How wide are Jupick's grade bands: 166 major stocks (29.4%) sat at exactly 50, the floor of grade B

Open a stock page and the grade badge sits at the top. It is one of five letters from S to D.

Grade B runs from 50 to 69 points. That is a 20-point band. So two stocks wearing the same B badge can in fact be 19 points apart. To someone who reads only the letter on the badge, the two are exactly the same stock.

Jupick knew about this problem but put it off for a long time. The grade boundaries were already all written on the methodology page, so the view was that readers could do the subtraction themselves. On September 16, 2026, Jupick set that view aside and began stating on the stock page where the overall score sits within its own grade band.

While doing that work, Jupick measured once where every stored stock sat within each grade band. Something unexpected came out. Among the major stocks, 166 (29.4%) were piled on exactly the same point: 50.0, the very bottom of the B band.

What follows is a record of what those 166 stocks were.

Jupick's grade boundaries are fixed values. S is 85 points or higher, A starts at 70, B at 50 and C at 30, and below that is D.

The band widths work out as follows. A runs from 70 to 84, a 15-point band. B runs from 50 to 69, a 20-point band. C runs from 30 to 49, a 20-point band. D runs from 0 to 29, a 30-point band. Only S is open at the top and has no upper limit.

The band widths differ. So the five letters are not the same kind of statement. An S grade means a single threshold has been crossed, while a D grade means the score is somewhere in a 30-point band. Each is one letter on a badge, yet the amount of information it carries differs by grade.

The top score of each band is a whole number such as 84 or 69 because stored scores are integers. Scoring results are rounded down before they are stored. So the highest score in grade A is 84, not 84.9.

As of September 16, 2026, the major stocks Jupick covers were counted by market: 240 on KOSPI (Korea's main market), 124 on KOSDAQ (Korea's second, growth-oriented market), 136 on the New York Stock Exchange and 64 on Nasdaq.

By grade, there were 9 in S, 106 in A, 358 in B, 75 in C and 16 in D. B alone made up 63.5 percent of the total.

Next, Jupick calculated at what percentage point of its own band each stock sat. The bottom of a band is 0 percent, and the 100 percent point is left empty because the top threshold belongs to the next grade. Since scores are integers, even 69, the top of grade B, stops at 95.0 percent. This percentage is a number for counting in what follows, not a number shown on the screen.

The median position within the band was 40.0 percent for A, 15.0 percent for B, 65.0 percent for C and 81.7 percent for D.

B alone stands out. In C and D the stocks cluster toward the top of the band, and A sits a little below the middle, but B has a median of 15 percent. That means half of the stocks in a 20-point band are within its bottom 3 points.

Counting the stocks sitting exactly at the bottom of each band gave the answer. There were 14 at 70, the bottom of grade A, 2 at 85, the bottom of grade S, and none for C or D. But at 50, the bottom of grade B, there were 166.

The 166 stocks are 46.4 percent of the 358 B-grade stocks. More than four in ten B-grade stocks have the same score.

To check whether this is a natural distribution, the area around each of the four thresholds was counted the same way: how many stocks were in the 3 points just below the threshold, and how many in the 3 points from the threshold upward.

The 85-point threshold had 14 below and 6 above. The 70-point threshold had 38 below and 36 above, close to symmetric. The 30-point threshold had 6 below and 4 above.

The 50-point threshold had 22 below and 175 above. That is a difference of nearly eight times.

The other three thresholds show no such skew. Scores are scattered fairly evenly near those boundaries. That means something is piling up separately at 50 alone.

In each update, Jupick tags and records the metrics it could not measure. So it is possible to count, for each stock, how many of the nine metrics carry a tag.

Of the 166 stocks at 50.0, 159 had tags on all nine metrics. These are stocks for which not a single metric could be measured.

Looking at the reasons on the tags, 159 of the 166 carried the reason that the data source provided no value, and 7 carried the reason that the company did not report it. These two reasons account for all 166. The reason that the previous value was carried over and the reason that the metric does not apply to the sector did not appear even once in this pile. Lumping all missing values together hides this skew.

The split also shows by market. Of the 166, 78 were on KOSPI, 84 on KOSDAQ, 3 on the New York Stock Exchange and 1 on Nasdaq. That puts 44.5 percent of major Korean stocks at this score, against 2.0 percent of major overseas stocks.

Widening the scope to the whole universe, 160 stocks have tags on all nine metrics. Of them, 159 are at 50 points. The remaining one is at 46, the same score that stocks with fill-in values for all nine metrics received before the neutral payout ratio value was put in place on September 3, 2026. It appears to be a record that has not been recalculated since then, and Jupick is checking it separately.

The distribution of tag counts also had something unusual. There were 174 stocks with 0 tags, 127 with 1, 33 with 2, 43 with 3, 21 with 4, 2 with 5 and 4 with 6. Then there is not a single stock with 7 or 8, and 9 jumps straight to 160. Between being partly blank and entirely blank, there is nothing.

In other fields that give scores from 1 to 100, the number 50 usually carries a meaning.

The US building energy efficiency score defines 50 as the national median. It means exactly in the middle among buildings of the same use, compared against national survey data, and the score states separately that 75 or above is in the top tier. In other words, that 50 is the result of a measurement.

Jupick's 50 is the opposite. It is the number that comes out when none of the nine metrics could be measured.

Metrics with no incoming value get a constant called a neutral anchor: ROE 7.0 percent, operating margin 7.0 percent, revenue growth 3.0 percent, debt ratio 110 percent, current ratio 128 percent, interest coverage 4.2 times, dividend yield 1.75 percent, payout ratio 15 percent and 4.2 consecutive dividend years. Each of the nine values was worked out backward so that it scores exactly 50 points on its own scoring table.

