Notice what the Lab never asks you: which dates to test. Every strategy is graded on the same kind of window — the two years leading up to its birth — and you can't move it. That isn't us being lazy — it's the most honest choice we could make.
You pick a coin, a timeframe, a side, and a leverage. What you can't pick is the date range — every strategy is tested on the two years leading up to its launch. On most tools that would feel like a missing feature. Here it's the feature.
Because the moment you let someone choose their own dates, you hand them a quiet way to fool themselves — and we'd rather not.
Picture a backtest tool with a start-date slider. What does everyone do with it? They drag it back and forth until the number looks good. "Start in March instead of January and the return doubles!" Nothing about the strategy changed — only the window you measured it in.
That's cherry-picking, and it's the oldest trick in the book: keep sliding until the past agrees with you. A fixed window takes the slider away. You see the strategy's record over our window, not the flattering little stretch you'd have hunted down.
There's a second reason, and it's about fairness of comparison. Every strategy card on CoinLAB is graded on a window of the exact same shape: two years long, ending just before its birth. When one card scores higher than another, they sat the same exam — same duration, same distance from the present. And the exam never quietly grows: a card born next year faces its own two years, exactly as long as the two years today's cards faced.
Two cards born months apart did meet different weather — no window design can erase that. What stays equal is the length of the test and how close it runs to the present. And the scoreboard that finally settles it is shared by everyone: the Since Launch record, written on the same live candles for every card.
Long enough to be hard. Two years of crypto is a lot of weather — rallies, crashes, and the long flat stretches where nothing works. A strategy that only ever saw one kind of market hasn't really been tested; it's easy to look brilliant in a market that only goes up. A short, cherry-picked window can hide a crash. Ours puts one right in the middle of the test.
Short enough to matter. Markets drift. A window that stretches back forever ends up grading strategies mostly on a market that no longer exists. Keeping the window at two years keeps every grade about the market as it is now — and keeps the shelf moving with it. These are strategies you rotate as conditions change, not heirlooms you hold forever.
Strategies created before mid-July 2026 were graded on an even longer window — the same two years plus the history behind them, back to early 2024. A longer exam, judged by the same gates: nothing about their record is softer. Nothing was re-scored, either — a card's backtest is frozen at creation, and it stays frozen. Every card's report prints its exact test dates, so you never have to take the window on faith.
Here's the honest truth about any backtest, ours included: the past can always be flattered. Even with a fixed, fair window, a backtest is still a story about history — and history is the easy part. So don't let it be the thing you trust most.
The number that earns your trust is on the Since Launch panel — how the same frozen strategy has done on the candles that arrived after it was born, updated every day. Anyone can show you a beautiful backtest. What's worth watching is whether the live record keeps pace with it. The fixed backtest is the fair starting line; the living track record is the real race.
A fixed window makes the comparison fair — it doesn't make the future certain. Every backtest here is hypothetical performance on past data, not a prediction, and nothing on CoinLAB is financial advice. A strategy that survived two hard years can still fail tomorrow. Always do your own research.