How Long Should You Paper Trade a Strategy?

Paper trade a strategy long enough to see it make a meaningful number of decisions and to live through at least one rough patch, which for most strategies means somewhere between one and six months. A strategy that trades every hour can show its behaviour in weeks; one that decides once a month needs far longer to say much at all. No paper test is long enough to prove future profits, so the real goal is to check behaviour, not to confirm returns.

The rest of this guide explains how to pick a length that fits the strategy in front of you, and what you can and cannot learn in that time.

Why there is no single right answer

People often ask for a number, such as "30 days" or "3 months." The trouble is that time is the wrong unit. What actually matters is how much evidence you collect, and evidence comes from three things:

1. The number of decisions the strategy makes. Each trade or rebalance is one observation.
2. The variety of market conditions it sees. Ten trades in a calm, rising market tell you less than ten trades spread across a rise, a fall and a sideways stretch.
3. What you are trying to learn. Checking that software works takes far less time than judging whether a strategy has an edge.

A month of paper trading might give one strategy 150 decisions and another just one. Those two tests are not comparable, even though the calendar says they are the same length.

Count decisions, not days

A practical first step is to estimate how many decisions a strategy will make in your test window. Here is how that looks for AlgolaneBots's five strategies over a 90-day free trial, using how often each one checks the market (checks are not always trades; a check can end with "keep holding").

The counts are rough, because markets close on weekends and holidays, and gold trades nearly around the clock on weekdays. The point is the scale. A three-month window is a deep look at an hourly strategy and a very shallow look at a monthly one.

You can read what each of these strategies does on the strategies overview.

What different test lengths can tell you

One to two weeks: does it work at all?

The first weeks are about plumbing. Does the strategy trade on the days it should? Does it buy what its description says? Are positions sized sensibly? Are there any errors or odd fills?

A week or two is usually enough to catch setup problems. It is nowhere near enough to judge profit or loss.

One to three months: does it behave as described?

With a daily or weekly strategy, a few months gives you dozens of decisions. You can start checking whether its behaviour matches its description. If a strategy claims to step aside in weak markets, you may get to see it do that, or you may not, depending on what the market does during your window.

This is the range where most people form a first impression. Treat that impression as provisional.

Three to twelve months: how does it feel through ups and downs?

Longer tests are more likely to include at least one meaningful drop. That matters, because the moment most people abandon a strategy is during a drawdown, not during a good run. Living through one on paper, and checking whether you would have stuck with it, is valuable.

Beyond a year: still not proof

Even a year of paper trading is a small sample compared with the decades of history a backtest can cover. It is real forward evidence, which backtests can never provide, but it is limited evidence. We compare the two approaches in backtest vs forward test.

Why a paper test cannot cover a full cycle

It helps to be honest about the timescales involved. Real strategies go through losing stretches that can last far longer than any reasonable paper test.

Consider Gold Breakout at 2% risk. In its historical backtest from June 2004 to November 2026 (a computer test on past prices, not money anyone earned), it returned +15.1% a year against +11.0% a year for simply holding gold. Its worst fall was 41%, compared with 45% for gold held over the same period. That deepest fall ran from May 2018 to October 2019, and the strategy did not get back to its previous high until March 2023, 41 months after the bottom. Its published caveat is that doubling trading costs cuts its yearly result roughly in half. Future results will differ.

The lesson is not about gold specifically. It is that a strategy can spend years underwater even when its long-run backtest looks reasonable. No paper test will sit through every phase of a strategy's history, so you need the backtest for the long view and the paper test for the present-day view. Our article on how long strategies take to recover looks at this in more detail.

Signs your paper test has run long enough

Rather than watching the calendar, look for these signals:

• You have seen a reasonable number of decisions. For a daily strategy, dozens; for an hourly one, hundreds.
• You have seen at least one losing stretch. If the test has only gone up, you have not yet learned the most important thing.
• Behaviour matches the description. The strategy did what it said it would do, in the conditions it met.
• Fills and costs look realistic. Spot checks against real prices do not reveal anything surprising.
• You can explain every trade. If you cannot say why a position was opened, keep watching until you can.
• You know what you would do in a bad month. Write it down before you go live.

Signs you should keep testing, or stop

Keep going if the market has been unusually calm, if the strategy has barely traded, or if anything about its behaviour still surprises you.

Stop and rethink if the strategy did something its description does not explain, if paper fills look far better than real prices would allow, or if a modest paper loss already made you uncomfortable. That last one is useful information about you, not just the strategy.

A list of common traps during this period is in paper trading mistakes that fool beginners.

A simple way to plan your test

1. Write down what you want to learn. "Does it work?" and "Would I stick with it?" need different test lengths.
2. Estimate decisions per month using how often the strategy checks the market.
3. Set a minimum. For example, at least 30 decisions or three months, whichever is longer. These are illustrative numbers; pick ones that make sense for your strategy.
4. Set a review date in your calendar so you judge the test deliberately, not on a whim after a good or bad day.
5. Compare against a benchmark over the same window. If a stock strategy rose 3% while the S&P 500 rose 6%, that is a different story from rising 3% while the index fell.

How AlgolaneBots's free trial fits

AlgolaneBots's free trial lasts 90 days and runs all five strategies on a $10,000 paper account. It starts when you press start, not when you sign up, so you can create an account and begin the clock when you have time to pay attention. No card is needed, and the trial never turns into a charge by itself. After 90 days you can subscribe for $29 a month if you choose; see pricing.

Ninety days is a solid look at the hourly and daily strategies and a light look at the weekly and monthly ones. If you are mainly interested in a monthly strategy, be clear with yourself that three months will mostly show you that it runs correctly, and lean on the backtest, with its caveats, for the longer picture.

Key takeaways

• Measure a paper test by decisions and market conditions seen, not just by days elapsed.
• One to two weeks checks the plumbing; one to three months checks behaviour; longer tests show how drawdowns feel.
• A monthly strategy makes only a handful of decisions in three months, so its paper test proves little about returns.
• Historical backtests show losing stretches that lasted years, far beyond any practical paper test.
• Decide in advance what you want to learn and when you will review the results.

See it for yourself

The quickest way to find the right test length for you is to start one. Begin a free 90-day paper trial of AlgolaneBots's five strategies, note how often each one acts, and review them on a date you choose. Read the risk disclosure before considering real money.