Glossary

The terms of the analysis, explained plainly. To see how they come together into a score, read the methodology.

Scores and signals

Score (0 to 100)

A single figure given to each asset, combining technical analysis and financial soundness. The higher the score, the more favourable the setup is judged at the time of the analysis.

What it does not tell you. It does not predict tomorrow's price and knows nothing of your situation: it reads the set-up on the day of the analysis, and can change from the next session.

Signal: Strong Buy / Buy / Neutral / Avoid

The score sorted into four levels, from most to least favourable. A fragile signal (a one-day spike, a trendless market, extreme overbought) is moved down a level: see Robustness. Beginner mode writes it in plain words (Buy, Watch, Avoid). It is a summary, not an instruction or advice.

What it does not tell you. It says neither when to act nor how much to invest, and a favourable level guarantees no rise.

Robustness

A 0 to 100 rating of how reliable the signal is, not how strong: how many days it has held, whether it is an isolated spike, whether the trend is real (ADX), whether the holding is overbought (RSI) or very volatile. Low robustness caps the signal, even when the score is high.

What it does not tell you. It does not say whether the move will continue: a robust signal can still reverse; it has only held longer, on less fragile ground.

Technical / fundamental weighting

The share of each analysis in the score. By default 65% technical and 35% fundamental, adjustable from 100% technical to 100% fundamental to suit your style.

What it does not tell you. It does not make the score more accurate: it chooses what matters most to you. Changing the weighting changes the score, not the stock.

Technical score

The part of the score that comes from prices. It is a sum of steps: each condition met (trend, momentum, volume, Ichimoku, market regime) adds its points, and the score window shows which ones are earned. The ML bonus is added afterwards.

What it does not tell you. It reads only past prices and volumes: it ignores the accounts, the valuation and company news.

ML bonus

An adjustment to the score from machine-learning models trained on the history of each broad zone (Europe, United States, Japan, crypto, commodities). It is added to the technical score, up or down, depending on the probability the model returns.

What it does not tell you. It knows only what the history contains: an unprecedented event escapes it, and a probability is not a certainty.

Fundamental score

The part of the score that comes from the company's accounts: valuation, profitability, debt, growth. Not computed for commodities or crypto.

What it does not tell you. It rests on the latest published accounts, so on the past, and compares each ratio with fixed bands, not with its sector: a P/E that is normal for a bank can be low for a software firm.

High growth

A 0 to 100 score, separate from the main score, kept for companies that combine strong revenue growth, volume backing the price, a score that is not negative and debt under control. A company that does not meet every condition has no score, which is not the same as zero.

What it does not tell you. It does not say whether the growth will last or whether the price paid is reasonable: a company can grow fast and already be expensive.

Technical indicators

Trend

The underlying direction of the price, read from the moving averages: the price above its 200-day average, the 50-day average crossing above the 200-day (golden cross), and the way they slope.

What it does not tell you. It describes the recent past, not the future, and reacts late: a moving average confirms a reversal after it has begun.

Momentum

The strength and speed of the price move, read from the RSI and the MACD histogram. Positive momentum means an upward dynamic is under way, without saying how long it will last.

What it does not tell you. It does not say how long the momentum will last, and strong momentum can come just before it fades.

RSI (Relative Strength Index)

An oscillator from 0 to 100 measuring whether an asset has risen or fallen too fast. Often read as overbought above 70 and oversold below 30.

What it does not tell you. Above 70 it does not announce a fall: in a strong trend it can stay there for a long time. It says nothing about what the company is worth.

MACD

The gap between two moving averages, and the average of that gap. It spots changes of dynamic: crossings, and moves above or below zero.

What it does not tell you. Built on averages, it lags the price, and it gives false crossovers in a market going nowhere.

Bollinger bands

An envelope around the price based on volatility. A price near the upper band is stretched to the upside, near the lower band to the downside.

What it does not tell you. Touching a band is not a reversal signal: in a trend, the price can ride the upper or lower band for weeks.

ATR (Average True Range)

A measure of volatility: the average size of the moves. It is used in particular to place a protective stop in proportion to the holding's risk.

What it does not tell you. It measures range, not direction: a high ATR does not say whether the stock is rising or falling.

OBV (On-Balance Volume)

Volume accumulated according to the direction of each session. A rising OBV suggests accumulation, real buying interest behind the move.

What it does not tell you. It adds up volumes without knowing who is buying: one large one- off volume, such as an index rebalancing, can move it with no real interest behind it.

Ichimoku

A set of lines forming a cloud. Where the price sits relative to the cloud gives a quick reading of the trend and of support.

What it does not tell you. The cloud is still drawn from past prices: it sees neither results nor news, and follows a trendless market poorly.

ADX

It measures the strength of a trend, not its direction. A high ADX means a clear trend, a low ADX a market going nowhere.

What it does not tell you. It does not give the direction of the trend: a high ADX comes with a steep fall as readily as with a steep rise.

Slope of the 200-day average

The way the long moving average points. A positive slope confirms an underlying uptrend.

What it does not tell you. It changes slowly: by the time it turns, much of the move has often already happened.

Volatility

The size of the price swings. High volatility means wider moves, so both more risk and more potential.

What it does not tell you. It does not say which way the price will move, and a calm period can come before a large swing.

Market regime

The general context (a clear trend or a market going nowhere, calm or jumpy) which adjusts the score to avoid false signals.

