in the two markets. In the Korean list, 60 percent is close to the top quarter; in the overseas list, it is below the median. Calling a single number high or low on its own is half wrong either way.

This difference does not say that one market is better. The two lists differ from their selection rules onward. The Korean list is reselected every day from the largest stocks by market cap, while the overseas list is a fixed roster of large US-listed stocks. With different rosters and different paths over the past year, different distributions are to be expected.

If position says where a stock stands within the band, band width says how long that ruler is to begin with. It can be read as the high divided by the low. A width of 2x means the top and bottom of the year were twice apart; 1.2x means the stock moved within a narrow channel.

Jupick measured this width for the same stocks. The Korean median was 2.48x and the overseas median 1.51x. That means half the Korean stocks spread more than twofold between top and bottom over the year. The top 10 percent boundary was 5.52x for Korean stocks and 2.19x overseas, and the widest stock was 15.91x in Korea and 10.98x overseas.

The counts at both ends diverge too. Stocks wider than 5x numbered 29 in Korea and 2 overseas, while narrow stocks below 1.3x numbered 1 in Korea and 20 overseas. Within the same set of major stocks, one side is crowded at the wide end and the other at the narrow end.

The wider a stock's band, the more sensitive its position becomes. In a band spread 15x, even a small move in the current price makes the position jump by several percentage points. Conversely, in a narrow 1.2x band, the actual price gap between a 90 percent position and a 30 percent position is small. Looking at position without width leads to overestimating or underestimating the same percentage.

Measuring the size of price swings by the gap between high and low is itself an old approach. A calculation set out in 1980 showed that a day's high and low alone measure volatility more accurately than closing prices alone. Jupick does not compute such statistics, but placing band width next to position follows the same idea. The distance between two end points holds information that a single close cannot.

Fields other than prices often use the same kind of number. Agencies that handle weather records label a month's temperature with band names: its rank since records began, whether it falls in the top third, whether it is in the top 10 percent, or whether it set a new record. Rather than giving only the absolute temperature, they also give its position within the record.

This approach is useful for two reasons. One is that it allows comparison between things measured in different units. A KRW 80,000 stock and a $200 stock cannot be lined up by price, but both positions fall between 0 and 100. The other is that the wording does not exaggerate. “Top 10 percent” is a statement that holds only within the observation period, and the phrase itself already states that scope.

It comes with a caveat. A band name depends on the length of the record. Fifty-two weeks is one year, and one year is short next to the history of a company. Widen the window to three or five years and the same stock's position comes out as a completely different number. A 90 percent position means near the top within one year, not high by historical standards.

That is also why Jupick writes the position only as a percentage and attaches no verdict words saying a stock is priced too low or too high. Such words only work by bringing in one more reference point, and if the screen sets that reference point on the reader's behalf, the reader has a harder time knowing what their judgment rested on.

Earlier this article noted that only 198 (99%) of the major Korean stocks were calculated. The other two had a stored 52-week low of zero. The price had not been KRW 0; the collection had gone wrong.

The cause lay in how trading halt days and data gap days looked. The source of price data returns the candles for such days as rows where the open, high and low are all zero. Over a year of rows, the high takes the largest value, so mixed-in zeros have no effect, but the low takes the smallest value, so a single zero row is enough for it to land on that zero. The structure broke only one side of the band, not both. One stock had just one such day and the other had 29, yet both ended up with a low of KRW 0. One row was enough.

More troublesome, a zero once stored did not heal itself. Each update ran a correction that compared the low with the current price and set it to the smaller of the two, but zero is always smaller than the current price, so it passed straight through that correction. The check for a broken band also looked only for a low above the current price, so zero did not trigger it. Three fixes went in this time: rows of zeros are no longer counted, a stored low of zero or below is fetched again immediately without waiting for the rotation cycle, and the stock page now uses the same validity rules as the market page.

The overseas side had a hole in the opposite direction. The current price correction that existed for Korean stocks was missing for overseas stocks, so on days when the 52-week high was updated later than the current price, some stocks had a current price standing outside the band. One stock had its position calculated at 128.5 percent. The same rule now applies overseas as well.

This story is recorded here because the position is tied directly to the quality of its inputs. The position is built from just two end points, the high and the low. It is not a structure like an average, where many values blend and errors get diluted; if one end point is off, the whole result is off. That is why leaving the value blank is better than displaying a plausible-looking number.

The 52-week position compares the past year's record with today's price. So the range of questions it can answer is narrow.

It cannot answer whether the company is priced low. The position contains no earnings, no assets and no debt. A stock that came down over a year of worsening results and a stock whose results held steady but came down along with the whole market can show exactly the same position.

It also says nothing about what comes next. Jupick does not calculate the future direction or level of prices, and does not write sentences such as “the position is low, so a rise is due” or “the position is high, so a fall is due.” Whether a low position is a sign of recovery or one scene in continuing weakness is not contained in this number.

Comparing the positions of two different stocks in the same list also calls for care. If band widths differ, the same position is the result of different swings. As seen earlier, even within the Korean list alone, widths range from below 1.3x to 15.91x.

So what is this number for? It is for checking where the price you are looking at sits within the past year. Anything beyond that is a job for other metrics. Asking one number to answer many questions leaves it unable to answer any of them properly.

The decline from the high uses only the upper end as its reference. The 52-week position uses both ends to express where the current price sits within the one-year range, on a scale of 0 to 100.

The same position number means different things in different markets. In the August 31, 2026 measurement, the median was 37.7 percent for 198 Korean stocks and 65.4 percent for major overseas stocks.

Position needs to be read together with band width. The median width was 2.48x in Korea and 1.51x overseas, and 29 Korean stocks and 2 overseas stocks were wider than 5x.

Because the number is built from just two end points, it is vulnerable to collection accidents. Two Korean stocks had a low stored as zero and one overseas stock had a position above 100 percent; both the collection and the screens were fixed this time.

Position does not answer whether a company is priced low or high, or what comes next. It only tells where today sits within the past year's price record.

This content is for information only and is not investment advice. Jupick gives no trading advice on any particular stock and does not forecast the future direction of prices. The stocks and figures used as examples serve only to explain how the 52-week position is calculated and where its limits lie, and are not an assessment or advice of any kind. The figures cited were measured on August 31, 2026 and change as the data is updated. Investment decisions and responsibility for them rest with the user.