Methodology
RankEq gives every company in its universe a score from 0 to 100, every week. This page says what that number is, what goes into it, how a feature earns a place in it, and what a strategy does with the ranking. It is written to be read in order, but each section stands on its own.
The score
A RankEq score is a percentile. A company at 90 scores higher than 90% of the 6,007 companies in the universe on the combined evidence the model weighs; a company at 100 is at the very top that week. Scores are shown with one decimal because at the top the first decimal is the whole difference between two companies.
The score is refreshed every week after the weekend, on the latest financial statements, analysts' estimates and prices. It moves gradually, because the data underneath it does: a company rarely crosses from the bottom to the top of the list in one step.
Beneath the overall score sit seven category scores, each also 0–100, so you can see where a company is strong or weak. Beneath those sit hundreds of individual features. The categories are public vocabulary; the features are not (see what it is not).
The universe
The ranking covers 6,007 companies in 6 countries. A company is in the universe when it has a primary listing (no depositary receipts, dual listings or preferred shares), at least three years of continuous price and fundamental history, and enough daily trading to be bought and sold. Companies that fail those tests are not scored at all rather than scored badly.
The universe is not forced to be sector-neutral or style-balanced. If the evidence favours one industry in a given period, the ranking says so; a strategy can add its own limits on top.
Seven categories
Every feature belongs to one of seven categories. Each category asks one question about a company; its score answers it on a 0–100 scale.
Growth
Is the business getting bigger, and is the growth real?
Revenue and gross profit growing steadily and faster than peers; operating income, earnings per share and free cash flow rising with them, so the growth is funded by the business rather than by leverage.
Valuation
What are you paying for it?
Price against earnings, cash flow, sales and enterprise value, judged against the company's sector. A modest entry multiple, on several measures at once, at the moment of purchase.
Quality
Does the business earn its keep?
High and stable returns on the capital it employs, margins that hold up when the environment changes, and profits that turn into cash.
Stability
Can the accounts be trusted?
A balance sheet that moves in step with the income statement and with its own history. Sudden jumps in tangible assets, or shifts out of line with sales, count against a company.
Sentiment
What is the market saying about it?
Changes in analysts' estimates, short interest and similar signals of what other investors expect, which tell you where opinion is moving before the numbers do.
Technicals
What do price and volume say?
Information the financial statements do not carry: which industries are being re-rated, when a move has overshot, how calm or erratic a share price is, and where trading is arriving.
Size
Is anyone else looking?
Smaller and less-followed companies, which large institutions cannot use and which therefore are more often mispriced. Size is where much of the ranking's edge lives, and where trading costs matter most.
How a feature earns its place
A feature is one measurable idea about a company, such as a growth rate or a valuation multiple. Hundreds are tested; only those that pass every test below are used, and a feature that stops passing is removed. This is the part of the method that matters most, and it is deliberately strict: it is easy to find a pattern that worked once.
- Works across time. The feature's payoff has to hold in each of several multi-year periods since 1999, not just on average over all of them.
- Works across regions. It has to work in North America and in Europe separately, so a feature cannot earn its place by tilting toward the region that happened to run.
- Works in random sub-universes. The universe is split into random fifths and the feature is judged in each. A payoff carried by a handful of names fails here.
- Is not driven by outliers. Payoffs are measured with the extremes trimmed, and a feature whose whole edge sits in a few extreme observations is out.
- Survives trading costs. Every payoff is measured net of the cost of trading it, at the size of account it is meant for. A feature that only works if trading were free does not work.
- Has economic logic. There has to be a reason a feature should predict returns. Patterns without one are the ones that stop working first.
The same tests are applied when RankEq changes its own ranking and when a Plus member adds a feature of their own: a member's feature is measured the same way, at the member's account size, before it can carry weight in a ranking.
From scores to a ranking
Each feature is turned into a percentile within the week, oriented so that higher is better, and the features in a category are combined into the category score. The overall score combines the categories. RankEq's own ranking weights them by what the tests have shown each to be worth, net of trading costs; a member's ranking uses the weights they choose.
