Home / Risk Management / Portfolio Heat
PortfolioIntermediate

Portfolio Heat

Portfolio heat is the sum of the risk currently at stake across all open positions, expressing how much of the account is exposed to loss at any moment.

Quick answer: Portfolio heat is the sum of the risk currently at stake across all open positions, expressing how much of the account is exposed to loss at any moment.

In simple words

Portfolio heat measures how much of your account is on the line right now across every open trade combined. A single trade may risk only one percent, but ten open trades risking one percent each put ten percent at stake, and if they are correlated it is effectively one large bet. Heat is the running total that tells you whether the whole book, not just each trade, is within safe limits.

Purpose

It exists because per-trade limits alone do not control aggregate exposure; heat is the portfolio-level ceiling that stops many individually reasonable trades from combining into one dangerous one.

Professional explanation

Summing open risk

The simplest heat measure adds up the current stop-based risk of every open position: for each trade, the rupees it would lose from its current stop to entry (or to the current price if the stop has moved to lock in profit), summed across the book. Expressed as a percentage of equity, this is the fraction of the account that could be lost if every open stop were hit at once. It is a real-time number that changes as positions open, close, and as stops trail, and it is the natural quantity for a risk engine to cap.

Why correlation clusters heat

Raw summation treats positions as independent, but correlated positions do not lose independently. Ten long positions across Nifty constituents are largely the same bet on the index, so their true combined risk is close to the arithmetic sum rather than the diversified, smaller figure independence would imply. Heat management therefore groups positions into correlation clusters (by sector, by factor, by underlying) and caps risk per cluster, not just per position, so that a book cannot accumulate ten one-percent trades that are really one ten-percent trade on the same driver.

Heat caps and their hierarchy

A typical risk engine enforces a hierarchy of caps: a per-trade cap (say 1 percent), a per-cluster or per-underlying cap (say 3 to 5 percent), and a total portfolio heat cap (say 6 to 10 percent). When a new signal would breach any cap, the engine either reduces its size to fit or rejects it. These numbers are rules of thumb that depend on strategy and risk tolerance, not universal constants, but the structure of nested caps is standard because it controls concentration at every level simultaneously.

Heat and available capital for new trades

Heat directly constrains how many new positions the system can take. As heat approaches its cap, new signals are throttled or skipped, which is a feature, not a bug: it prevents piling into a market that is already heavily represented in the book. This creates an implicit prioritisation problem, since not every signal can be taken, and systems resolve it by ranking signals or by first-come allocation, always subject to the heat ceiling. It also means backtests must model heat-based rejection or they will overstate how many trades were actually takeable.

Dynamic heat and trailing stops

Heat is not static: as trailing stops move to breakeven or into profit, the risk of those positions falls, freeing heat for new trades. A well-built engine recomputes heat continuously, so a position whose stop has trailed above entry contributes zero or negative risk and no longer consumes the budget. This dynamic view rewards letting winners reduce their own risk and is a meaningful difference from a naive count of open positions, which would wrongly treat a risk-free runner the same as a fresh full-risk entry.

Heat under gaps and stress

The heat figure assumes stops fill near their levels, so in a gap or a fast market the realised loss can exceed the computed heat, sometimes badly for a book concentrated in one correlated cluster. Prudent heat limits therefore leave headroom below the theoretical maximum and are complemented by worst-case gap scenarios and by circuit breakers that halt trading when losses accelerate. Heat is a control for normal conditions; it must be backed by harder stops for abnormal ones.

Formula

Heat = Σ (per-position stop risk) ÷ Equity

Sum each open position's rupee risk (distance from current price or trailing stop to the stop, × size) and divide by equity. Positions whose stops have trailed into profit contribute zero or negative risk. Cap heat per cluster and for the whole book.

Practical example

Illustrative example (Indian market)

On a ₹5,00,000 account you hold five open positions, each originally risking 1 percent (₹5,000), so naive heat is 5 percent (₹25,000). But three of them are long positions in Nifty-heavy stocks that move together, so you treat them as one correlated cluster and cap that cluster at 3 percent; their combined ₹15,000 exactly meets the cluster cap, blocking a sixth correlated long. Meanwhile one of the five has trailed its stop to breakeven, so it now contributes zero heat, freeing ₹5,000 of budget for an uncorrelated new trade. Your live heat is therefore 4 percent, within a 6 percent total cap, with one cluster maxed out.

A common NSE trap is holding several F&O positions that all depend on Bank Nifty direction (a long future, a bull call spread, sold puts); individually each looks small, but their heat clusters into one large directional bet, so a single gap down on expiry can hit all of them at once far beyond the per-trade budget.

