Write a Crypto Plan and Define Testable Rules

An idea that cannot be described precisely is difficult to test honestly. The rule must identify what information is observed, when it becomes available and what action follows. This lesson builds a paper-test specification and distinguishes a short-term trading hypothesis from an asset thesis or a recurring purchase policy.

Education only. Crypto assets can lose substantial value or become worthless. No lesson requires buying an asset, depositing money or sharing real passwords, private keys or recovery words.

Why write a plan before you need it

A couple in Guadalajara planning a wedding agree a budget before visiting any venue. In a beautiful hall with a smiling salesperson, saying no is much harder. The calm decision at the kitchen table protects them from the excited one.

Professionals use the same idea. CFA Institute's 2010 guide to investment policy statements for individual investors describes a written policy that sets out objectives, risk management and how a portfolio will be governed. CME Group's education makes a similar point: define your risk rules, such as a maximum loss per trade, before you place a trade.

You have already written two pieces. In Level 0 you made a learning plan, and in the recurring purchase and records lesson in Level 8 you drafted a portfolio-risk policy. This lesson folds both into one wider personal crypto plan, adds the psychology tools from Lesson C9.1, and shows you how to review it honestly.

The analogy has a limit: a wedding budget is spent once, but a crypto plan runs for years through rallies, crashes and life changes, so it needs reviews as much as rules. Start with the reason behind it all.

Start with purpose and time horizon

A train ticket only makes sense once you know where you are going. Level 0 separated using, owning, investing and trading. Your plan's first line names which one you are doing, and why. "To learn how self-custody works with a small amount" leads to very different rules from "to hold a long-term position alongside my pension", and both differ from "to trade short-term".

The second line is your time horizon: how long you expect to hold before you might need the money. Money you need within a year or two, for rent, fees or a deposit, does not belong in an asset that can fall roughly 75% in a year, as bitcoin did from November 2021 to November 2022. A long horizon does not make crypto safe, but it reduces pressure to sell at the worst moment.

It is also fine to write "Purpose: to understand crypto; I am not investing yet." Choosing not to invest is a complete plan.

With a purpose and horizon written, the next lines set how much money the plan can ever touch.

Set money limits you can check

The sizing and exposure lessons in Level 8 taught you to set a loss budget, and the recurring purchase and records lesson in Level 8 to write it into a portfolio-risk policy. Copy those figures into your plan rather than re-deciding them, and add one line you can check at a glance: a maximum allocation.

A maximum allocation is the largest share of your investable savings that crypto may ever reach, written as a percentage and as a cash figure. Investable savings means money left after your emergency fund and essentials, which stay out completely, as Level 0 showed. Your figure is personal. Regulators such as the UK FCA warn that you should be prepared to lose all the money you put into crypto, and that you are highly unlikely to be covered by a compensation scheme if something goes wrong.

Money limits say how much. The next lines say what you will and won't do with it.

Write what you will and won't do

A restaurant in Lyon keeps a short menu on purpose, so every dish can be done well. Your allowed activities work the same way. Crypto offers spot buying, staking, DeFi, leverage, memecoins and much more, and each brings its own risks and checks. Fewer activities means fewer ways to be caught out.

The "won't" column is often the more useful one: it lists decisions already made, so you need not remake them at 2 a.m.

Allowed activities still need protecting. The next part of the plan is about security.

Add security and custody rules

A family in Toronto keeps passports in one drawer, spare keys with a trusted neighbour, and never tells strangers when they are away. These are routines, not tricks. Your plan should list the security routines from Levels 2 and 3 the same way.

Real support staff never need your seed phrase or private key. If anyone asks for it, the plan's answer is already written: no.

Security keeps your coins safe from others. A research standard keeps them safe from your own rushed decisions.

Research standards and observable rules

Name the asset universe, venue, product, quote currency, dates, timezone and data source. A test on one surviving token differs from a test on all tokens available at the time. A spot series differs from a derivative mark-price series. Choose the environment before selecting favourable results, and preserve the version of the dataset used.

State the question in observable terms. For example, does a specified rule produce a better net result than a defined benchmark over a fixed sample? Avoid words such as strong trend unless you define how they are recognised. Record missing data and product availability. If a pair did not exist or was inaccessible at the supposed time, an assumed trade on it cannot represent feasible practice.

