A friend sent me a screenshot of his open positions a while back and asked if his sizing looked right. Seven trades, each risking 1 percent of his account to its stop. Individually, every one of them was textbook. Then I asked what happens if all seven stop out in the same week, and he went quiet, because the answer was 7 percent of his account and he had never thought of it as one number.
That one number has a name. Portfolio heat is the total amount of equity you lose if every open position hits its stop at once. It is the simplest portfolio-level risk measure there is, it takes about two minutes to compute by hand, and almost nobody tracks it. Per-trade sizing rules are everywhere, the habit of adding them up is rare, and the gap between those two things is where a lot of accounts quietly die.
Computing heat, and why it moves
For each open position, take the distance between the current price and the current stop, multiply by position size, and you get that trade's open risk in currency terms. Sum it across every position and divide by account equity. That is your heat. If you are long 200 units of something trading at 50 with a stop at 46, the position carries 800 in open risk, which on a 40,000 account is 2 percent of heat all by itself.
Two details matter. First, use the current stop and the current price rather than the original entry levels. A position whose stop has trailed up to breakeven contributes zero heat, because from here it cannot cost you anything beyond slippage. This is why heat is a living number rather than a diary entry. It drops as stops trail, rises when you add, and tells you at any given moment how much of your account is actually on the table.
Second, heat measures the worst case for today rather than an expected loss. Most weeks you will not get stopped out of everything at once. But the weeks where you do are exactly the weeks that decide whether you are still trading next year, so the pessimistic number is the one worth capping.
On the cap itself, somewhere around 5 to 8 percent of equity is the range I keep coming back to. The logic is unromantic. A full portfolio stop-out at 6 percent hurts and is recoverable. Two or three of those back to back, which happens in genuinely bad regimes, puts you down roughly 15 to 20 percent, and that is near the edge of what most traders can absorb without starting to trade differently, which is its own compounding damage. Where exactly you set the cap matters less than the cap being hard, because a soft cap is a suggestion, and suggestions lose to good-looking charts every time.
Correlated positions are one trade
Here is where the arithmetic lies to you. Say you are long five different altcoins, each sized to risk 1 percent. On the spreadsheet that reads as five independent small bets. In practice all five ride the same tide, and when bitcoin drops hard, altcoin correlations converge toward one and the five stops trigger within hours of each other. Five separate 1 percent risks collapse into a single 5 percent bet on crypto beta, and the diversification you thought you had turns out to be mostly cosmetic.
Crypto is the clearest case but hardly the only one. Five semiconductor longs behave the same way in a sector selloff. Long gold miners plus short the dollar is closer to one position than two. My rule of thumb is crude but it has served me well: if a single headline would stop all of them out, they are one trade, and their combined risk counts as one line against the heat cap.
You do not need a correlation matrix for this. Crude buckets catch most of the danger, and unlike the matrix you will actually maintain them. One bucket for crypto as a whole (I used to split majors and alts, and in a real drawdown that split turned out to be flattering fiction), one per equity sector, one for anything that is fundamentally a dollar bet. Then cap each bucket at roughly half your total heat cap, so with a 6 percent total no single theme carries more than 3. Worth remembering that correlations measured in calm markets understate stress correlations, always in the direction that hurts. What you actually care about is how these assets move during a liquidation, and that is nearly always tighter than whatever trailing window you measured suggests.
The pre-entry gate
None of this is worth much as an after-the-fact report. The point is to run the check before the order goes in, every single time, as a gate the trade has to pass. Mine looks like this:
- Compute current total heat: the sum of price minus stop, times size, across everything open, divided by equity.
- Assign the candidate trade to a bucket and note that bucket's current heat.
- Compute the candidate's risk at the intended size, using the stop you would actually place rather than a placeholder.
- Add it all up. If total heat after entry would exceed the cap, or the bucket would exceed its cap, the trade does not go on.
When the gate blocks you, there are three honest options. Skip the trade. Size it down until it fits, though below a certain size a trade stops being worth its attention cost. Or free up heat by closing or trimming something already open. The third option is the interesting one, because it forces a comparison you would otherwise never make: is this new idea better than the weakest thing I am currently holding? If the answer is no, you also have your answer about the new trade. One caution here, tightening an existing stop to manufacture room is technically legal under these rules and usually a mistake, since you are degrading a trade you presumably still believe in to make space for one you are merely excited about.
The gate earns its keep through what it prevents at the worst moments. Maxed-out heat tends to coincide with the stretch where you feel most confident, because you have been adding as things worked. That is exactly when the next idea looks irresistible, and exactly when adding it turns the eventual reversal into a portfolio-wide event instead of a bad trade or two.
If you want a starting point, just measure. Take whatever is open right now, bucket it crudely, and compute the number. Everyone I have walked through this with found their real heat higher than their guess, usually because of correlation rather than sizing. Then pick a cap, write it somewhere you will see it before entries, and let it block at least one trade before you judge whether it is too strict. In my experience the cap that feels slightly too tight is usually the one that is right.