A backtester is a machine for answering one question: would this have worked? The uncomfortable truth is that most backtesting tools are built to make the answer look like yes. Polished datasets that forget failed companies, single-stock tests that dodge portfolio reality, results pages that lead with the best-looking number — each choice nudges the answer toward flattery.
Well Street's backtester is built on the opposite instinct. The entire platform exists so strategies can be judged on evidence instead of vibes — and that only works if the evidence is allowed to say no. Every design decision below follows from that one commitment.
The most common style of backtest tests one rule on one stock: buy this ticker when an indicator crosses a line. It's easy to build and easy to game — run it across enough tickers and something will look brilliant by luck alone.
Real investing doesn't work one ticker at a time. You hold a portfolio; positions interact; winners offset losers; rebalancing forces decisions. So Well Street simulates at the portfolio level: multiple assets, position sizing, scheduled rebalancing, and portfolio-wide metrics — total return alongside volatility, Sharpe ratio, and maximum drawdown.
The difference shows up immediately. A strategy that looks heroic on its best ticker often looks ordinary across a portfolio — and that ordinary number is the honest one, because a portfolio is what you'd actually have held.
Most historical data quietly describes the past as we know it today — restated earnings, cleaned databases, delisted failures removed. Test on that and your strategy gets information no investor had, which manufactures returns out of hindsight.
Well Street's simulations run on point-in-time data: on each simulated day, the strategy sees only what was actually known that day. Original filings rather than restatements. The full universe of stocks that existed then — including the ones that later went to zero. No peeking at tomorrow.
It's less glamorous than a highlight reel, and that's the point. The methodology pages explain the machinery in detail; the philosophy fits in a sentence: a time machine only tells the truth if it can't see the future.
Two strategies can post the same return and deserve completely different trust. One earned it steadily across decades and market regimes; the other got lucky in one stretch, or was tuned until it fit the past perfectly. A single performance number can't tell them apart — which is why Well Street grades backtests on performance and confidence together.
Confidence asks the skeptical questions: how long is the evidence, how consistent is it across different periods, how sensitive is the result to the exact parameters chosen, how bad was the worst stretch? A strategy with modest returns and high confidence is often a better find than a spectacular backtest that survives only in one carefully chosen window.
And always read drawdown before return. A strategy that fell fifty-five percent at its worst needed more than a double just to recover — and very few people hold through that, whatever the end-of-chart number says.
The last piece of the philosophy is social: evidence improves when other people can poke at it. Backtest results on Well Street can be shared to the Street feed, where the community sees the same honest numbers you saw — the drawdown included — and can question the assumptions, suggest variations, or run their own versions.
That's the culture the whole system is built to support: strategies proposed openly, tested honestly, and judged on what history actually says. A good backtest here isn't a sales pitch; it's a claim with receipts.
And the standing caveat applies to everything above: a backtest is evidence about the past, not a promise about the future. Markets change, edges decay, and past performance does not guarantee future results. We'd rather show you that plainly than let a flattering chart imply otherwise.
← All guides · Investing concepts library