Daily limit and drawdown

This lesson answers three questions: how many transactions fit the day’s budget, where to get the limit from, and what you do once it is used up.

Thesis of this lesson: a limit is a decision made in advance, not in the middle of emotion

On a bad day the brain does not want to finish — it wants to recover. The daily limit and maximum drawdown are rules that make the decision for you before emotion takes over. There is no maybe in a limit: there is a number, a moment of interruption, and a return procedure.

Convert the limit into R — otherwise you do not know how many transactions you have

A daily limit in per cent is useless until you know how many transactions fit inside it:

transactions_to_limit = daily_limit_% / risk_per_transaction_%

This table changes how you think about the limit: at 1% risk and a 2% limit, the third transaction of the day is already outside the rules — even if it is the best setup of the week. The limit stops being an abstract number and becomes a counter.

A limit in R gives a fixed number of transactions and a variable amount; a limit in per cent works the other way round.

Full calculation: $20,000, 1% risk, 3R limit

Capital $20,000, risk 1% per transaction, daily limit 3R:

risk per transaction = 1.0% x $20,000 = $200 = 1R
daily limit = 3R x $200 = $600
daily limit in % = 600 / 20,000 = 3.0% of capital
losses to the limit = 3R / 1R = 3 transactions

The third loss closes the budget, so a fourth entry appears only if an earlier transaction closed better than a full -1R. Count the day’s result after costs, because breaking even also costs money.

Variant with 2% risk

risk per transaction = 2.0% x $20,000 = $400 = 1R
daily limit = 3R x $400 = $1,200
daily limit in % = 1,200 / 20,000 = 6.0% of capital
losses to the limit = 3R / 1R = 3 transactions

Three losses in a row now cost $1,200, that is 6% of capital, because R scales with size.

Result check

equity path   20,000 - 600 = $19,400   19,400 / 20,000 = 0.97 → -3.0%
              20,000 - 1,200 = $18,800 18,800 / 20,000 = 0.94 → -6.0%
cost of recovery   0.03 / 0.97 = 3.1%     0.06 / 0.94 = 6.4%

Both routes agree with the limit in per cent. The percentage notation, however, changes the number of transactions along with risk: at a 3% capital limit and 2% risk, the second R takes the day outside the limit.

How to set a daily limit

Starting point: 2-3% of the account per day (more in the 1-2% rule). Concretely:

The key point: you enforce the limit with tools, not with a resolution. Phone out of reach, platform closed, notification set. An internal I will hold back this time loses under stress.

A limit that is used up once in a hundred sessions is insurance, while one that breaks every third session points to a problem in your setups.

A limit from data, not from impression

You take the limit from the distribution of daily results. Three numbers from the journal: the daily average, the daily standard deviation, and the longest losing streak.

daily_limit ≈ daily_average + 2 x daily_std_deviation

The normal model is a starting point, because daily results are skewed; you round it off on your own data.

A distribution from ten sessions says nothing. A decisive measurement begins with a sample of about 62 transactions, derived in the lesson expectancy and drawdown.

Three levels of limits and their order

The order is hierarchical. Drawdown from peak is superior, because it describes the state of capital and applies even when a new day provides a fresh budget in R. Then the daily limit operates, and finally risk per transaction.

peak $20,000, today $18,600 → -7.0% from peak → half size
1R at half size = $100; 3R = $300 = 1.6% of capital (not 3%)
the day is closed by the daily limit, and the next day still at half size
return to full size: leaving the range below -5% AND 10 plan-compliant transactions

When two levels act together, the more conservative rule wins: the limit closes the day, but the drawdown threshold still keeps size at half. Resetting the daily limit does not cancel the drawdown threshold, and that is where people most often break their own procedure.

Drawdown measured from peak

Drawdown is the fall from the account peak, not from the start of the day. The maximum drawdown, which switches you off (for example 10%), protects against a spiral: it turns a losing streak into a pause, not into recovering. After the pause you return to smaller size or to the journal.

Drawdown from peak is a state and does not reset with a new day, while a day’s loss is an event. Confusing these notions ends with a new start from Monday.

Why drawdown is more expensive than it looks

A loss and its recovery are not symmetrical. The formula for the return needed to get back to the peak:

required_return = drawdown / (1 - drawdown)

That is the whole arithmetic that justifies a limit: at -50% you must double what is left to get back to zero on the curve. The market does not know you are recovering — it will not increase your rate of return. That is why maximum drawdown is not a setting, but a lower bound below which the game is over.

The practical consequence: three days of -2% are better than one day of -6%, because the distance to the pause threshold is different.

Pause thresholds

The numbers are illustrative — what matters is the structure: the larger the drawdown, the longer the pause and the tougher, countable return criterion. The criterion until I feel confident is not a criterion, because it cannot be verified.

Pause thresholds and the daily limit operate on two time scales. A day with an exhausted limit less often than once in twenty sessions is a normal cost of trading, while every other session is a problem in the entry system.

What to do after hitting the limit

A number alone will change nothing without a procedure for the moment after it is exceeded. You make four decisions before the session, not after the limit.

