Quant Investing for Beginners: How It Works

In a coin-flip test, the best of 1,000 random rules beat holding US stocks by 1.9 points a year on average from 1927 to 1976, before costs. On new data from 1977 to August 2026 the same rules trailed by 3.9 points a year.

That is the central trap in quantitative investing: a rule picked from a large search can look skilled and be luck. This guide explains how quant investing works, its four parts and how to read a claim.

Quantitative investing, or quant investing, picks investments with statistics, computer code and data. The rules are written down, tested on past prices and followed the same way each time.

A rule might say: hold the cheapest tenth of stocks measured by the ratio of book value to market value. Code applies it to thousands of stocks in the same way every month. Systematic vs discretionary investing compares this with deciding case by case.

Systematic investing is the wider idea of investing by fixed rules. Rules-based investing covers the same ground. Quant investing is one way to write the rules.

Algorithmic trading is different. It automates how orders are placed, while quant investing decides what to hold.

Most quant strategies have four parts.

Trend following and momentum investing are two published signals built from past prices.

A factor is the return gap between stocks that score high and low on one signal. The value factor is built from company accounts.

Book value is a company's assets minus its debts as the accounts record them. A stock with a high ratio of book value to market value is a value stock. One with a low ratio is a growth stock.

Kenneth R. French's Data Library publishes the factor as high minus low (HML). The library refers to Eugene Fama and Kenneth French's 1993 paper "Common risk factors in the returns on stocks and bonds" (Journal of Financial Economics) for the description.

Each June, US stocks are split in two by size and into three book-to-market groups, making six portfolios. The size split is at the median of stocks listed on the New York Stock Exchange (NYSE). The book-to-market splits are at the NYSE 30th and 70th percentiles. HML is the average return of the two value portfolios minus the average return of the two growth portfolios, before costs.

Overfitting means fitting a rule so closely to past data that it learns noise. To show it, each rule here flips a coin every month.

Heads holds the US stock market and tails holds one-month Treasury bills, loans to the US government. A coin has no skill, so any edge a rule shows is luck.

The rule that returned most from 1927 to 1976 (600 months) was kept and judged on 1977 to August 2026 (596 months). The test was repeated 200 times for 1, 10, 100 and 1,000 rules. The chart shows the average.

The market made 8.63% a year in the first window and 12.07% in the second. Trying more rules improved the past. The best of 1,000 beat the market in all 200 repeats from 1927 to 1976, against 4 of 200 for a single rule.

Afterwards none of the 800 rules kept at any size beat it. The gap was about 4 points at every size.

A coin-flip rule spends about half its time in Treasury bills, so even one unchosen rule trails a stock market that beat bills. The point is that searching 1,000 rules gave no better result afterwards than one.

Out-of-sample testing keeps some data back. The rule is chosen on the first part and judged once on the second.

What backtesting is shows one such run on an S&P 500 fund. A golden cross tested against a coin-flipping monkey uses chance as the yardstick.

The chart shows the value factor's average yearly return by decade, from 1927 to 2025. Over the 99 calendar years it averaged 4.4% and was positive in 57 of them.

The 1940s averaged 9.9% and the 2000s 8.3%. The 2010s averaged −2.2%, with 3 positive years of 10 and a compound loss of 23.3% over the decade.

A signal that the whole record supports can still lose for a decade. A test over 99 years does not say which decade comes next.

A claim is a measured record over a stated window. Four questions show how much weight it can carry.

How to read a backtest goes through a report line by line.

In this test the best of 1,000 rules beat the market by 1.9 points a year on the data that chose it. On new data it trailed by 3.9 points. The value factor averaged 4.4% a year over 99 years and averaged −2.2% a year in the 2010s.

This explainer is part of the YX Insights Academy. Each one takes a single idea and checks it against real data.

The same approach runs through everything else we publish:

Good places to start on the website:

Quant investing picks investments with statistics, computer code and data instead of personal judgement. The rules are written down, tested on past prices and followed the same way each time. It is one way to invest by fixed rules, which is called systematic investing.

A quant strategy has four parts. It starts with data, then a signal that scores each asset, then a rule that turns scores into holdings, then a test on past data. The test is the step that most often goes wrong, because trying many rules finds luck.

Overfitting means a rule is fitted so closely to past data that it learns noise. In the test here, the best of 1,000 coin-flip rules beat the market by 1.9 points a year on the data that chose it. On new data it trailed the market by 3.9 points.

No: quant investing decides what to hold using models run on data, while algorithmic trading automates how orders are placed. A quant strategy can use algorithmic trading to place its orders, but the two answer different questions: what to hold and how to trade it.

DISCLAIMER: This article is strictly educational. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Any opinions, analyses, or probabilities expressed in this note are those of the author as of the note's date of publication and are subject to change without notice.