Why Risk Management Beats Strategy Selection

Most traders obsess over entries: which indicator to use, where to buy, when the signal fires. Professional algorithmic traders know a different truth: how you manage risk determines whether you survive long enough to profit.

Consider two algorithms. Algorithm A has a 65% win rate but risks 10% of the account per trade. After 10 consecutive losses (which happen with any strategy), the account is down 65%. Recovery requires a 186% gain. Algorithm B has only a 55% win rate but risks 1% per trade. After 10 consecutive losses, the account is down 9.6%. Recovery requires a 10.6% gain.

The superior algorithm is almost certainly B — not because it has better entry logic, but because it sizes positions appropriately. This is the core insight of professional risk management: position sizing and drawdown control are more important than the entry signal itself.

Position Sizing: The Fixed Fractional Method

The standard approach in algorithmic trading is fixed fractional position sizing. You risk a fixed percentage of your current account balance on every trade — typically 1-2%.

The formula: Position Size = (Account Balance × Risk Per Trade %) / (Entry Price - Stop Loss Price)

Example: Account = $10,000, risk 1% = $100. Entry = $100 (per unit), Stop Loss = $95 (5% below entry). Position Size = $100 / ($100 - $95) = $100 / $5 = 20 units.

If the trade hits the stop loss, you lose exactly $100 (1% of account). If it hits the take-profit at $110, you gain $200 (2% of account) — a 2:1 reward-to-risk ratio.

The key property of fixed fractional sizing is that position sizes scale automatically with your account. As your account grows from wins, you trade larger (compounding gains). As it shrinks from losses, you trade smaller (protecting against ruin). This prevents the common mistake of trading the same fixed size whether your account is up or down.

FerroQuant's execution engine calculates position sizes dynamically using this formula for every trade, with per-market leverage limits ensuring notional exposure stays within configured risk tolerance.

Maximum Drawdown Controls

Drawdown is the decline from a portfolio's peak value to its current value. Maximum drawdown (MDD) is the largest such decline over a given period — the worst-case loss an account experiences.

Professional risk managers set hard drawdown limits:

- Daily drawdown limit: if the account drops more than X% in a single day, stop trading for the rest of that day. This prevents a bad day from becoming a catastrophic day. - Total drawdown limit (kill switch): if the account drops more than Y% from its peak (or from the starting balance), halt all trading and review. FerroQuant uses a 5% portfolio drawdown threshold as a kill switch — when breached, all signal engines halt and no new positions open.

Drawdown limits serve a psychological function too. They prevent the destructive cycle of doubling down on losing trades trying to "get back to even" — a behaviour that turns manageable losses into account-destroying ones.

A key principle: the stop is the truth. Your maximum drawdown is determined not by your take-profit targets but by how quickly your stop-loss gets hit. Wide stops = large drawdowns per trade. Tight stops = small drawdowns but more stop-outs.

Correlation and Portfolio-Level Risk

Individual position sizing is necessary but not sufficient. You also need to manage correlation risk — the danger that multiple positions move against you simultaneously.

If you hold long positions in BTCUSDT, ETHUSDT, BNBUSDT, and SOLUSDT simultaneously, you do not have four independent trades. These assets are highly correlated: when Bitcoin drops 10%, altcoins typically drop 15-20%. Your "diversified" portfolio is effectively one large Bitcoin long.

Professional algo traders manage correlation through:

- Market caps on concurrent positions per correlated group (e.g., maximum 3 simultaneous long crypto futures) - Limiting exposure to a single sector or instrument family (e.g., maximum 20% of portfolio in any single asset class) - Cross-market diversification: holding positions in crypto, forex, and commodities simultaneously reduces correlation because these markets are driven by different fundamentals

FerroQuant's risk engine tracks the number of open positions per market and per direction, enforcing configurable limits. The system intentionally spreads across crypto futures, spot, forex, and commodities to reduce the correlation inherent in holding only crypto.

Volatility-Adjusted Sizing and ATR

A limitation of fixed fractional sizing is that it uses a fixed stop-loss percentage. In practice, different assets have different natural volatility: BTC might move 3% in a day while EUR/USD moves 0.5%. Setting the same 2% stop on both means BTC positions get stopped out constantly while EUR/USD positions almost never are.

The solution is volatility-adjusted stops using Average True Range (ATR). ATR measures the average daily range of an asset over a lookback period (typically 14 bars). Setting your stop-loss at 1.5× or 2× the ATR means you are allowing for the asset's natural volatility before considering the trade failed.

This approach ensures stops are neither too tight (getting stopped out by noise) nor too wide (taking excessive risk per trade). The position size calculation remains the same — you still risk 1% of account — but the stop distance is now proportional to the asset's actual volatility.

FerroQuant uses ATR-based stops by default across all strategies. Stop distance = ATR-14 × multiplier (configurable per strategy). This prevents the common mistake of applying cryptocurrency-calibrated stop distances to forex pairs or vice versa.