Lesson 01: Portfolio-Level Risk

Track 5 sized one trade correctly. This track starts where that stops being sufficient: several positions open at once, each sized correctly, adding up to something nobody decided. The shift is from asking "is this trade the right size?" to "what is the whole book worth if today goes badly?"

A budget, then allocations

Work top-down. Set the total first, then divide it, rather than adding positions until it feels like enough.

Level	Question	A workable answer
Total open risk	What do I lose if everything stops out today?	4% to 6% of the balance
Per driver	How much of that on one theme?	3%
Per currency	How much net on one currency?	2%
Per trade	How much on one idea?	1%

The numbers cascade: a 6% total with a 1% per-trade rule permits at most six positions, and the driver and currency caps will usually bind before that. The per-trade rule is the least restrictive of the four, which is precisely why sizing one trade at a time is not portfolio risk management.

Choosing the total

Start from the daily limit. If a 3% day ends your session, a 6% total open risk means half the book stopping out ends it.
Check it against the drawdown you will accept. Track 5's drawdown arithmetic applies: two consecutive full-budget days is about 12%, and the recovery from that is about 13%.
Reduce it for correlation. A 6% budget across six genuinely independent ideas is different from 6% across six versions of the same bet — and the second is the common case.

Try it now

The drawdown calculator turns a sequence of full-budget days into the recovery each one demands, which is the honest way to choose a total rather than picking a number that sounds prudent.

Drawdown Calculator

Exposure by driver

Track 5 aggregated by currency. At portfolio level that is necessary and not sufficient, because positions with no currency in common can still be one bet.

Group by driver — the thing that would have to happen for the trade to work.

Driver	Positions that share it
The dollar weakens	Long EUR/USD, long AUD/USD, short USD/CAD, long gold
Risk appetite improves	Long AUD/JPY, long equities, short gold, short CHF
The Fed turns dovish	Long EUR/USD, long gold, short USD/JPY
Oil rises	Short USD/CAD, long NOK, long CAD/JPY

The first row has four tickets, four different instruments, and one question behind all of them. A currency-level check catches three of the four and misses gold entirely, which is the gap that driver-level grouping exists to close.

Doing it in practice

For each open position, write the one sentence that has to be true for it to work.
Group the sentences that are the same sentence.
Total the risk in each group.
Compare against the per-driver cap.

Note

This takes two minutes and is worth doing before each new position rather than at the end of the day. The question it answers — does this add a new bet, or a fourth copy of one I already hold? — is the one that stops a book quietly becoming a single trade.

Breach rules

Budgets get breached, usually without a rule being broken: correlations rise, a position is added to, a stop is moved to break-even and frees up budget that then gets used. Decide the response in advance.

Breach	Response
Total open risk above budget	No new positions until it is back inside
One driver over its cap	Close or reduce the weakest position in that group
Correlations converged, pushing effective risk up	Reduce across the group, not just one
Daily limit hit with positions open	No new entries; manage the open ones by their own rules

What to close first

The position with the weakest original reason. You wrote it down; use it.
The one that duplicates another. Closing a fourth copy of a bet costs the least diversification, because it was adding none.
The one furthest from its target, which has the most left to be wrong about.
Not the one that is losing. That is the loss-aversion reflex from Track 7, and the losing position may be the one with the most intact thesis.

Caution

Point 4 matters because the instinct runs the other way. Closing winners to free budget while keeping losers is exactly the behaviour that turns a positive expectancy negative, and it is easiest to rationalise when it is dressed up as risk management.

Rebalancing a book of trades

A book drifts. Stops move to break-even, positions are partly closed, price moves change what each position is actually risking. Rebalancing is the periodic act of recomputing what you hold rather than what you opened.

Risk is not what you opened with. A position whose stop is at break-even risks nothing and is using none of the budget — it has freed room, which is a genuine thing to notice.
A position in profit with a trailing stop has negative risk. Include it as zero rather than as its original 1%; counting it at 1% leaves budget unused.
Correlations change. The book you assessed on Monday may be more concentrated on Thursday with no new positions added.

A weekly rebalance

Recompute the risk on each open position from the current stop, not the original one.
Regroup by driver, since the reasons may have changed with the market.
Compare totals against the caps.
Act only if something is outside. A book inside all four limits needs nothing, and rebalancing for its own sake pays spread to achieve nothing.

Step 4 is the discipline. Portfolio management at retail scale is mostly a monitoring activity with occasional action, and the trader who rebalances every week regardless has added a cost and a decision to a process that was working.

Key takeaways

Set the total open risk budget first, then divide it by driver, currency and trade — the per-trade rule is the loosest of the four.
Group positions by the sentence that has to be true for them to work; a currency check misses cross-asset copies of the same bet.
Write breach rules in advance, and close the weakest reason or the duplicate — never default to closing the winner.
Recompute risk from current stops weekly, and act only when something is actually outside a cap.

By the end of this lesson