Why this article exists

Stop losses and drawdown caps are the unglamorous core of risk management. They get less attention than entry strategies because they do not feel like the source of edge — they feel like a cost imposed on edge. In practice, the rate at which retail traders blow up accounts is driven less by bad strategy choices than by undisciplined or absent stop-loss frameworks. A mediocre strategy with disciplined stops outperforms a brilliant strategy with poor stops over any period long enough to include normal variance.

This article covers the practical implementation of stops and drawdown caps for UK practitioners running AI-driven systematic strategies. The principles are universal but the venue-specific details (UK-regulated brokers, FCA leverage caps, guaranteed stops in spread betting) are the focus.

Stop loss methodologies

Three methodologies cover the bulk of professionally used stop-loss frameworks. Each fits different strategy types; mixing them within a single strategy is generally a sign of confusion rather than sophistication.

Fixed-percentage stops

The simplest approach: set the stop at a fixed percentage below the entry price (for long positions). A 2% stop on every trade, applied uniformly. The advantage is simplicity and consistency; every trade has the same risk profile. The disadvantage is that the same percentage is too tight on high-volatility assets and too wide on low-volatility ones — a 2% stop on a Bitcoin position is likely to be triggered by normal noise; a 2% stop on a FTSE 100 ETF is unusually wide. Fixed-percentage stops are appropriate when an entire strategy operates on assets with comparable volatility profiles.

ATR-based stops

A volatility-aware approach: set the stop at a multiple of the asset’s Average True Range. A common configuration is 2x ATR(14), giving the position room equal to twice the average daily range over the past 14 periods. This naturally widens stops on high-volatility assets and tightens them on low-volatility ones. It is the standard professional approach for systematic strategies trading multiple assets, because it keeps the per-trade risk profile (in absolute terms, after position sizing) consistent across very different markets.

Practitioner note. Combine ATR stops with the 1% position-sizing rule and you have a clean framework: position size = (1% of capital) / (2x ATR). This sets every trade’s maximum loss at the same percentage of capital regardless of which asset is being traded. The bookkeeping is consistent across the portfolio.

Structural stops

Stops placed at meaningful technical levels: previous swing lows for long positions, previous swing highs for shorts, key moving averages, or established support/resistance zones. The advantage is that the stop level is justified by market structure rather than arbitrary distance — if the stop is hit, the technical thesis behind the trade has been invalidated, which is genuinely useful information. The disadvantage is that structural stops are often farther from entry than ATR or fixed stops, which forces smaller position sizes and reduces strategy capacity.

Structural stops are most appropriate for discretionary or longer-timeframe systematic strategies. For high-frequency or short-timeframe strategies, ATR stops are usually the better choice because the technical structure at small timeframes is too noisy to give clean stop levels.

Stop loss execution: standard vs guaranteed

Once the stop level is decided, the next question is execution. Two structures exist at UK-regulated brokers, with different costs and reliability profiles.

Standard stop loss orders

A standard stop loss converts to a market order when the stop price is touched. Execution then occurs at whatever the next available bid (for sell stops) or offer (for buy stops) is. In normal market conditions on liquid assets, this is essentially the stop price. In fast-moving conditions — a sudden gap, a major news event, or an asset that has thinly traded overnight — the execution can be substantially worse than the stop level. The 2015 Swiss franc unpegging produced execution prices several percent worse than stop levels for many traders, including those at major brokers; this is the type of event a standard stop does not protect against.

Guaranteed stop loss orders

Available at the major UK spread betting and CFD providers (IG, CMC Markets, others) for an additional premium. The broker undertakes to execute at the stop level regardless of market conditions, taking on the gap risk themselves. The premium varies by asset and market conditions but is typically reasonable for the protection provided. For traders who cannot tolerate the unbounded loss potential of standard stops during gaps — most retail traders should be in this category for at least their large positions — guaranteed stops are worth the cost.

A pragmatic policy: use guaranteed stops on positions large enough that a gap-filled execution would meaningfully damage the account, use standard stops on smaller positions where the additional premium is not justified by the protection. The threshold depends on the trader’s account size and risk tolerance; many UK practitioners use guaranteed stops on any position representing more than 5% of account capital.

See xTradeGrok’s stop-loss configuration options. Open an xTradeGrok account in minutes →

Trailing stops

A trailing stop moves with the position when the trade is profitable, locking in gains as the price advances. Configured properly, trailing stops let winners run while taking the trader out of trades that reverse. Configured poorly, they cause premature exits on routine pullbacks within sustained trends.

Two trailing-stop methodologies are professionally common. The first is a fixed-distance trail: stop trails 2x ATR below the highest price reached since entry (for long positions). This is mechanical and consistent. The second is a structural trail: stop moves to each new significant swing low as the trend develops. This requires the underlying market structure to be cleanly identifiable, which is reliably the case on daily and 4-hour timeframes and unreliably the case on lower timeframes.

The mistake to avoid is using a trailing stop on every trade by default. Some strategies are designed for fixed-target exits and benefit from trailing stops only in specific conditions (typically only after the position has reached the first target). Bolting trailing stops onto a strategy that was not designed for them often degrades performance rather than improving it.

