A setup clears a solid-looking backtest and the temptation is to start trading it live immediately, at normal size, the very next session. This is the single most common way a genuinely promising setup gets a worse live track record than it deserves — not because the setup was wrong, but because the transition from backtest to live skipped several checks that exist specifically to catch the ways backtesting and live trading differ.
Check one: does the sample size actually clear the threshold
Before anything else, confirm the backtest sample is large enough to trust — generally 30 to 50 trades minimum, more for setups with high result variance. A promising 15-trade backtest is a hypothesis worth continuing to test, not a setup ready for real capital; the win rate and expectancy at that sample size are still unstable enough that a live-money graduation is premature regardless of how good the numbers currently look. This is exactly why mysetuplap's Strategy Lab gates a hypothesis's verdict behind an explicit minSample — a strategy under its sample threshold stays INCONCLUSIVE by design, not SUPPORTED, no matter how strong the raw numbers look in a small sample.
Check two: does expectancy beat your baseline, not just zero
A positive expectancy alone isn't the bar — a setup showing +0.3R expectancy sounds fine until it's compared against a trader's baseline expectancy across everything else they already trade, which might be +0.5R. A new setup that underperforms your existing baseline isn't worth adding complexity for, even with a technically positive number. The comparison that matters is against your own baseline, not against zero — exactly the comparison mysetuplap's Strategy Lab builds in explicitly, tracking baselineSample and baselineExpectancyR alongside every strategy test rather than judging a setup in isolation.
Check three: has the setup been forward-tested, not just backtested
A backtest, by definition, is tested against data you already know the outcome of — which creates room, even unintentionally, for hindsight to influence exactly where entries and exits get marked. A forward test — paper-tracking the setup in real time, on unfolding price action, without knowing the outcome in advance — is a meaningfully different and more honest test, closer to the actual conditions live trading will present. A setup that only exists as a historical backtest, never forward-tested even briefly, is skipping a step that specifically catches setups whose backtested edge was partly an artifact of knowing the answer in advance.
Check four: is the setup fully defined in writing, not just proven in aggregate
A backtest can show a positive expectancy while the actual entry rules used to generate it remain loosely defined in the trader's head — meaning the "edge" that was proven belongs to a slightly different, better-specified version of the setup than what a trader will actually be able to execute consistently, live, under pressure. Before going live, the setup needs a real Playbook entry — explicit entry rules, invalidation, minimum RR, required confirmations — precise enough that the backtest and the live version are provably the same setup, not approximately related ones.
Check five: start smaller than your normal size, on purpose
Even a setup that clears every check above deserves reduced size for its first live trades — not because the backtest was untrustworthy, but because live execution introduces variables (slippage, hesitation, the psychological weight of real capital) that no backtest fully captures, and the first live trades are effectively still part of the testing process, just with real money attached. Sizing down deliberately during this phase limits the cost of discovering an execution gap the backtest couldn't have shown.
Putting the checklist together
Before any setup graduates from Backtesting to full-size Live Journal trading: confirm the sample clears 30-plus trades, confirm expectancy beats your actual baseline (not zero), confirm the setup has been forward-tested even briefly, confirm it's written down as a real Playbook entry with explicit rules, and start live execution at reduced size. None of these checks are complicated individually — the value is in running all five, in order, before capital is at real risk, rather than skipping straight from "this backtest looks good" to full-size live trading on the strength of a number that hasn't yet been tested against everything live trading is about to add to it.