What it does not tell you. It describes the market on the day of the analysis: a calm market can turn jumpy overnight.

Fundamentals

P/E (price / earnings)

The share price divided by earnings per share: how many years of current earnings you are paying for. A low P/E can reflect a reasonable valuation, or a market expecting earnings to fall. It is compared between companies in the same sector.

What it does not tell you. It says nothing about future growth or debt, and means nothing when earnings are negative. The score compares it with fixed bands, not with its sector.

PEG (P/E / growth)

The P/E divided by expected earnings growth. It adjusts the P/E of a fast-growing company: around 1, the price paid is in line with the growth. It rests on an analyst forecast, so on an estimate.

What it does not tell you. It depends on a growth forecast that can prove wrong, and means nothing if expected growth is zero or negative.

EV/EBIT

Enterprise value (market value plus net debt) divided by operating profit. Close to the P/E, but it accounts for debt: two companies on the same P/E, one indebted and one not, do not have the same EV/EBIT.

What it does not tell you. Like the P/E, it looks at a single year of profit: an exceptional result, up or down, distorts it.

P/B (price / book value)

The price divided by the book value of equity. Below 1, the market values the company at less than its net book assets.

What it does not tell you. Book value does not say what the assets are really worth: it often leaves out brands, patents and software, and can overstate outdated assets.

ROE (return on equity)

Earnings divided by shareholders' capital. A high ROE means a company that puts its equity to good use.

What it does not tell you. High debt can inflate it without the company being more profitable, and it says nothing about how steady the earnings are.

D/E (debt / equity)

The level of debt relative to equity. The higher it is, the more the company depends on its borrowings.

What it does not tell you. It says neither whether the debt is expensive nor when it falls due, and some sectors, such as banks or utilities, normally run with a lot of debt.

Revenue growth

The rise in sales from one year to the next. Sustained growth is a driver of valuation.

What it does not tell you. Selling more does not mean earning more: it says nothing about margins or profitability, and an acquisition can inflate it at a stroke.

Factors (asset allocation)

Six angles scored 0 to 100 by rank within the chosen universe: valuation, technical score, low volatility, profitability, fundamentals and free cash flow, growth. Missing data removes the factor for that holding rather than giving it a score.

What it does not tell you. A rank is relative to the chosen universe: first in a weak universe is not necessarily strong, and the rank changes if the universe changes.

Decision and validation

Backtest

A simulation of a strategy over past data, to measure what it would have given. It measures the past; it does not predict the future. The site reruns it every week.

What it does not tell you. A good past result guarantees nothing: market conditions change, and a simulation never exactly reproduces real costs and execution.

Control

What a measurement is compared against: the same dates, the same market, without the condition being tested. In a rising period any selection looks like a winner; only the gap with the control says whether the condition adds anything.

What it does not tell you. Beating the control over one period does not prove the gap will repeat: it takes enough observations, and the period may be unusual.

Drawdown

The largest fall from a peak to the trough that follows, as a percentage. It is the loss you would have lived through by getting in at the worst moment; a return is always read next to its drawdown.

What it does not tell you. It does not say how long it took to get back to the peak, or whether the next fall will be deeper.

Correlation

Between -1 and +1: how closely two series rise and fall together day to day. +1 means identical moves, 0 no link, -1 the exact opposite. It does not say that one causes the other.

What it does not tell you. It changes over time: two stocks barely linked in calm markets can fall together in a crisis. And it measures whether moves go the same way, not how large they are.

Sample size (n)

How many observations a figure rests on. The same result over 30 days and over 300 are not worth the same; below a floor, the site publishes nothing.

What it does not tell you. A large sample does not make a measurement right if the observations all look alike: a thousand sessions of one rising market are still a single context.

Percentile

Where a value ranks among every value in the history. 96th percentile means it has been exceeded only 4 times out of 100 in the past.

What it does not tell you. It places a value within the past; it does not say whether it will return toward the average, or when.

ATR stop-loss

An exit level placed below the price, at a distance proportional to volatility (ATR), to cap the loss on a position. The site watches it during market hours and alerts you; it never places an order.

What it does not tell you. It does not guarantee the exit price: a stock that opens sharply lower can gap straight through the level. And the site alerts you; it does not sell for you.

Profit target

The exit level aimed for on the upside. Combined with the stop and a maximum holding time, it defines a position's rules in the backtest.

What it does not tell you. It is an exit rule, not a forecast: nothing says the price will reach it.

Sector rotation (RRG)

A chart placing each sector by its strength relative to the market and how fast that strength is changing. Four quadrants: leading, weakening, lagging, improving. Sectors usually rotate in that order, with no guaranteed timing.

What it does not tell you. The rotation has no timetable: a sector can stay in one quadrant for months, or skip one.

News sentiment

The tone, positive or negative, of recent articles about a holding. When it is clear-cut it adds or removes a few points from the score; a neutral or unknown tone changes nothing.

What it does not tell you. It reads the tone of articles, not their accuracy or importance: a very positive article can be about a minor piece of news.

Candlestick pattern

A shape drawn by one to a few candles (hammer, engulfing, morning star…). The site does not treat it as a signal: it measures what followed it in the history, against a control, and stays silent when there are too few cases.

What it does not tell you. A pattern announces nothing on its own: its measured effect is often small, and the site stays silent when there are too few past cases.

Read the methodology Create an account