How many measured values sit under a grade: in a count of major stocks, 150 (37.5%) had all nine metrics filled

The first thing on a stock page is the overall grade badge. One of five letters from S to D appears in large type, with a score out of 100 next to it.

That letter comes from scoring nine metrics. On the profitability side: ROE, operating margin and revenue growth. On the financial stability side: debt ratio, current ratio and interest coverage. On the dividend side: dividend yield, payout ratio and consecutive dividend years. Each metric is converted to a score from 0 to 100, and the averages of the three groups are weighted 40% for profitability, 35% for financial stability and 25% for dividends, then added up.

The catch is that the nine metrics are not always filled. For some stocks the data source supplied all nine in this update; for others it supplied none. A metric with no value gets a fill-in value, and scoring runs as usual.

Yet on screen both stocks appear with the same single badge. To tell which was which, a reader had to scroll down to the area-by-area scoring and count by eye how many hyphens were shown. The badge sits at the top of the page and the hyphens far below it, so in practice most people did not count.

So the first step was to measure how common this situation actually is.

Using the September 2, 2026 update, Jupick counted, for each stored major stock, how many of the nine metrics were measured in that round.

All nine were measured for 150 stocks. Eight were measured for 148. Together that is 298, about three quarters of the total. For most stocks, at most one metric is blank.

The rest split as follows: seven for 29 stocks, six for 45, five for 20, four for 1 and three for 5. No stock had only one or two measured.

And 2 stocks had none of the nine measured. Both are listed on KOSPI, Korea's main market, and both carried a grade of C with a score of 46.

The matching numbers are no coincidence. With no measured values, all nine metrics are filled by the same rule, so the result converges on the same score whatever the company does. Why it is 46 in particular becomes clear only from how the fill-in values are set.

Why a single blank metric is so common, and which metric it tends to be, differs by market. That is covered in a separate article.

When a metric does not come in for a given round, Jupick does one of two things.

First, if a value measured in the previous round is stored, Jupick keeps it. The number there is not a fill-in value but one that was actually measured last time. It has not, however, been confirmed again in this round.

Second, if there is no measured value to bring back, Jupick inserts a constant called a neutral anchor. These values are not arbitrary: each was worked backward so that it scores exactly 50 points on its scoring curve.

ROE 7.0%, operating margin 7.0%, revenue growth 3.0%, debt ratio 110%, current ratio 128%, interest coverage 4.2 times, dividend yield 1.75%, payout ratio 15% and 4.2 consecutive dividend years. Each of the nine scores 50 points on its own scoring curve.

Of these, the payout ratio anchor stayed empty for a long time and was filled only on September 3, 2026. Why it came late and what happened in the meantime is described below.

The reason for 50 points is to give neither reward nor penalty for what the data source did not say. Inserting a high score for an unknown value invents a strength that is not there; inserting a low score pins on a flaw that is not there.

Using the sector median as the anchor was also considered and dropped. Inserting the median amounts to saying the company does at least as well as a typical large company, and that too is ultimately a made-up value.

Each anchor was worked backward to score 50 points. So with no measured values, all three groups average 50 points, and however the weights are applied, the total should be 50.0 points.

Yet the two stocks counted on September 2 scored 46. The arithmetic does not add up. The reason is that, of the nine metrics, only the payout ratio had no anchor.

At the time the dividend group looked like this: dividend yield 50 points and consecutive dividend years 50 points, but payout ratio 12.5 points. The average drops to 37.5 points. Applying the weights gives 50 times 0.4 plus 50 times 0.35 plus 37.5 times 0.25, for a total of 46.875 points. The display rounds down, so it shows 46, and since the threshold for B is 50 points, the letter becomes C. The two stocks' C 46 was exactly this number.

In other words, that grade did not come from measuring the company; it came from how the fill-in values were set up. If a hundred stocks had been in the same state, all hundred would have been C 46.

Where those 3.125 points came from and how they were removed comes next.

There was a reason only the payout ratio lacked an anchor.

The other eight metrics each have a set direction, where higher is better or lower is better, so exactly one value scores 50 points. The payout ratio is different. Too low means stingy returns to shareholders and too high means paying out more than earnings, so it is scored with about 35% as the ideal and points deducted the further it moves away.

Under this band-style scoring, two values score 50 points: 15% and 55%. Choosing one of them as the default looked like dressing up a coin toss as a default, so no anchor was set.