No single category can dominate. A company that is cheap but shrinking, or growing but paying for it with a stretched balance sheet, scores in the middle, which is the point: the ranking rewards companies that look sound from several directions at once.
A score is not a prediction of next week's return. It is a statement about the evidence today, which is why strategies hold groups of high-scoring companies rather than betting on one.
Strategies
A strategy is a rule for holding the top of a ranking: a fixed number of the best-scoring companies, weighted toward the top of the list, rebalanced every week. The rules are set once and followed; nothing is decided week to week by hand.
Rebalancing rules limit trading: a position is not touched for a change of less than about one percent, a position may drift some way from its target before it is trimmed, and the whole book is only rebalanced when its drift adds up. Costs are modelled from the liquidity of each name, so a strategy for a small account and one for a large account do not hold the same companies.
RankEq publishes strategies on its own rankings and tracks each one live against the S&P 500 from the day it went live; every earlier point on a chart is a backtest and is marked as such. Plus members build strategies of their own, on a ranking they set the weights for, and test them the same way before following them. A member who follows a strategy gets a weekly list of the trades that would bring their account in line with it; the member places every trade themselves.
Backtests
Backtests in the Strategy Lab are historical simulations over point-in-time weekly rankings since 1999: roughly 6,000 global companies a week, delisted companies included, so there is no survivorship bias. Each simulated week runs Monday close to Monday close, never overlapping.
Trades fill at real prices.A buy or sell executes at the next session's average of its high, low and twice its close, with a backup session for holidays. A company that did not trade the next session cannot be bought that week, and broken price bars are refused rather than trusted.
Trading costs are charged, three ways.Liquidity tiers charge each traded dollar an all-in rate by how thinly the company trades; flat basis points charge one rate on turnover; and size-aware pricing uses the square-root impact law (volatility times the square root of the trade's share of the day's volume, on top of an estimated spread), so the same strategy shows an honest curve at youraccount size. No position may use more than 15% of a company's daily volume per execution day. The exact rates:
| Median daily $ volume | Liquidity tier, per side | Size-aware spread, per side |
|---|---|---|
| ≤ $15K | 5.5% | 2.5% |
| $15–50K | 4.0% | 2.0% |
| $50–100K | 3.0% | 1.7% |
| $100–350K | 1.5% | 0.7% |
| $350K–1M | 1.0% | 0.5% |
| $1–5M | 0.5% | 0.3% |
| > $5M | 0.2% | 0.1% |
Both models add $0.01 per share, which bites low-priced stocks, and a name with missing liquidity or price data costs a prohibitive 6% per side. Non-US listings carry per-country spread additions, from +0.1% round trip in the Nordics to +1.7% in Hungary, and +0.7% for the UK including its 0.5% stamp duty.
The remaining constants, in the open.Universe hygiene drops closes under $1.10 and names trading fewer than about 100 shares a day. Per-name volatility is its own trailing 26-week realised figure. Drift-band rebalancing uses 50% position drift, 15% portfolio drift and a 1% minimum transaction. Cash earns nothing, the same-rules benchmark is the equal-weighted eligible universe, and multi-day fills assume the week's move accrues evenly over its five sessions. Every one of these exists because a too-good-looking number was chased to its cause and fixed; the Lab's job is to compare strategies honestly, never to flatter them.
What it is not
Not advice.A score, a ranking or a strategy on RankEq is the output of a quantitative model, published for information. It is not a recommendation to buy or sell any security and it does not take your circumstances into account. Every trade a member makes is the member's own decision.
Not a forecast.A strategy's backtest is a simulation of the past under stated rules and costs. It says how the rules would have done; it does not say how they will do. Past performance, simulated or real, does not predict future results, and every strategy has periods, sometimes long ones, of trailing the market.
Not fully public. The categories, the tests, the universe rules and the backtest mechanics are described here in full. The individual features and their weights are not: publishing a formula spoils it for the members who pay for the ranking. What a member sees is the seven category scores; what is public for a company is its band in the ranking, never its exact score.