Advantages

  • Controls aggregate exposure that per-trade limits alone miss
  • Correlation clustering prevents many small trades becoming one big bet
  • Nested caps limit concentration at trade, cluster and book level simultaneously
  • Dynamic heat rewards trailing stops by freeing budget as winners de-risk

Limitations

  • Accurate clustering needs a correlation view, which is noisy and unstable in crises
  • The sum assumes stops execute near their levels, which gaps violate
  • Heat can rise faster than expected when correlations converge under stress
  • Setting cap levels is judgement, not science, and wrong caps either strangle or over-expose the book
  • Backtests that ignore heat-based rejection overstate the number of takeable trades

Why it matters in practice

  • Heat is the portfolio-level guardrail that stops correlated accumulation
  • It is the difference between a diversified book and a disguised single bet

Common mistakes

  • Summing per-trade risk while ignoring correlation, so clustered bets look diversified
  • Treating open-position count as heat, ignoring that trailed stops carry little risk
  • Setting no cluster cap, allowing ten same-direction trades to aggregate
  • Assuming heat is the worst case, when gaps can exceed it
  • Not recomputing heat in real time as stops trail and positions close
  • Building a backtest that lets the system take every signal, ignoring the heat ceiling that would have blocked some live

Professional usage

Institutional risk engines track heat continuously and enforce nested limits by position, by cluster or underlying, by sector and by risk factor, rejecting or shrinking orders that would breach any of them. They define correlation clusters explicitly and stress them under scenarios where correlations spike to one, and they pair heat limits with hard circuit breakers for abnormal conditions. Heat is a pre-trade check the strategy cannot override, ensuring aggregate exposure stays governed no matter how many independent signals fire.

Key takeaways

  • Portfolio heat is the summed open risk across all positions as a fraction of equity
  • Correlated positions cluster into one bet, so cap risk per cluster, not just per trade
  • Recompute heat live; trailed stops free budget, fresh entries consume it
  • Heat controls normal conditions; back it with circuit breakers for gaps and stress

Frequently asked questions

What is portfolio heat?
Portfolio heat is the total risk currently at stake across all open positions, usually the sum of each position's stop-based risk expressed as a percentage of equity. It measures how much of the account is exposed to loss right now.
How is portfolio heat calculated?
Add up the rupee risk of every open position (distance from current price or trailing stop to the stop, times size) and divide by account equity. Positions whose stops have trailed into profit contribute little or no heat.
Why is correlation important for heat?
Because correlated positions do not lose independently. Ten long positions in index-heavy stocks are effectively one bet on the index, so their real combined risk is close to the sum, which is why heat is capped per correlation cluster.
What is a good portfolio heat limit?
There is no universal number; common rules of thumb use around 6 to 10 percent total heat with tighter per-cluster caps, but the right level depends on strategy, win rate and correlation. Treat published figures as illustrative, not advice.
How is heat different from risk per trade?
Risk per trade caps a single position; heat caps the sum of all open positions. A book can respect every per-trade limit yet still be over-exposed in aggregate, which is exactly what heat controls.
Does a trailing stop reduce portfolio heat?
Yes. As a stop trails to breakeven or into profit, that position's remaining risk falls to zero or below, so it stops consuming the heat budget and frees room for new trades.
What is a correlation cluster in heat management?
It is a group of positions that move together (same sector, factor or underlying) and are therefore treated as one risk unit with its own cap, preventing the book from stacking many trades on the same driver.
Can portfolio heat exceed its computed value?
Yes, in a gap or fast market where stops fill far from their levels, especially in a concentrated cluster. That is why heat limits leave headroom and are backed by circuit breakers.
How does heat limit new trades?
When heat nears its cap, the engine throttles or rejects new signals, or shrinks their size to fit. This prevents piling into an already heavily represented market at the cost of skipping some signals.
Why must backtests account for heat?
Because a heat cap would have blocked some trades live. A backtest that takes every signal overstates the number of takeable trades and the achievable exposure, flattering the results.
Is portfolio heat the same as margin used?
No. Margin is what the broker requires to hold the positions; heat is how much you would lose if stops were hit. Two positions can use similar margin but carry very different heat depending on stop distance.
How does heat relate to portfolio diversification?
Diversification aims to make open positions less correlated so that their combined heat is genuinely smaller than the sum; heat is the measurement that reveals whether the diversification is real or illusory.

Voice search & related questions

Natural-language questions people ask about Portfolio Heat.

What is portfolio heat?
It is the total risk across all your open trades at once. Even if each trade risks one percent, ten of them put ten percent on the line.
Why do all my positions lose together?
Because they are correlated, really the same bet. Group trades that move together and cap the whole group, not just each trade.
How much total risk should I have open at once?
As a rough guide, keep total open risk in single digits, maybe six to ten percent, with tighter caps on any group of similar trades.
Does moving my stop to breakeven free up risk?
Yes. Once a stop is at breakeven that trade risks nothing, so it stops using your risk budget and you can take another trade.
Why did small trades add up to a big loss?
They were all the same direction on the same driver, so a single gap hit all of them together, far beyond what one trade should have cost.
Is heat the same as the margin I am using?
No. Margin is what the broker holds; heat is what you would lose if your stops hit. They can be very different.

Sources & references

    Last reviewed 11 July 2026. Educational content only — not investment advice. Markets and rules change; verify current conventions with SEBI, NSE/BSE and your broker.

    ← Previous
    Maximum Drawdown
    Educational content only — not investment advice. Examples use illustrative numbers and simplified models. Algorithmic trading and derivatives involve substantial risk. See our Risk Disclosure and SEBI Disclaimer.