A backtest answers one question: if I had followed these exact rules in the past, what would have happened? The Backtest tab is Well Street's time machine for that question. Give it a strategy — no code required — and it replays market history at daily resolution, stepping through each trading day in order, checking your rules against what was known that day, and recording every buy, sell, and portfolio value along the way.
When the run finishes, you get the strategy's whole simulated life: an equity curve, performance and risk numbers, and a grade on both performance and confidence.
Before your first run, it helps to set expectations honestly, because this tool is easy to over-trust. A backtest is evidence, not a promise. It can tell you an idea would have failed. It can tell you an idea survived history. It cannot tell you the future will cooperate.
A strategy is just a set of rules: what to buy, when to buy it, when to sell, and how much to hold. Your first backtest should be a rule simple enough to say in one breath. Buy the cheapest profitable large companies and rebalance periodically. Hold stocks above their long-term average price, step aside when they fall below. Simple is not a training-wheels compromise — it's the honest starting point, for a reason that matters more the longer you do this.
The reason is overfitting. Add enough conditions and knobs and any strategy can be tuned to fit the past perfectly — and fitting the past perfectly is exactly what makes it fragile in the future. A rule with few moving parts, tested across long periods and different market regimes, deserves more trust than a many-parameter strategy that looks flawless on one stretch of history.
If strategy ideas feel unfamiliar territory, the Concepts library covers momentum, value, and friends in plain English.
Here's the quiet thing that makes a backtest trustworthy or worthless: what data the simulation is allowed to see. Well Street's backtester runs on point-in-time data, meaning every value carries the date the market actually learned it, and the simulation can only use what was truly known on each simulated day.
That blocks two famous ways backtests lie. Look-ahead bias: a company's quarter ends in March but the numbers aren't filed until weeks later, and a sloppy backtest lets an April trade use them anyway — trading on news before it existed. Survivorship bias: test only on today's stock list and you've quietly excluded every company that went bankrupt or was delisted, handing your strategy a silent tailwind by skipping history's disasters. Both can manufacture several points of fake annual return.
On Well Street, the simulation sees originally-reported fundamentals, the universe of stocks actually listed on each date — graveyard included — and prices adjusted correctly. Nobody gets to peek at the future, including us. The methodology pages document all of this in detail.
Your finished backtest reports growth and risk together, because either alone can mislead. Total return and CAGR — compound annual growth rate — tell you how fast the strategy grew. Maximum drawdown tells you the worst peak-to-trough fall along the way. Volatility and risk-adjusted measures like the Sharpe ratio relate the reward to the ride. Every one of these has a plain-English definition in the glossary.
Read drawdown first. Two strategies can both show 12% CAGR while one lost 20% at its worst moment and the other lost 55%. A 55% loss needs a 122% gain just to get back to even, and very few people hold on through that. The equity curve tells you what the strategy did; the drawdown tells you what it would have cost you to live through it.
The confidence side of your grade speaks to how much the evidence should be trusted — how robust the result looks, not just how pretty. A modest, steady result across decades often means more than a spectacular one from a short lucky stretch.
Real trading isn't free, so honest reading of a backtest can't be either. Commissions, bid-ask spreads, and slippage — the gap between the price a signal saw and the price a real order would have filled at — all compound. A strategy that trades often and gives up a fraction of a percent per round trip can quietly lose several points of annual return to friction. So when a high-turnover strategy narrowly beats a patient one, assume friction closes the gap until proven otherwise. And when comparing two results, check they make the same cost assumptions.
Remember too that markets change. A pattern that worked for twenty years can fade once it's widely known. The honest way to hold a backtest is as a filter, not a fortune-teller — past performance doesn't guarantee future results, here or anywhere.
When a result genuinely impresses you, share it on the Street with the evidence attached and let others poke at it. That's the whole idea of this place: strategies checked against history, in the open, together.
← All guides · Investing concepts library