ADVANCED POSITION SIZING & PORTFOLIO HEAT

Portfolio heat is an estimate of planned loss if the stops on open positions are reached under the stated sizing assumptions. It can be calculated by summing those planned stop risks, while also considering correlated outcomes, gaps, slippage and changes in equity. A portfolio cap should fit your account rules and risk capacity; there is no universal 6% maximum. EXAMPLE: Position 1 has a planned stop risk of 2% of account equity. Position 2 has a planned stop risk of 1.5%. Position 3 has a planned stop risk of 2%. The sum of these planned stop risks is 5.5% before execution differences, gaps or other exposure. This is an illustrative calculation, not a recommended portfolio limit. Before adding another position, reassess the total planned stop risk and plausible joint-loss scenarios. Summing stop risk is useful, but it does not perfectly predict a portfolio loss when markets move together or stops execute away from their levels.

Correlated positions can create concentrated exposure because losses may occur together, but correlation is not fixed and does not translate directly into a single portfolio-risk multiplier. Relationships vary by measurement period, timeframe and market regime. When holding positions such as EURUSD and GBPUSD, account for each position's stop risk separately and then consider the possibility of joint losses, changes in correlation, gaps and slippage. Do not assume an 80% or 90% correlation means a fixed reduction or increase in risk. Portfolio rules may limit the number or size of correlated positions, but they should be calibrated to the instruments, account constraints and tested strategy. SMT or relative-strength observations can be a hypothesis to test; they do not guarantee that one market will lead or that the other will reverse.

Scaling into or out of positions changes exposure and the distribution of realized wins and losses. Adding to a winning position does not automatically protect capital, and adding to losing positions is not universally wrong or right outside a defined model; either choice must be specified and tested. If your strategy permits scaling in, define the trigger for each add, the stop and exit for each position, and the maximum combined planned loss before placing orders. Recalculate total exposure after every add and account for correlated positions, gaps and execution costs. A stop moved to breakeven does not guarantee a risk-free result. Partial closes are also a strategy choice. A 50% exit at 1:2 does not guarantee profit on the full trade or account; results depend on the entry, realized fill, cost, remaining-position stop and subsequent exit. Compare partial-exit schedules with a full-position baseline using the same signals and costs.

The Kelly Criterion is one model for estimating a growth-maximizing fraction when the return distribution and edge are known accurately. Its output is extremely sensitive to estimation error and may be inappropriate when outcomes are nonstationary, correlated or poorly measured. FORMULA FOR A SIMPLIFIED TWO-OUTCOME MODEL: f = (p × b − q) / b, where p is the win probability, q = 1 − p, and b is the average net win measured in units of the average loss. HYPOTHETICAL EXAMPLE: With p = 0.40, q = 0.60, average win = 3R and average loss = 1R, f = (0.40 × 3 − 0.60) / 3 = (1.20 − 0.60) / 3 = 0.20 or 20% full Kelly under these assumptions. This mathematical output is not a recommendation to risk 20% on a trade. Fractional Kelly simply scales the estimate down, but quarter, half or other fractions are not automatically safe or suitable. Validate the inputs, account for costs and correlation, and compare with conservative fixed-risk limits; if the edge is uncertain, do not rely on Kelly output for live sizing.

PRE-TRADE CHECKLIST (ADAPT TO YOUR STRATEGY): - Confirm the scenario and required entry conditions are defined and observable. - Mark stop, target and invalidation before entry. - Calculate size from planned risk, stop distance, instrument value and execution costs. - Check aggregate exposure, including positions that may be correlated. - Confirm the order type and slippage assumptions fit the model. - Record the setup and the reason for any no-trade decision. POST-TRADE REVIEW: - Record entry, exit, result and execution costs. - Note whether rules were followed and classify any errors. - Review wins and losses using the same criteria. - Update performance statistics when the sample is sufficient and uncertainty is reported. A checklist can reduce omissions, but no fixed risk percentage, payoff ratio, portfolio-heat threshold or checklist duration guarantees safety. Tailor it to your tested strategy and account constraints.