Screens · Price change
Biggest Losers and Percentage Rankings
Percentage change is the right measure for a losers list, and the list still misleads, because a split, a spin-off or a special dividend produces exactly the same number as a collapse.
The measure is right, which makes the defect interesting
Most of the screens on this site are compromised by their choice of measure. A most-traded list ranked by share volume is substantially a list of cheap stocks; a points-based price ranking would be a list of expensive ones. A losers list ranked by percentage change has neither problem: the percentage is scale-free, and an instrument at four dollars and one at four hundred are compared fairly.
So this is the best-constructed of the common screens, and its remaining defect is the subtlest one in this section. The percentage is computed from a price series, and a price series can fall for reasons that are not a decline.
The four cases in full
| Instrument | Previous | Now | Change | What happened |
|---|---|---|---|---|
| Failed trial result | $18.40 | $6.10 | -66.8 % | Also genuine, and the largest real fall of the four. |
| 3-for-1 split | $96.00 | $32.00 | -66.7 % | Nothing happened. Every holder has three times as many shares at a third of the price. |
| Profit warning | $41.20 | $27.60 | -33.0 % | A genuine decline: the business disappointed and the stock repriced. |
| Spin-off distribution | $74.50 | $55.80 | -25.1 % | Value left the listed entity as shares in another company, distributed to holders. |
The worst genuine fall here is the failed trial result at -66.8 per cent, and the split prints -66.7 per cent. On an unadjusted series the two are indistinguishable, and the split ranks above one of the real declines. This is the one screen defect on this site that produces a plausible wrong answer rather than an obvious one, which is what makes it worth a page of its own.
Three checks that separate them
Use adjusted prices. An adjusted series divides historical prices by the split factor and accounts for distributions, which removes almost all of this automatically. Most data providers offer both series and default to one without saying which, worth establishing once.
Look at the volume. A split requires nobody to trade: the share count changes overnight and the following session can be perfectly ordinary. A genuine repricing after news arrives with several times the usual turnover. So a large fall on unremarkable relative volume is the signature of an artefact, and it is the fastest of the three checks.
Check the corporate-action calendar. Splits, spin-offs, special dividends and rights issues are announced in advance and dated. This is the check that is certain rather than suggestive, and it takes a moment for a single instrument.
The other population the list collects
Corporate actions are the interesting failure; the ordinary one is the same problem every screen in this section has. A percentage move is easier to produce on a small, low-priced, thinly traded instrument.
At eighty cents a share, one or two ticks is a ten per cent move. On a wide spread, a print at the bid following a print at the ask can show a several-per-cent fall with no news and no change in what anyone thinks the instrument is worth. And a single small trade at a low price can set a closing print on an instrument that barely traded all day.
All three are removed by the gates the screen-design page puts in front of any ranking: common stock only, a price floor, and a dollar-volume minimum. None of them is sophisticated, and together they turn a list of microcaps and rounding into a list of instruments that were actually repriced.
The mirror list is not symmetric
A biggest-gainers list is built by the same arithmetic and collects a different population, which is worth knowing before the same interpretation is applied to both.
Percentages are asymmetric to begin with: a fall from 100 to 50 is −50 per cent, and the recovery from 50 to 100 is +100 per cent, so the same two prices produce a much larger number on the way up. A gainers list is therefore dominated by instruments that had already fallen a long way; the percentage is computed from a small base. Add the populations that produce large one-day rises, mostly small speculative issues and squeezes in heavily shorted names, and a gainers list skews further towards the low-priced end than a losers list does.
The corporate-action problem is milder on that side, though it exists: a reverse split multiplies the price and shows as an enormous gain in unadjusted data. It is rarer than a forward split, which is exactly why it catches people; the check nobody runs. Adjusted prices handle both, and the volume check works identically: a reverse split requires nobody to trade.
Halts, and the falls that are missing
One more thing worth knowing, because it works in the opposite direction: the largest declines in a market are sometimes absent from the day’s list altogether.
An instrument halted pending news does not print a closing price in the ordinary way, and when it reopens, possibly the following session, the whole move arrives in one step. So a losers list is not a complete ranking of the day’s falls: it omits the instruments where something happened that was serious enough to stop trading, and then reports their move on a later date in a size that belongs to two sessions. Anyone treating the list as a census of the day’s worst outcomes is missing exactly the cases that were worst.
Frequently asked questions
Is percentage change the right way to rank a losers list?
Yes. It is scale-free, so it does not simply return the most expensive instruments the way a points ranking would. That makes this the best-constructed of the common screens, which is why its remaining defect is worth a page: the percentage is computed from a price series, and if that series has not been adjusted for corporate actions, some of the largest falls on the list did not happen.
Which corporate actions look like a collapse?
Anything that changes the share price without changing what a holder owns. A split is the clearest: a three-for-one split shows as a fall of two thirds in raw prices while every holder is exactly as well off. A large special dividend, a spin-off distribution and a rights issue all do the same thing to a lesser degree. Each produces a plausible-looking percentage decline and none of them is a decline.
How would I tell the difference?
Three checks, in order. Use an adjusted price series, which handles almost all of it automatically. Look at the volume: a split does not require anyone to trade, so a large fall on ordinary volume is suspicious, while a genuine repricing arrives with several times the usual turnover. And check the corporate-action calendar for the instrument, which is published and takes a moment.
Why does the list fill with small, low-priced issues?
Because a percentage move is easier to produce when the price is small and the instrument is thinly traded. A stock at eighty cents moves ten per cent on a tick or two, and one with a wide spread can appear to fall several per cent between a bid-side print and an ask-side one with no news at all. Add a price floor and a liquidity gate before ranking and most of that disappears.
Should the ranking use the close or the intraday low?
The close, unless you have a specific reason otherwise, and whichever you choose has to be stated. A ranking by intraday low returns instruments that traded briefly at a price nobody could actually have sold much stock at, which is a different question, and on a thin instrument the low can be a single small print. Two lists built on the two definitions disagree substantially, and neither is wrong.
Do trading halts affect the list?
They affect what is missing from it. An instrument halted for pending news does not print a closing price in the usual way, and when it reopens the move arrives in one step that may span two sessions. So the largest falls in the market are sometimes absent from the day’s list entirely, and appear on the following day’s in a size that belongs to two sessions.
Is a big fall a buying opportunity?
The screen cannot tell you, and this is where it is most often misused. A percentage decline is a description of what the market decided, and the reason for the decision is not in the number: a profit warning, a failed trial, a legal outcome and a split all look the same on a losers list. Sorting them requires reading about the company, at which point the screen has done its actual job, finding something to look at.
What does a well-built version of this list look like?
Adjusted prices, common stock only, a price floor, a dollar-volume gate, closes rather than intraday extremes, and a note beside any instrument with a corporate action on the date. That is five decisions and a data-quality requirement, and once they are made the list is a genuinely useful description of where the largest repricings happened.