Account-level drawdown caps

Stop losses limit individual trade losses; drawdown caps limit account-level losses. These are separate mechanisms operating at different levels and both are necessary. A daily drawdown cap pauses all trading once accumulated losses for the day reach a threshold. A weekly cap does the same at the weekly level. A maximum drawdown cap terminates strategy operation entirely if the account falls below a threshold below the historical peak.

Daily caps

A 3% daily cap is the standard floor for systematic UK strategies. Some implementations use 2%; tighter is generally better. The cap stops the bot once cumulative day losses exceed the threshold, regardless of whether individual trade stops have been hit. The purpose is catching cases where a strategy fires repeatedly into a deteriorating market and individual trade losses respect 1% rules but accumulate damage exceeding what a single bad trade could produce.

Weekly caps

A 6–8% weekly cap, depending on strategy aggressiveness. The weekly cap catches sustained adverse runs across consecutive days that the daily cap individually permitted. Without a weekly cap, an account can grind a 25% loss across multiple weeks while never hitting any individual day’s circuit breaker. The cap forces the trader to step back and assess whether the strategy is suffering normal variance or is in a regime mismatch that warrants pausing.

Maximum drawdown cap

A 15–20% maximum-drawdown cap from any historical equity peak. If the account falls this far below its high-water mark, all systematic strategies pause until the trader has reviewed and explicitly chosen to resume. This cap exists because at this level of drawdown, normal strategy behaviour is no longer the right hypothesis; either the strategy has decayed, the market regime has changed beyond what the strategy is built for, or there is an implementation issue that has not yet surfaced. Continuing to trade without explicit review at this drawdown level is how recoverable losses become unrecoverable ones.

Practitioner note. Most UK-regulated brokers do not implement these caps natively at the account level — they are implemented in the bot’s logic. This places responsibility on the trader to configure them in whatever framework is generating signals and executing. Manually-traded accounts depend on trader discipline, which is an unreliable substitute. Either accept the unreliability or run the strategy through a framework that enforces caps mechanically.

Trade with mechanical drawdown enforcement through xTradeGrok. Get started with xTradeGrok →

When stops fail to execute as expected

Standard stops on UK-regulated venues execute reliably in normal market conditions on liquid assets. Three categories of conditions can produce unexpected behaviour, and the practitioner should understand each before being surprised by it.

Weekend gaps in FX. Forex markets close from Friday evening (UK time) to Sunday evening. A position open over the weekend can gap on Sunday open by an amount substantially exceeding any Friday-evening stop level. Stops execute at the next available price, which can be several percent worse than the stop. The defence is to flatten weekend exposure or use guaranteed stops on positions held over the weekend.

Earnings releases on individual equities. UK equities reporting earnings outside market hours can gap dramatically on the next open. Same dynamic, same defence: either avoid holding the position over the release or use a guaranteed stop.

Crypto exchange outages. Crypto markets are notionally 24/7 but individual exchange operations are not always continuous. An exchange suffering a temporary outage during a fast move can result in stops failing to execute at any price during the outage window. The defence is using exchange-side stops where available and not concentrating capital on a single exchange.

Frequently asked questions

Should every trade have a stop loss?

For systematic strategies, yes — the stop is part of the trade specification. For long-term core holdings (multi-year positions in companies with strong fundamentals or in broad-market ETFs), stops are often inappropriate; the position is held through volatility and the exit decision is made on fundamentals, not price. Mixing the two in a single account requires clear delineation: a “core” sub-portfolio without stops and a “tactical” sub-portfolio with stops. The mistake is letting a strategy with stops drift into one without.

How tight should stops be on AI trading strategies?

Tight enough to limit damage, wide enough to avoid being shaken out by normal noise. ATR-based stops with a 2x multiplier are the standard starting point. If the strategy is being stopped out frequently by what turn out to be transient moves (the price reverses to the entry level shortly after the stop fires), the stop is too tight relative to the asset’s normal volatility. Widen to 2.5x or 3x ATR and reduce position sizing to keep maximum loss constant.

Are guaranteed stops worth the cost?

For positions large enough that a gap-filled execution would meaningfully damage the account, yes. The premium is typically modest (often 0.3 pips on major FX, more on volatile assets) and the protection covers the rare but consequential cases where standard stops fail. For very small positions, the premium-as-percentage-of-position is high enough that standard stops are usually the better economic choice. The threshold depends on account size; many UK practitioners use guaranteed stops on positions above 5% of capital.

What is the right relationship between position size and stop distance?

Position size = (intended account risk per trade) / (stop distance from entry). A 1% account-risk-per-trade rule on a £50,000 account allows £500 maximum loss per trade. If the stop is 50 pips away on a forex pair, the position size that risks £500 is calculated from pip value: £500 / (50 × pip value). This calculation is the core of position sizing and must be done for every trade. Spreadsheet templates and bot frameworks can automate it; the calculation must not be skipped on any individual trade.

Should trailing stops be used by default?

No. Trailing stops fit some strategies and degrade others. Use trailing stops where the strategy specification explicitly includes them and where backtesting has shown them to improve outcomes. Adding trailing stops to a strategy not designed for them is a common source of self-inflicted underperformance.