A signal is a defined observation. Entry specifies when and how a hypothetical fill is modelled. Exit specifies the condition and fill assumption. Invalidation states what undermines the idea. Sizing states quantity or the loss-budget procedure. No-action rules cover missing evidence, unavailable liquidity or existing exposure limits.

For a teaching specification, a signal could be a completed daily close above the previous three completed daily highs. Entry could be the next recorded interval's executable ask under a supplied fill rule, not the signal candle's already-past opening price. An exit might occur after a fixed number of intervals or a defined condition. This is a test example, not a claimed profitable strategy. Its purpose is to make timing and feasibility inspectable.

A doctor in Berlin does not prescribe a medicine because a patient saw an advert; evidence comes first. Your plan needs a standard too.

Write which checklists must pass before you buy or use anything: the provider checklist from the provider checking lessons in Level 3, the token research template from the token research lesson in Level 5 and the DeFi checklist from the DeFi dependency lesson in Level 6. Add source rules: primary sources first (project documents, explorers, regulator registers) and no decisions based only on social-media posts. Add a written thesis: why you are buying and what would prove you wrong, the thesis invalidation from the recurring purchase and records lesson in Level 8.

Research rules protect each decision. The next tools protect you when emotions run high.

Cost feasible execution and the right test

Include trading fees, spreads, estimated slippage, network or withdrawal charges where relevant, and funding for derivative examples. State which are embedded in assumed fills and which are added. Size must fit available depth and order increments. A historical high or low inside a candle does not prove your limit order filled, because trade sequence and queue position may be unknown.

Use conservative assumptions when evidence cannot establish execution, and mark the uncertainty. Partial fills require tracking the filled amount and remaining instruction. Stops require a defined trigger reference and adverse-fill treatment. Avoid giving the test perfect buying at lows and selling at highs simply because those prices appear on a chart. That would test hindsight rather than a usable rule.

A trading hypothesis concerns an action sequence and net outcomes under defined execution. An asset thesis concerns rights, use, supply, control and other evidence that may be reviewed as facts change. A recurring purchase policy concerns cash-flow timing, affordability and long-term review conditions. These questions need different benchmarks and evidence.

Do not declare an asset thesis proven merely because a short trading test made money. Nor does a DCA schedule prove that the asset is sound. State which proposition is being examined and what would refute it. A thesis can fail even during a temporary price rise if the promised right does not exist. Research quality requires that the outcome measure fit the actual claim.

Precommitment and a premortem

Lesson C9.1 introduced personal circuit breakers and matched each bias with a countermeasure. Write the ones you chose into your plan, with a trigger and an action for each: "If I feel FOMO, I wait 24 hours." "If I have an unplanned loss, I make no new trades for a week." "If I check prices more than my two daily windows, I delete the app from my home screen for a week."

A pause rule is not a universal safe interval and does not cancel existing positions; it stops new decisions until your written restart conditions are met. A written pause rule makes a checking step part of the process before an urgent decision arrives.

The psychologist Gary Klein suggested a useful planning tool in a 2007 Harvard Business Review article: the premortem. Before a project starts, the team imagines it has already failed, then lists the reasons why. This makes it easier to voice doubts early, while there is still time to fix the plan.

Rules are only half a plan. The other half is a record of what you actually did.

Freeze the rule and keep the decision journal

Write the rule version and date, then test that version. If you change it after seeing results, record the change and treat the prior sample as development information rather than untouched validation. Identify information unavailable at the decision time, including revised supply or entity labels and final candle values not yet formed.

Keep a simple research log of every variation tried. Testing many variations increases the chance of finding an attractive result by coincidence. The next lesson explains why a holdout and forward observation help but do not remove all overfitting. A clear rule specification is the beginning of honest testing, not a certificate that the rule will work.

A pilot's logbook in São Paulo records every flight. Years later, it shows patterns no single flight could reveal.

The transaction records from Lesson C8.4 show what happened; the journal records why, so you can review your thinking, not only your results. Include decisions not to act: a skip, a pause or a missed move is a decision too.

Write feelings in ordinary words; "felt FOMO after group chat" is enough, with no score or diagnosis needed. Add dated notes later rather than rewriting entries.

A journal also lets you test ideas without risking money, as long as you guard against hindsight.

Try the following questions aloud or on paper before opening the worked answers. Compare your reasoning as well as your final answer.

References and further reading