ADVANCED RISK MANAGEMENT & POSITION SIZING

The Kelly formula estimates a growth-maximizing fraction only when its inputs accurately describe a repeatable process. For a simplified binary-outcome model, f = (b × p − q) / b, where b is the average net win per unit lost, p is win probability and q = 1 − p. Real trading involves changing odds, costs, slippage and non-identical outcomes, so the estimate can be very wrong. ILLUSTRATIVE CALCULATION: With p = 0.45, q = 0.55 and average win b = 3R versus a 1R loss, f = (3 × 0.45 − 0.55) / 3 = 0.80 / 3 ≈ 26.7%. This is a theoretical full-Kelly estimate under those assumptions, not a recommendation to risk 26.7% of an account. Fractional Kelly scales the estimate down (for example, one-quarter or one-sixteenth), but a fraction is not automatically safe. Estimation error, small or unrepresentative samples, correlation and drawdowns can make Kelly sizing unsuitable. Do not choose a risk percentage merely because a formula returns it; compare conservative fixed-risk approaches and account restrictions, and validate assumptions before any live use.

Correlation between instruments changes with timeframe, market regime and measurement period. Instruments that have often moved together can diverge, and seemingly unrelated markets can become correlated during stress. Do not convert a correlation coefficient directly into a portfolio-risk multiplier. A 1% planned loss on each of two positions remains two separately defined stop risks, while the likelihood and timing of those losses may be related. Assess aggregate exposure with volatility, stop distances, gap risk, account currency and plausible joint-loss scenarios instead of assuming a fixed correlation or that both positions will stop out together. Portfolio rules should fit the strategy and account constraints. Review rolling relationships and stress scenarios relevant to the instruments and period you trade; no universal limit on the number of correlated positions or standard reduced risk percentage is safe for every account.

Drawdown is the decline from an account's prior peak to a later trough. Prepare a response plan before losses occur, using thresholds appropriate to your strategy, account constraints and personal risk capacity. EXAMPLE OF A REVIEW PROCESS (NOT UNIVERSAL THRESHOLDS): • At a predefined warning level, check whether losses fit the tested distribution and whether you followed the rules. • At a larger predefined drawdown, consider reducing risk, pausing new entries or switching to simulation while reviewing the data. • Before resuming, check for execution mistakes, changed market conditions, data-quality issues and evidence that the strategy's edge may have weakened. THE RECOVERY FORMULA: A 20% drawdown requires a 25% gain to return to the starting balance. A 50% drawdown requires a 100% gain to return to the starting balance. These figures illustrate the arithmetic, not a forecast or a recommended drawdown allowance. Define stop and review rules in advance rather than increasing risk to recover losses.

Account size alone does not determine the right percentage risk, order size, trade frequency or return target. Position size should follow the stop distance, instrument specifications, planned account risk and aggregate exposure. Smaller accounts may face practical minimum-size constraints; larger accounts may face liquidity, slippage or execution constraints. HYPOTHETICAL COMPOUNDING MATH: If $10,000 grew by a constant 3% every month, before costs, taxes and withdrawals, the arithmetic would be approximately: Year 1: $14,258 Year 2: $20,328 Year 3: $29,004 Year 5: $59,092 Year 10: $348,912 This is an illustration, not an achievable or expected return target. Trading returns vary and include losing periods; a constant positive monthly return is not guaranteed. Evaluate performance using a representative record that includes costs, drawdowns and out-of-sample results.

Risk management combines sizing, execution, strategy design and behavior; a fixed 80/20 split is not a universal measurement. THE FEAR OF LOSS: People may react asymmetrically to losses and gains, but the size of the effect varies by study and context; avoid treating a single multiplier as universal. This causes: - Moving stop losses to avoid realizing a loss - Closing winning trades too early to secure profits - Revenge trading after losses - Avoiding valid setups due to recent losses THE GREED TRAP: After a winning streak, traders often: - Increase position size (I am on fire!) - Take lower-quality setups - Remove stop losses EXAMPLE: A PRE-TRADE CHECKLIST: Before entering, ask whether the setup meets your written criteria, whether the stop and target are defined, whether expected payoff is supported by testing, and whether the planned risk fits your limits. Thresholds such as a 1:3 risk-to-reward ratio or 1% risk may be used in a specific personal model, but they are not universal requirements; define subjective grades such as A+ in observable terms before testing. If a required rule is not met, follow the plan you set in advance.