Simply put (maybe), stop-orders are designed to exit a position when it moves against the trader beyond a predetermined threshold. Stop orders can be seen to be the last line of defense as an automated risk manager built into the execution process. And yet, despite their apparent utility, stop orders remain one of the more divisive elements in trading.
Ask five fund managers how they use stops and you’re likely to get six different answers. Some rely on them religiously, weaving them tightly into their signal logic and execution protocols. When I first entered the CTA space, I was struck by what seemed a super-sophisticated use of stops. It looked like these systems could perform magic tricks I had never seen before as a trader. I must admit I am still taken but that initial awe I experienced some 25 years ago. That made it all the more surprising to learn that some systematic managers and CTAs avoid using stops altogether. Instead, they rely on discretion, manual oversight, or soft, internally tracked risk levels that remain invisible to the market. Or indeed, different approaches that vary again. The reason for these differences is not purely philosophical; often, the choice is shaped just as much by technology.
At the heart of the stop order debate (maybe) is a tension between control and execution quality. Another argument is that traditional stop orders, as commonly used in trend-following systems, can weaken a strategy’s robustness. A traditional stop market order, where the position is sold at the next available price once the stop level is breached, can protect a portfolio from further losses, but also risks poor fills in illiquid or volatile conditions. Conversely, stop-limit orders may offer better price control but run the risk of not executing at all, leaving managers exposed.
Harold de Boer of Transtrend recalls how, in the early 1990s, protective stops were considered essential to trend-following. “Many potential investors wouldn’t consider a CTA that didn’t use stops,” he said. “However, our research led us to question the benefits of such stops.” De Boer´s skepticism was validated during the Soviet coup of August 1991. Transtrend was running a test portfolio at the time, and while some positions moved sharply against them, others moved in Transtrend´sfavor, leaving the net effect close to neutral, just as intended in a well-diversified portfolio. “Two days later, the coup was undone, and the market made mirrored moves. The winners of the coup became the losers of the ‘re-coup.’ And the losers of the coup became the winners of the ‘re-coup.’ That is, provided these positions weren’t stopped out on the first day,” he noted, illustrating how the use of protective stops could have undermined diversification and caused losses due to slippage and reduced exposure. “This event confirmed for us that investment decisions should really be made and evaluated from a portfolio perspective, rather than focusing too much on positions in individual markets”, de Boer underlines.
The way stops are implemented can vary widely depending on a manager’s technology stack. In systematic firms, stop logic is often embedded directly into the signal engine or order management system (OMS). These systems will automatically trigger an order once the stop condition is met, sometimes routing it through execution algorithms that account for volume and market conditions. At more discretionary shops, stops may be managed manually or through alerts that prompt a trader to act. Often the manager may use more discretion and market awareness than a purely automated system can provide.
There is also the question of where the stop lives. Stops can be placed at the exchange level, visible in the order book once triggered, or held “synthetically” at the broker or EMS level, where they aren’t exposed to the market until activation. This is not a trivial distinction. Some managers avoid exchange-visible stops out of concern they will be “hunted” and orders get triggered by market participants who detect clusters of stop orders and intentionally move prices to activate them. Whether such behavior is systemic or more anecdotal, the perception is real enough to influence execution strategy.
Technology has evolved to give managers more nuanced control over how and when stops are used. Execution management systems (EMS) now allow for highly customized conditional orders, incorporating time-of-day logic, volume thresholds, volatility filters, and more. A manager can instruct the system to only trigger a stop if the price breaches a level and volume drops below a certain threshold, or to delay a stop during news events. For those managing hundreds or thousands of positions, this granularity is no longer a luxury but a necessity. De Boer also emphasized a responsible investing angle: “In addition to potentially paying large liquidity premiums and harming portfolio diversification, stop-loss orders can significantly destabilize the market. This is especially the case when multiple parties use such orders simultaneously, as the execution of one stop-loss order can trigger a chain reaction.” This underlines the systemic risk stops can pose under stress scenarios.
However, even with these technological advancements, stops remain a point of contention in performance attribution.While often associated with exits and risk control, stop orders can also play a more active role in entering positions. Some systematic strategies use stop orders to initiate trades only once a price threshold has been breached, treating stops as confirmation signals rather than fail-safes. In these cases, stops are not simply defensive tools but serve as part of a broader tactical logic to filter noise and validate momentum.
On the exit side, stops are commonly employed to protect profits once a position moves in the manager’s favor. Trailing stops, for instance, offer a way to systematically lock in gains without fully exiting a position prematurely. The effectiveness of this approach hinges on how well the stop parameters are calibrated to the volatility and liquidity of the asset.
And while stops do not generate alpha in the traditional sense, believers believe they can improve a strategy’s efficiency and consistency. By minimizing tail risk and enforcing a degree of trading discipline, well-designed stop frameworks supported by strong technology can enhance overall portfolio performance, proponents argue.
Some argue that they cap downside but also limit potential upside by exiting positions prematurely that may have recovered. Others see them as essential discipline, a way to automate exits in fast-moving markets where hesitation can be costly.
An emerging compromise in the stop-order debate is the use of soft stops. These are internally tracked levels that, when breached, trigger a review or a discretionary execution decision rather than an automatic market order. These soft stops may live entirely inside a portfolio management system (PMS), invisible to the market, but visible to traders, portfolio managers, and risk officers. For some, this offers the best of both worlds: the discipline of predefined exit levels with the flexibility of human oversight.
Of course, soft stops place greater demands on the firm’s internal tech stack. Systems need to not only track and alert but also escalate breaches in a way that fits the firm’s workflow. Integrating this kind of conditional monitoring across PMS, EMS, and risk tools isn’t trivial, and often becomes a project in itself which blends investment process, technology design, and organizational behavior.
One area where technology has yet to catch up is in backtesting the real-world effects of stop orders. Many strategies that look attractive on paper begin to fray once realistic stop order logic, including potential execution delays, slippage, and partial fills, are introduced. Sophisticated managers now simulate not only the signal but also the path the order takes through the execution system, modeling not just whether a stop would be hit, but how and at what cost.
Ultimately, the question isn’t whether to use stop orders, but how consciously they are implemented and supported by technology. Even among managers trading similar models or signals, the way stop logic is implemented (or not) can lead to significant performance dispersion. Differences in stop placement, activation logic, and order routing create subtle but persistent deviations. One manager might be stopped out and miss a recovery, while another captures the full move. Over time, these micro-level differences accumulate often enough to separate top-quartile performers from the rest. As strategies become more crowded and alpha harder to find, how a firm handles stops may quietly become a key source of differentiation.
And increasingly, it’s the strength of their technology stack that determines how well that match holds up under pressure.But we’ll stop here.