No leverage, no shorting, no derivatives. Strategies hold long positions in listed shares and nothing else, and the liquidity rules exist so that every position can be sold.
Questions
- What is a RankEq score?
- A number from 0 to 100 for every company in the universe, refreshed each week. It is a percentile: a score of 90 means the company scores higher than 90% of the universe on the combined evidence of hundreds of tested features in seven categories. It is model output, not a forecast and not a recommendation.
- What do the seven categories mean?
- Growth (is the business getting bigger), Valuation (what you pay for it), Quality (returns on capital and cash conversion), Stability (whether the accounts hold together), Sentiment (what estimates and positioning say), Technicals (what price and volume say) and Size (how overlooked the company is). Each has its own 0–100 score; the overall score combines them.
- Which formulas are behind the scores?
- The categories are public; the individual features inside them are not. Publishing a formula spoils it for the members who pay for the ranking, so RankEq describes what each category measures and keeps the leaf formulas private.
- How often does the ranking change?
- Every week, after the weekend, on the latest financial statements, estimates and prices. Scores move gradually because the underlying data does; a company rarely jumps from the bottom to the top in one week.
- Is a score of 100 a buy signal?
- No. A high score says the company looks strong on the evidence the model weighs. Whether to own it depends on your strategy, your account size, the cost of trading it and everything the model does not see. RankEq strategies hold groups of high-scoring companies under fixed rules; nothing on RankEq is advice about a single stock.
- What is a RankEq strategy?
- A rule for holding the top of a ranking: a fixed number of the best-scoring companies, weighted toward the top, rebalanced weekly under rules that limit trading. RankEq publishes several strategies and tracks them live; Plus members can build their own from a ranking they set the weights for.
Get Access Today.
See every stock ranked, follow a strategy — or build your own and test it on 27 years of history.
★★★★★
“Like X-ray lenses for the market.”“It's like slipping on X-ray lenses for the market — observe every company through the Ranking, see the example Strategy, and mirror the top stocks on your own.”
Free
PER MONTH
- ✓The e-book's ranking on every US stockThe method from the e-book, run weekly · shown two weeks behind
- ✓ The free Example StrategyHoldings are shown with a two-week delay for free users
- ✓Build a strategy: RankEq's strategies as publishedTheir track records and books, sized to your account · nothing to run
- ✓ The e-book, free articles, the feed and the community
- ✓ Broker linking — $2/mo add-onAt cost, one broker · included free with Membership
Membership
Most popularPER MONTH
or $390 yearly — 2 months free
- ✓ Everything in Free
- ✓ Four rankings, liveThe e-book's on every US stock, and the three behind the large-cap strategies on the stocks each strategy picks from
- ✓ Access to paid articles
- ✓ Three strategies of big US companies
- ✓ No capacity limit — join any time
- ✓ Rebalance alerts by email
- ✓ Broker sync + one-click rebalancing
- ✓ Follow any strategy a member shares
- ✓ One ranking of your own, simulated on the US universeThe standard RankEq book on our data partner's simulator, paid per run · follow it like ours
- – Building your own features, and the global universe, start with Plus
Plus
PER MONTH
or $1,990 yearly — 2 months free
- ✓ Everything in Membership
- ✓RankEq's own ranking, with every factor scoreUS and international, every stock above $500k a day — the universe the microcap strategies pick from
- ✓ Build your own: unlimited features, 3 rankings, US and globalFeatures tested on 27 years of data, paid per run at cost (quick from $0.25, thorough under $1 each in a batch) · rankings simulated on the US or the global universe, quick on our data partner's simulator or in full on our engine · swap and share them · keep them private for $100/mo
- ✓ A seat in the microcap strategies, as capacity allowsSmall companies, where the edge is largest — and room is limited, so a full strategy means a waitlist
- ✓ Run several strategies at once
- ✓ Blended into one target portfolio
- ✓Recommended trades & rebalancing
- ✓ Cash-flow rebalancing on new deposits
7-day free trial · 30-day money-back guarantee · cancel anytime