Most trading plans are wish lists. "I will be disciplined. I will only take high-probability setups. I will manage risk and let winners run." Every line is agreeable and none of it is a plan, because none of it can be checked. A plan you cannot check against your own trades is not a plan, it is an affirmation, and the market does not care about your affirmations. The test of a trading plan is whether it survives a bad Tuesday, and the only plans that do are the ones written in terms specific enough to catch you breaking them.
So build the plan backwards from that test. Every rule should be phrased so that, after the fact, your journal can tell you plainly whether you followed it. If a rule cannot be graded, rewrite it until it can.
A plan is a set of rules that produce a tag
Here is the mental model that makes plans useful instead of decorative. Every rule in your plan should map to a field you can record on a trade. Not a sentiment, a field. "Only trade with the higher-timeframe trend" becomes a with-trend or against-trend tag on every entry. "Risk no more than 1% per trade" becomes a logged risk figure you can sort by. "No trading in the first fifteen minutes of the session" becomes a timestamp you can check. When your rules are shaped like fields, your journal becomes a scorecard for your own discipline, and you lose the ability to lie to yourself about how well you followed the plan.
This is the connective tissue between planning and journalling, and it is why the two subjects are really one subject. If you have not set your journal up to capture the context of a trade rather than just its result, the plan has nothing to grade against, which is the case argued in keeping a trading journal that works .
What actually goes in the plan
The setups, defined tightly enough to be falsifiable. Not "I trade breakouts" but the specific conditions that make a breakout one of yours: what has to be true before, at, and after entry. If two traders reading your setup definition would disagree about whether a given chart qualifies, it is too loose. Tighten it until they would agree.
Entry, stop, and target rules a stranger could execute. The test is mechanical: could someone else, handed your plan, take roughly the trade you would have taken? If your entries depend on a feel you cannot put into words, that feel is not yet a rule, and it will not survive pressure.
Risk per trade and per day, as hard numbers. A fixed fraction of the account per trade, and a daily loss limit that ends the session. These are the rules that keep a bad day from becoming a bad month, and they are the first ones traders abandon in the exact moment they matter most, which is why they belong in the plan in ink.
The conditions under which you do not trade. Most plans list reasons to enter and forget to list reasons to stand aside. Name them. Major news you do not trade through, market states your edge does not work in, personal states (tired, tilted, distracted) that historically wreck your numbers. A plan that only says yes is half a plan.
How you size, and whether it varies. If you size up on higher-conviction setups, define the tiers and what qualifies for each, so that "high conviction" is a category with criteria and not a mood you assign after you have already decided to take the trade.
Base it on evidence, not aspiration
A plan invented from scratch is a hypothesis, and most hypotheses are wrong. The plans worth trading are built on evidence that the setups actually pay. That evidence comes from one of two places: your own journalled history, if you have enough of it, or testing before you risk money.
If you already have a track record, mine it. The setups that earn a place in your plan are the ones your data shows carry positive expectancy over a decent sample, which is the edge-finding exercise applied forward. If you do not have the history, or you want to try something new without paying tuition to the market, test it first. You do not need to code to do this, and the mechanics of testing a strategy on past price without writing a line of Python are laid out in backtesting a strategy without code . Either way, the setups in your plan should have earned their place, not just sounded good in your head.
The plan has to name its own success metric
A plan without a scoreboard drifts, because you have no way to tell whether it is working or whether you have simply grown attached to it. Pick the metric before you start, and make it expectancy on a rolling sample, not your account balance and not your win rate. Balance is too noisy and win rate is too easily gamed by cutting winners early. Expectancy over a meaningful number of trades tells you whether the plan produces an edge, and the reasoning for why that is the number to trust is spelled out in the expectancy explainer . Write the target into the plan so future you cannot quietly move the goalposts.
Plan adherence is its own metric
This is the part that separates a plan that survives from one that gets quietly abandoned. Grade your adherence separately from your results. On every trade, log whether it was in-plan or off-plan, honestly, at the time. Then, once a month, compare the expectancy of your in-plan trades against your off-plan ones.
One of two things will be true, and both are worth knowing. Either your in-plan trades outperform your off-plan ones, in which case you now have hard evidence that following the plan makes you money, which is far more persuasive at three o'clock on a losing afternoon than a promise you made to yourself. Or your off-plan trades are quietly outperforming, in which case your plan is wrong about something and your instincts have found an edge it has not captured yet, so you should study those off-plan trades to work out what the plan is missing. Logging the habit of plan adherence is exactly what the Diary in TradeSave+ is for, because a behaviour you record every day is one you can eventually measure.
Revise on evidence, not on emotion
A plan is a living document, but there is a right and a wrong way to change it. The wrong way is to tear it up after a losing streak, because losing streaks are guaranteed even for good plans, and a plan you rewrite every drawdown is not a plan. The right way is to change it when the data, over a real sample, tell you a rule is not doing its job. Raise the bar on a setup whose lowest-conviction tier keeps losing. Cut a session that never pays. Add a stand-aside rule for a condition that keeps hurting you. Every revision should trace back to something you measured, not something you felt after a bad week. A plan built that way does not just survive contact with the market. It gets sharper every time it makes contact.