The problem was that setting no anchor meant scoring the metric as 0%. That is 35 away from the ideal of 35% and scores 12.5 points. In trying not to toss a coin, Jupick was using a default far lower than either candidate. That runs directly against the principle of not penalizing what the data source did not say.

On September 3, 2026, this value was filled. Instead of choosing by preference, Jupick used numbers it had already counted. Among stored major stocks, 324 (81%) actually had a payout ratio, with a median of 31.19%. More than half, 55.2%, sat below the ideal of 35%. Measured from the median, 15% is closer than 55% to the group of stocks Jupick covers. So the anchor is 15%.

As long as a blank is filled with a single constant, this choice cannot remove arbitrariness entirely. But the value was chosen by numbers Jupick counted, not by a coin.

Now a stock with all nine metrics as anchors totals 50.0 points rather than 46.875, and its letter is B rather than C. No score was created out of nothing; the 3.125 points that were being taken away because data was not supplied have disappeared.

A payout ratio of 0% for a company that pays no dividend is a real 0%, so no anchor goes there and it is scored as 0. The anchor is used only when whether the company pays a dividend at all could not be confirmed, or when it pays a dividend but the payout ratio line alone did not come in.

In the September 2, 2026 round, a line was added beneath the overall grade. Its name has changed since then, and it now gives the number of “metrics with values” out of nine.

When all nine have values, only that line appears, with nothing added, because there is nothing more to say.

If even one is blank, the line adds why the remaining ones did not make the count. If no metric has a value, the sentence is written differently altogether: it says the grade was calculated without any metrics that have values. Zero of nine is not nine metrics each falling slightly short; it is a different kind of state.

The sentence does not say the grade was calculated with neutral values, for the reason seen above. Some of the flagged metrics hold not a neutral anchor but the previous round's measured value. Calling that value neutral would say something false about that number.

For a while, the words “in this update” were also dropped from this line, because one more flag was added on September 4, 2026. A metric that does not apply to a sector in the first place, such as interest coverage for financial companies, stays unfilled in the next update too, and wording it as a circumstance of this round implies that waiting is enough. The only fact the four flag reasons shared was that the number cannot be used as is, so only that was stated.

For documents stored before flag records existed, the line is hidden entirely. Writing that all nine have values for a document that cannot be counted would promote the unknown to measured. The fact of not knowing is left blank rather than covered with a number.

On September 30, 2026, the name of this line was changed. Until then it read “metrics measured this update,” but the counting rule was narrower than whether a value was measured. A company whose debt ratio turned negative because of capital impairment, meaning its equity fell below zero, is an example. Jupick confirmed the value, so it is measured, but a negative number cannot be ranked as high or low, so it is left out of this count. Metrics were dropping out this way while the name spoke as though it counted measured values. So the name became “metrics whose numbers can be used as is,” a name that says exactly what the counting rule decides. It says “can be used” rather than describing past use for the same reason: what the rule decides is whether the number qualifies for use, not where and how the number actually went in. The matching line in the area-by-area scoring and the verdict wording in the per-metric coverage table were fixed at the same time, because sentences sitting side by side on one screen must not read as though they count different things.

That evening, the stock description paragraph was moved to this rule as well. That paragraph alone was counting flagged metrics, so for a company with a negative debt ratio and one more flag, the line said seven while the paragraph said eight. For a company with no flags and only a negative value, the paragraph treated all nine as used as is and said nothing. The passage that called the remaining metrics ones the data source did not supply was fixed too, since a negative value is one the source did supply, so the wording was wrong and also implied that waiting would fill it. After that, the paragraph said only that those numbers cannot be used as is, because that was the one statement true of both the four flag reasons and negative values.

There is precedent for this choice. The Bank of Korea publishes an annual statistical compendium, Financial Statement Analysis, that explains how to read its tables. There, blank cells are marked with different symbols by reason. “Not applicable” and “unknown” get different symbols, and a separate slot is kept for “not calculable.” The Bank of Korea describes the not-calculable case as one where the denominator of the ratio formula is negative. The denominator of the debt ratio is equity, so a company with capital impairment is exactly this case.

Another passage in the same compendium is clearer still. It says companies with zero or negative equity are excluded from ranking. They are excluded not because the value is unknown but because ranking them would distort interpretation. Jupick leaves such metrics out of the count for the same reason, which is why the name was also moved away from talking about measurement.

