An average across the whole market hides as much as it reveals. The 90/80 Rule works on the headline numbers - but a sensible investor doesn't just want to know that it works. They want to know where it works, because that's where you put your attention and, just as usefully, where you don't.
The uncomfortable truth is that the edge is deeply uneven. It is enormous in some corners of the market and non-existent in others. Four lenses bring that map into focus: size, sector, factor and - the one most people never think to check - geography. Let's take them in turn.
Size: a sub-£350m phenomenon
This is the single biggest constraint on the strategy, and it's worth being blunt about it up front.
The jump effect is a small and mid-cap phenomenon. For +10 to +40 point jumps, 12-month mean returns ran a healthy 13% to 31% in companies under £350m - and a flat ~5% in companies above it, no matter how big the jump was (before dividends of course). Below the £350m line, a bigger jump buys you a bigger return: the signal scales. Above the line, the signal is simply gone - a +40 leap in a large-cap earned about the same limp 5% as a +10 creep. The edge doesn't fade politely with size. It falls off a cliff.
The explanation is one Peter Lynch would have recognised on sight. Lynch ran Fidelity's Magellan fund to legendary returns by buying small, dull, under-followed companies before Wall Street woke up to them, and his whole philosophy rested on a single asymmetry: a company covered by thirty analysts has no secrets left to find, while a company covered by none is full of them. A large-cap that jumps has already been dissected by a research industry; the good news is in the price. A neglected small-cap that jumps has been noticed by almost nobody, so the price has room to drift up for months as reality slowly dawns on the few people watching.
So the edge is liquidity-constrained alpha - it exists precisely because nobody large is looking, and nobody large is looking precisely because these shares are too small and too illiquid for institutional money to bother with. That is the catch and the opportunity in one breath. The returns are real, but they live in shares that are less liquid and hard for institutions to trade.
The first filter, then, is a hard one: small and mid-cap only, roughly £10m to £350m. A large-cap jump is interesting context and occasionally a beautiful story - but it is not a very tradeable signal.
Sector: mind the postcode
The edge clusters in certain neighbourhoods of the market and is scarce, or outright negative, in others. Here are 10-plus-point QVM jumps measured against same-sector incumbents - the fair test, because a health stock should be judged against other health stocks, not against banks.
The top of that table is the quality end of the market. Healthcare and Technology jumps were the strongest, with hit rates of 65% to 71% - when a neglected health or tech small-cap re-rates hard, it tends to mean it, because the thing that changed was usually a real improvement in the business. Industrials were reliably positive too, and being the workhorse sector of the UK small-cap market, they supply a lot of the signals.
Utilities tops the edge table, and it's tempting to make a headline of it - but hold your fire. That +12.7% rests on just 40 events across twelve years and heavily dependent on a few big winners like Yu Group. Treat it as suggestive, not as a rule.
The bottom of the table is a warning. Financials are the single largest source of jump events in the entire study - and the only sector with a negative edge. A financial stock that jumped into the top decile actually underperformed the financials already sitting there by 2.3%. Energy is barely better: its jumpers scraped a positive return, but only because the sector's base rate is itself negative, so beating it is no great honour. Between them, Financials and Energy were where the left tail lives - the levered lenders and the boom-and-bust resource names that turn into the -90% disasters. A simple "exclude Financials and Energy" filter would have removed about 27% of all events and a disproportionate share of the worst losses.
Of course, that's all historic, and the future may be very different. I would be careful not including these sectors just because the past behaved like this - but it's useful information.
Factor: it's a quality story
The StockRank blends three ingredients - Quality, Value and Momentum - into one 0-to-100 score. So when a rank jumps, which of the three is actually doing the work? We ran the same study on each individual factor rank, and the answer came back clean.
The jump effect is a Quality story. A +20 to +40 point jump into a Quality rank of 90+ beat quality incumbents by +11.1%, and it is statistically significant - the one factor result that clears the bar with room to spare. Momentum and Value, by contrast, showed no reliable edge at all.
The story the table tells is the intellectual heart of the whole series. A genuine, sudden step-change in the quality of a business - margins turning, cash flowing, returns on capital jumping - is a real event that the market is slow to fully price. Analysts anchored to their old models nudge forecasts up only partway, the company beats again, and the rank drifts higher for months. That is a signal.
A jump powered by Value is a different animal, and a treacherous one. Value jumps carried the highest delisting rates in the entire study, 9% to 13%, a median return of exactly zero and a hit rate below a coin toss. Look at what a Value jump often is: a rank rising not because the business improved, but because the share price fell, which mechanically makes the stock look cheap. That is the classic value trap - a company quietly fading, its falling price flattering its Value score - arriving dressed up as a buy signal. The 13.4% of these that delisted are the ones that finished the job.
Put the three lenses together and the map sharpens considerably. The richest hunting ground is a small or mid-cap, in Healthcare, Technology or Industrials, whose jump is powered by quality rather than by a falling price. That isn't a stock screen I'm handing you to run blind - it's the contour map of where the edge concentrates, so you know which jumps to lean into and which to treat with more suspicion.
Geography: the neglect premium
If this effect is real, would the world's biggest, smartest, most heavily-arbitraged market have competed it away? So we asked directly, and re-ran the entire study on the US main boards - NYSE, Nasdaq and AMEX - over 2014 to 2026, nearly 8,000 decile jump events.
The result is one of the most revealing in the whole project. The pattern is there in America - jumpers beat incumbents, the drift shows up, quality does the work, size still matters. But it runs at roughly half the magnitude.
Read that carefully, because it says two things at once. First, the mechanism is universal - the behavioural forces behind the drift, the anchored analysts, the reluctant re-rating, the holders who sell winners too soon, are human, not British, so the effect appears wherever there are companies and the people who price them. But second, the deeper, faster, more crowded US small-cap market arbitrages the edge away far quicker. More eyes, more capital and more algorithms are hunting the same inefficiency, so it gets closed before you can harvest it.
Which means the UK edge is, substantially, a neglect premium. It is large in Britain not because Britain is special, but because British small-caps are ignored - thinly covered by analysts, largely un-owned by institutions, invisible to the global money that would otherwise compete the drift away. And here is the part that should change how a UK private investor sees themselves: neglect is your structural advantage. The very thing that makes these shares too small and too obscure for the big funds is what leaves the edge on the table for you. You are not competing with the giants in this corner of the market - you are fishing in the pond they've decided isn't worth their time. It is the same lesson as the size lens, seen from a different angle: the edge lives wherever the crowd isn't.
That is also why the performance figures in this study are more UK-specific. The mechanism travels though the magnitude does not.
Reading the map
Four lenses, one consistent message. The jump effect is not a broad market truth to be applied everywhere - it is a concentrated one, and knowing where it concentrates is most of the skill. It lives below £350m, in the quality-led sectors, in jumps driven by improving businesses rather than falling prices, and in markets the world has stopped watching. Everywhere the crowd gathers, the edge thins.
Contour maps are useful. But nothing makes a finding land like watching it play out in real, named companies - the triumphs and the disaster alike. Let's look at some case studies.
Next: Chapter 6 - Case studies. Rolls-Royce, Kromek and Yü Group on the winning side - and Carillion, the loser we keep in to keep ourselves honest.