Stop and Restart Rules Using Equity Curve Moving Average Filters

Trading Strategies

25 October 2026

The core logic of the stop and restart rules using equity curve moving average filters is just one sentence: stop opening new positions when the curve falls below the moving average, and resume after it climbs back above; but stopping is easy, while restarting is the part that truly decides whether this rule is useful. Most discussions focus on "which line to break before stopping" but rarely explain clearly "what conditions count as truly being able to come back." If the restart conditions are written vaguely, this filter will switch on and off repeatedly near the moving average, turning a normal pullback into repeated stop-outs.

OKX Exchange
A leading global cryptocurrency platform,suitable for both beginners and experienced traders.
New user benefit: 20% off trading fees upon registration!！

Stop Rules: First Make Clear What You Are Stopping

The role of the stop rule is to limit the boundary of losses and drawdowns, not to replace your entry system. It governs "whether you should still take on new risk," not "what to do with current positions."

An executable stop rule needs to clearly define three things at the same time:

Trigger condition. The most common one is the equity curve falling below a certain moving average. A historical case recorded by TradersWeek used a 150-period simple moving average to stop and restart trading in a certain market. For higher-frequency traders, a 20-period daily moving average is a defensible starting point; lower-frequency traders can consider 25 to 50 periods. The key is not to find the "optimal parameter," but to choose a round number and test its stability against small changes. If changing the moving average period from 20 to 22 causes a dramatic change in results, the signal itself is fragile.

Action after triggering. When the curve falls below the moving average, do you stop all new positions, or only cut position size by half? NexusFi's advice is to write a specific risk ladder, for example: "When daily equity falls below the 20-day moving average, reduce all position sizes by half starting from the next trading day; when the drawdown from peak exceeds 20%, stop new entries and return to simulated trading." A tiered response is less likely to be repeatedly triggered by noise near the moving average than a binary on/off switch.

Handling of existing positions. Stop rules usually only constrain new positions. Existing positions should be handled according to your original stop-loss and exit rules. Do not temporarily change the exit logic of your positions just because the equity curve has fallen below the moving average. Mixing stop rules with position management will cause the two sets of rules to interfere with each other.

Restart Rules: More Important Than Stopping, and Easier to Write Badly

Stopping is only the brake; restarting is the part that determines whether this filter helps you or hurts you in the long run. Arrow Algo's compilation lists "restart conditions" as the most often overlooked step.

An overly loose restart condition—such as "resume when the curve climbs back above the moving average"—will create a large number of false restarts near the moving average. When the equity curve repeatedly crosses above and below the moving average, you will switch back and forth between "stop—restart—stop again," and each restart may be immediately followed by the next drawdown. The solution is not to abandon the filter, but to add confirmation conditions to the restart.

The example given by NexusFi is worth referencing: the curve stays above the moving average for 5 consecutive trading days, and the drawdown from peak has narrowed to within 10%. Only when both conditions are met at the same time should you restore the original position size. Arrow Algo suggests adding a buffer band above the moving average, for example, the curve needs to be 2% above the moving average to trigger a restart, in order to reduce sensitivity near the boundary.

There is also a more conservative approach: after stopping, first return to simulated trading for at least 10 trading days to confirm that the strategy still produces acceptable results in the current market environment, and then resume live trading with half position size. The cost is missing part of the early rebound, but the benefit is avoiding repeatedly losing money when the strategy has truly failed.

What This Rule Can Really Help You Do, and What It Cannot Do

The value of the moving average filter is not in statistical alpha, but in behavioral constraints and tail risk control. NexusFi's judgment is direct: academic literature is not optimistic about performance persistence, and past performance has only weak predictive power for the future. The honest use of equity curve trading is to "prevent a catastrophic drawdown," not to "precisely determine when a strategy's edge is online and when it is dormant."

Actual test conclusions on TradingView regarding moving average filter strategies also point in the same direction: in most cases, after turning on equity curve trading, the win rate and net profit actually decline. The strategy author's original words were "using a simple indicator for equity curve trading has more downside risk than upside benefit." This does not mean the rule is useless, but rather shows that you should not expect it to improve returns. Its role is to help you lose less when the strategy is clearly not working, at the cost of possibly stopping frequently and missing part of the recovery in choppy markets.

A counterexample worth noting comes from research recorded by the CMT Association: directly adding a 200-day moving average to the equity curve resulted in lower profits and larger drawdowns; switching to a more sensitive 50-day moving average made the results even worse. This shows that moving average filtering is not a universal layer that is "safe once added." If your strategy itself frequently causes the curve to cross the moving average during normal volatility, the filter will only increase friction costs.

A Cleaner Alternative

If what you want is dynamic control of risk exposure, volatility targeting is a cleaner mechanism than equity curve moving averages. Its signal comes from the market's current realized volatility, not your own profit and loss curve, avoiding the endogeneity problem of "the signal and the strategy output coming from the same data." Changes in market volatility are smoother than the equity curve crossing the moving average, and also produce fewer false signals.

The proper positioning of the equity curve filter is as a behavioral guardrail: when you find yourself thinking about "adding positions to recover losses" after consecutive losses, a written rule can stop you. If you are already using volatility targeting to control position size, the moving average filter can serve as a second-layer check, but do not use it to replace position management.

What You Should Do Now

First write down your stop conditions (moving average period, action after triggering, how to handle existing positions), then separately write down your restart conditions (how many consecutive days above the moving average, what level the drawdown needs to recover to, whether simulated trading confirmation is needed). After both conditions are written, backtest them using your own trading records from the past 12 to 24 months: did the filter really reduce the maximum drawdown? How much profit was missed during the stop periods? If the backtest shows that the filter worsens risk-adjusted performance, do not use it. If it shows that it helps, post the rules somewhere you can see them every day, and then give it at least one complete drawdown cycle in live trading to verify.

OKX Exchange
A leading global cryptocurrency platform,suitable for both beginners and experienced traders.
New user benefit: 20% off trading fees upon registration!！