On October 12, 2026, the line was renamed again. “Metrics whose numbers can be used as is” fit the counting rule, but readers found it hard to follow and no stock page explained it. So it was changed to the everyday phrase “metrics with values.” The sentence that said only that the remaining metrics cannot be used as is was also changed to one that gives the reason. The sentence differs depending on whether the item was missing from the source data, whether the data was not received in this update, or whether the value does not arise given the nature of the sector. The words “in this update” return only for reasons where a value can arrive in the next update. When metrics with different reasons are mixed, they are not lumped under one reason; only wording true of all of them is used. A company whose debt ratio turned negative from capital impairment has a value, yet that metric does not make the count, which clashes with the new name. In that case the line adds that the measured value is negative and its size cannot be compared. The matching line in the area-by-area scoring was renamed along with it, and from the same day the financial summary paragraph on the page also groups the remaining metrics by reason.

The display idea came from an overseas service that aggregates review ratings. When there are too few reviews to produce a rating, it turns the icon itself gray. The shared point is stating, right next to the number, that the sample is thin. Jupick, however, left the grade letter itself untouched.

Nine of nine means all nine metrics are values confirmed in this round, entered into scoring as is with no fill-in values. It is no assurance that the grade is right. It is a statement about what inputs the grade stands on.

Eight of nine or seven of nine is the most common state. Which metric is blank shows up as a hyphen in the area-by-area scoring below. Under the hyphen, the reason it is blank is also given.

At zero of nine, the letter on the badge is better not read as an assessment of the company. As seen above, that value is a number produced by how the fill-in values were set up.

When this number is low, the cause usually lies in the collection path. Some companies are small and disclose fewer items, some lack the metric altogether because of how they are listed, and in some cases a free data source does not supply the item. In these cases a low count does not mean the company is in poor shape.

Not every case is about collection, though. A company whose debt ratio turned negative from capital impairment drops out of this count even though Jupick confirmed the value. Here the low count does point to the company's condition. The way to tell the two apart is to read the reason given in the area-by-area scoring below. If it says not disclosed or not applicable, it is a collection matter; if it says negative value, it is about the company.

One option was to hide or remove the grade letter for stocks with few measured values. That was not done this time.

There are two reasons. Lists and comparison screens sort by grade, so if only some stocks lacked a grade, the ordering would wobble. And hiding is itself a judgment made on the reader's behalf. The moment a threshold is set for how thin is thin enough to hide, stocks just above and just below it get entirely different treatment.

Dropping missing metrics from scoring altogether was also considered earlier and set aside. That is covered in a separate article.

So the choice this time was neither to hide nor to remove, but to say. The grade stays as it is, and what it stands on is stated in the same place. The judgment is left to the reader.

A count of major stocks on September 2, 2026 found 150 with all nine metrics measured in that round and 148 with eight. Together that is three quarters of the total. Seven were measured for 29 stocks, six for 45, five for 20, four for 1 and three for 5, and 2 stocks had none measured.

A metric with no incoming value either keeps the previous round's measured value or gets a neutral anchor set to score exactly 50 points on its scoring curve. The point is to neither reward nor penalize the data source's silence.

Only the payout ratio lacked an anchor. Because it is scored around an ideal of 35%, two values score 50 points, 15% and 55%, and choosing one of them seemed groundless. So when this metric was blank it was scored as 0%, or 12.5 points, and a stock whose nine metrics were all fill-in values came to 46.875 points, displayed as 46, and sat at C.

On September 3 this value was filled too, using the measured distribution of stored major stocks instead of preference. The 324 stocks (81%) that actually had a payout ratio had a median of 31.19%, and 55.2% sat below the ideal of 35%, so of the two candidates 15% is closer to the group Jupick covers. A stock whose nine metrics are all fill-in values now scores 50.0 points, grade B. No score was created out of nothing; the 3.125 points that were deducted because data was not supplied have disappeared.

So the number of metrics with values now appears beneath the overall grade badge. With nine, only that line appears; with fewer, it says why the rest did not make the count; with none, it states that the grade was calculated without any metrics that have values.

For documents stored before flag records existed, the line is hidden. Leaving the unknown as unknown is more accurate than writing that something uncounted was fully counted.

This content is for information only and is not investment advice. Jupick gives no trading advice and no price forecasts. A grade is a summary value that converts disclosed financial metrics with a fixed formula and is no assurance of investment results. The measured-value counts cited were taken on September 2, 2026 and the payout ratio distribution on September 3, 2026, and they change with updates. Investment decisions and responsibility for them rest with the user.