Moving the limit during the day invalidates the mechanism regardless of the result. You negotiate it at the only moment when you are inclined to, that is, after a loss; an extension that ends in profit teaches the habit more strongly than a loss, because it worked.

A limit exhausted by breaking the procedure means a break until the cause is named in writing, because a repeatable execution error costs more than a bad market day.

What to do after a losing streak (without revenge trading)

1. Stop — a reached limit means the session is over. There is no last trade to fix it.
2. Reduce size — after a drawdown of 5% or more you trade at half of normal size, until the journal shows a stable process.
3. Return to the journal — find what the losses had in common: too large a size? entering against the trend? changing rules mid-way?
4. Restore full size only on the basis of data — for example after 10 plan-compliant transactions, not after feeling good.

Revenge trading is not a weakness of character — it is a missing procedure. The psychology lesson explains the mechanism; this lesson provides the execution rule.

Decision table

Prose leaves room for interpretation, and in a session interpretation costs money. In no row does the correct reaction increase activity.

When the limit says do not trade

1. The day’s transaction budget is exhausted. The counter is absolute — not almost absolute. Three losses at a 3R limit end the session even half an hour before the close.
2. You are halfway to the limit and you must make it back. That is the state in which the limit exists precisely to stop you.
3. Drawdown exceeded the pause threshold. Trading at full size during a pause is a breach of procedure. A fresh daily budget does not cancel the drawdown threshold, because the two thresholds measure different things.
4. You change the limit during the session. Every today is an exception is proof that the limit was never a rule.
5. The only reason for entry is that the day’s result is negative. That is an attempt to recover a loss, not a setup from the plan; the same transaction assessed in the morning would usually not make it into the sheet.
6. A drawdown pause is active. During a pause you do not trade even minimal size, because the procedure covers all positions.

The same limit in a trend and in consolidation

The limit concerns the tail of the distribution, not the average, so the same number 3R behaves differently in each environment.

In consolidation, four attempts at 1R each give -4R, even though only one thesis was wrong, and the system’s expectancy remains unchanged.

You check this on data: a structure label for each session, the share of sessions with an exhausted limit in both environments, and the standard deviation of daily results. If the limit breaks three times more often in consolidation, you decide in advance: skipping those sessions, lower risk in that environment, or accepting the interruptions.

Typical mistakes

• A limit set in your head — without a record and an alarm it does not exist under stress.
• A limit in per cent without converting it to R — you do not know how many transactions really fit the day’s budget.
• A limit pulled from thin air — a number chosen because it sounds reasonable, without checking the journal.
• Raising the limit because it was a bad market day — that is changing the rules mid-game.
• Drawdown measured from yesterday — you measure drawdown from the peak, otherwise recovering masks a spiral.
• The limit calculated from full size during a drawdown — it then corresponds to twice the risk you actually bear.
• Returning to full size by feel — a return based on data, not on how you feel.
• A record without a cause category — the entry limit exceeded does not say whether the setup, execution, or psychology was at fault.

Measurable exercise

Write in a sheet: the daily limit in % and in USD, risk per transaction, the number of transactions that fit the limit (from the formula), maximum drawdown in %, and the pause thresholds with return criteria expressed as a number of transactions. Add a notification at the limit and calculate the required return with the formula DD / (1 - DD). Over the next 10 sessions, tick off limit not exceeded / exceeded and procedure executed.

Add a second sheet for the data: daily average, standard deviation, and the longest losing streak. Compare the limit from the plan with the threshold average + 2 x standard deviation.

Common questions

What daily limit makes sense?

A starting point is 2-3% of the account, but first convert it into a number of transactions: the limit divided by risk per transaction. At 1% risk and a 2% limit you have two transactions — not up to two percent. More important than the size is that the limit exists and is enforced.

Why does a 50% drawdown require a 100% return?

Because percentages are calculated from an ever smaller base. Out of $100, $50 is left, and getting back to $100 requires doubling what remains. The formula: required return = DD / (1 - DD). That is why a maximum drawdown protects the account better than any single decision.

Can I break the limit once because the setup was certain?

That is exactly the moment the limit stops existing. A better certain setup within a limited session than a certain setup outside the rules — and if breaking the limit works out, it rewards a habit that will take more the next time.

How does drawdown differ from a day's loss?

Drawdown is calculated from the account peak and shows the entire fall, not just the day. It is what sets the pause and the size reduction. A day's loss is a local event; drawdown is the state you are in.

Is the daily limit counted per instrument or for the whole account?

For the whole account. A 3R limit describes the day, not a single position, so two open positions with 1R of risk each occupy 2R of the same budget. With correlated instruments risk adds up faster than it looks on separate charts — the method for counting such cases is in the lesson on portfolio risk.

Do commissions and swap enter the daily limit?

Yes, because the limit concerns the account result, not the clean result from the platform. Keep two columns, however: the result before costs and after costs. If the limit is exhausted mainly after costs, the problem is not the market but too many transactions relative to the edge — that is what the round-trip cost lesson describes.