So when all nine metrics are anchors, each of the three groups averages 50 points, and however the weights are multiplied and added, the total lands at exactly 50.0.

The reason for this setup is to neither reward nor penalize what the data source did not say. Putting a high score into an unknown metric creates a strength that does not exist, and putting a low score in creates a flaw that does not exist. Jupick has no plan to change this judgment itself.

On screen, though, the result looks like this. A stock that was measured and turned out to be exactly in the middle and a stock for which nothing could be measured both wear the same B 50-point badge side by side. The number is the same and the basis is the opposite.

Once the band width is known, the next question is the distance to the boundary.

For the 555 stocks outside grade S, the points remaining to the next grade's threshold were counted. Within 1 point were 28 stocks, 5.0 percent, and within 3 points were 80 stocks, 14.4 percent.

That means roughly one stock in seven stands within 3 points of the threshold of the grade above. Because scores are stored as integers, near a boundary a 1-point difference decides a grade letter. On screen, 69 and 70 differ by 1 point, yet the badges split into B and A.

This is not written down to say the boundaries are arbitrary. The boundaries apply the same values to every stock and do not change from stock to stock. It is simply better to read knowing that 1 point near a boundary looks very different on screen from 1 point in the middle of a band.

Jupick is also clear about what it does not say. Jupick does not say that any stock is about to move up or down a grade. What it writes on screen is only a subtraction between the score stored today and a fixed boundary.

Since September 16, 2026, the “where this score sits” part of the stock page shows the position within the grade band. Three numbers are written together in one paragraph: the range of the grade band in points, how many points the score is above the lower boundary and how many points remain to the upper boundary.

A bar is drawn too, but the bar shows only the relative position within that band. Band widths vary at 15, 20 and 30 points, and drawing bars of the same length while hiding the width would lead the eye into a false comparison. So the width is written separately as a number.

For grade S, the sentence is written differently. There is no grade above it, so there is nothing to measure the remaining distance to. Forcing it into a single sentence pattern would talk about a next grade that does not exist for nine stocks as if it did.

The position within the band is not written as a percentage. As noted earlier, band widths vary, so relative position alone invites a false comparison. Percentages carry the same trap, so they are used only for counting in this article, and the screen shows points.

This line pairs with another line near the top of the page. Below the overall grade, the page already states how many of the nine metrics have values. The two lines sit in different sections and apart on screen, but they are best read together.

For the 159 stocks with all nine metrics blank, the upper line reads 0 of 9 metrics with values, and the lower line states that the score of 50 equals the lower boundary. Seen together, they reveal what kind of 50 that is.

First, there was the option of redrawing the grade boundaries or splitting grade B in two. This is the solution the EU energy label actually adopted. When the top grades became saturated, it replaced the whole grading scale.

Jupick did not do this. Jupick's scoring logic and grade boundaries are fixed and are not to be changed lightly. What the European case also shows is that redrawing the boundaries breaks comparison with the past. Jupick leaves the boundaries as they are and gives the same information by stating the position within the band.

Second, there was the option of removing stocks with no measured metrics from grading or hiding their badges. This was examined against measured data on September 1, 2026, dropped, and set aside for good. Lists and comparison screens are sorted by grade, so if only some stocks had no grade the ordering would wobble, and above all, hiding is also a way of substituting for judgment.

Third, there was the option of replacing the neutral anchors with sector medians. That would scatter the pile at 50. But it would give stocks that could not be measured numbers that look measured. It runs directly against the principle of not creating numbers that do not exist, so Jupick does not do it.

Fourth, there was the option of expressing the position within the band with color, giving each band its own color the way an air quality index does. Jupick already uses grade colors to tell grades apart, so expressing position within the band with color as well would use the same axis twice. Position is expressed only with numbers and bar length.

A single letter on a grade badge covers a wider range than it seems. A is a 15-point band, B and C are 20-point bands, D is a 30-point band, and only S is open at the top. Inside the same B badge, 50 and 69 live together.

Measuring major stocks as of September 16, 2026 gave 9 in S, 106 in A, 358 in B, 75 in C and 16 in D. The median position within the band was 40.0 percent for A, 15.0 percent for B, 65.0 percent for C and 81.7 percent for D, with only B leaning toward the floor.

The cause was the 166 stocks piled at the bottom of grade B. Around the 50-point threshold there were 22 in the 3 points below and 175 in the 3 points above, a skew of nearly eight times, while the other three thresholds show no such skew.

Of the 166, 159 were stocks for which none of the nine metrics could be measured, and the reason was mostly that the data source provided no value. A total of 162 were Korean stocks, putting 44.5 percent of major Korean stocks on this single point.

Metrics without values get neutral anchors, each worked out backward to yield exactly 50 points. This is to neither reward nor penalize the silence of the data source. As a result, a stock with all nine metrics on anchors totals exactly 50.0, and a stock measured to be exactly in the middle and a stock for which nothing could be measured wear the same badge.

The distance to the boundary was measured as well. Of the 555 stocks outside grade S, 28 (5.0 percent) are within 1 point of the threshold above, and 80 (14.4 percent) are within 3 points. Because scores are stored as integers, near a boundary 1 point decides the grade letter.

So the stock page now shows the width of the grade band and the position within it. Jupick chose neither to change the grade nor to hide it, and instead states on the same screen how wide a range that grade refers to. The judgment is left to the reader.

This content is for information only and is not investment advice. Jupick does not encourage trading in any particular stock, does not forecast the future direction of prices and does not say that any stock is about to move to another grade. The grade is a summary value that converts disclosed financial metrics with a fixed formula and is no assurance of investment results. The figures cited were measured on September 16, 2026 and change with updates. Investment decisions and responsibility for them rest with the user.