2. Price, volume and return: the three variables you will see on every chart
Level 1: Foundations of Algorithmic Trading
Everything a quantitative trader does is built, sooner or later, on three variables. Understanding them properly matters more than knowing any particular strategy, because they are the raw material from which all strategies are built.
Price is the most basic variable: the amount at which an asset changes hands between a buyer and a seller at a given moment. It is set by the tug of war between supply and demand: when more people want to buy than to sell, the price rises; when more people want to sell than to buy, it falls. Price is public data, available on any platform, and it is the basis of both technical analysis and the models that try to predict future movements.
Volume is the total number of shares or contracts traded over a given period. On its own it says little, but combined with price it is highly revealing: a price rise on heavy volume suggests a strong uptrend backed by many participants; the same rise on thin volume is more fragile and may reverse soon.
Return is the percentage change in the price of an asset between two points in time. Quants prefer to work with returns rather than absolute prices for an important technical reason: price is not stationary, meaning its mean and variance change over time, which complicates any statistical model. Returns, by contrast, behave far more stably and allow you to compare the performance of completely different assets, or of the same stock across different eras, on equal terms.
When researchers study thousands of daily return series across stocks, they find patterns that repeat again and again, so consistently that they are known as the 'stylized facts' of financial returns. Three of them matter especially to any trader: returns have far fatter tails than a normal distribution predicts, meaning extreme moves occur more often than basic statistical intuition suggests; volatility tends to cluster in time, with calm periods following calm periods and turbulent periods following turbulent ones; and the linear correlation between one day's return and the next day's is practically zero, which is one of the reasons why predicting the direction of price with simple models is so hard.
These three facts are not academic curiosities: they are the reason strategies that 'should work' on paper fail in practice, and the reason machine learning has become so relevant in quantitative finance — it is especially good at detecting the complex, non-linear patterns these stylized facts conceal.