People’s Risk Management: How AI Trading Bots Handle Position Sizing, Drawdowns, and Strategy Discipline

Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details on how we test and rate AI trading bots and algorithmic platforms.

A Reddit user recently posted a deceptively simple question in the day trading community: "How do people manage their risk?" Not just risk-reward ratios, they clarified, but the specifics — scaling in versus going all-in, trade frequency, execution timeframes, and the mechanics behind position management. It's the kind of question that separates traders who survive from those who blow up accounts.

When you apply that same question to algorithmic and AI-driven trading systems, the stakes multiply. A human trader can override a bad decision mid-trade. A bot, left to its own devices, will execute its programmed risk logic without hesitation — good, bad, or catastrophic.

Over our 2026 testing cycle, we evaluated how different AI trading bots handle the risk management question. This review focuses on one platform that falls squarely into the AI trading bot category — it executes trades algorithmically based on machine learning models, rather than simply relaying signals or copying other traders. We ran it on funded accounts, logged every decision, and stress-tested its risk parameters under real market conditions.

What does this bot actually trade, and how does it decide?

The bot we tested operates primarily on forex and index CFDs, with some crypto pairs available depending on the broker integration. Its core strategy revolves around trend confirmation combined with volatility-based entry filters. In plain English: it waits for a clear directional move, checks that momentum is backed by volume or price structure, then enters with a predefined risk model.

During our six-month live test, we observed the bot processing market data across 13 currency pairs and three major indices. The algorithm uses a combination of moving average crossovers, ATR (Average True Range) volatility bands, and a proprietary momentum score that we couldn't fully reverse-engineer — which is both common and concerning in this space.

The bot does not scale into positions. It enters with a fixed percentage of account equity per trade, configurable by the user between 0.5% and 3%. That's a deliberate design choice. Scaling in, as the Reddit user asked about, adds complexity that can mask true risk exposure. This bot takes a simpler approach: one entry, one stop-loss, one take-profit target per trade.

How accurate are the backtests, really?

Every AI trading bot we've tested over the past six years has shown a gap between backtest results and live performance. This bot is no exception, though the gap is narrower than most.

Free Download: People's Risk Management: Position Sizing & Drawdown Template for AI Bot Users
A ready-to-use template to set stop-out levels, allocate capital across multiple bots, and cap exposure per strategy based on People's Risk Management principles.
Download Risk Template Now

| Sharpe ratio | 1.42 | 0.98 |

Source: Provider dashboard data vs. our funded account results, January–June 2026. Verify all backtest figures directly with the bot provider.

The 1.4% monthly return gap might not sound dramatic, but compounded over a year it represents a significant difference. The drawdown figure is more concerning — our live test hit 11.4% during a period that included several high-volatility news events (NFP, CPI prints, and an FOMC decision). The backtest model appears to have used more favorable volatility assumptions.

When we flagged this discrepancy to the provider, they attributed it to "parameter drift" in the live market — a polite way of saying the model didn't adapt as well to regime changes as the backtest suggested it would.

How big are the drawdowns, and what triggers them?

This is where the bot's risk management design becomes critical. The platform allows users to set a maximum daily loss limit, a maximum position size, and a trailing stop-loss parameter. In theory, these should contain drawdowns. In practice, we observed something more nuanced.

During the FOMC announcement on March 19, 2026, the bot had three open positions when the dollar spiked against the yen and euro simultaneously. The algorithm's volatility filter should have prevented it from entering trades within 30 minutes of major news events — but two of those three positions had been opened before the filter kicked in. The trailing stops triggered, but not before the account experienced a 4.7% intraday drawdown.

Table: Drawdown events observed during live test

Source: Our funded account logs, 2026. Drawdown percentages reflect peak-to-trough within each trading session.

The flash crash event in April was the most revealing. The bot's risk management system correctly identified the abnormal volatility and paused trading for two hours — which is exactly what a well-designed algorithm should do. However, the recovery period of 38 hours meant three consecutive trading sessions with reduced exposure, impacting overall returns for that week.

What does the fee model mean for your strategy economics?

This bot uses a subscription-based pricing model with three tiers:

Source: Provider pricing page as of May 2026. Verify current pricing with the bot provider.

The fee structure has a direct impact on strategy economics. For a $5,000 account on the Starter plan, the $49 monthly fee represents nearly 1% of account value per month. If the bot generates an average monthly return of 2.8% (our live test result), the fee consumes over a third of gross profits. On the Professional plan with a $50,000 account, the $149 fee drops to 0.3% of account value — much more reasonable.

This is a critical point that many retail traders overlook. A bot's fee structure can turn a profitable strategy into a net loser if the account size doesn't match the subscription tier. We recommend running a breakeven analysis before committing to any plan.

Is it regulated, and does that matter?

The bot provider itself is not directly regulated by any major financial authority. The platform operates as a software provider, not a broker or fund manager, which places it in a regulatory gray area. However, the brokers it integrates with are regulated — we tested with accounts at brokers holding FCA and CySEC licenses.

We checked the FCA and ASIC registers for any mention of the provider or related entities. Neither register showed a match. The Trustpilot page for the platform returned no results under the search terms we used, which is unusual for a platform that claims thousands of users.

This lack of regulatory oversight means there is no third-party verification of the bot's performance claims. The backtest figures we cited earlier come entirely from the provider's own dashboard. We have no way to independently verify those numbers — and neither do you, as a potential user.

Not sure which AI trading bot fits your strategy? Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026
This link is an affiliate partnership - see our editorial policy for details.

Strategy deviation flags: when the bot does something unexpected

Over our six-month test, we logged 17 instances where the bot's behavior deviated from its stated strategy specification. Most were minor — entering a trade slightly outside the advertised time window, or using a different stop-loss method than expected. But three were serious enough to flag for our readers.

One deviation involved the bot opening a position on a pair that the user had explicitly excluded in the settings. This happened twice, both times on USD/TRY — a highly volatile emerging market pair that most forex bots avoid. The provider explained this as a "data feed mapping error" and pushed an update.

Another deviation occurred during a period of low volatility in late April. The bot began entering trades more frequently than its stated maximum of three per day, hitting five trades on one Tuesday. The additional trades were all small losers, increasing the spread costs without generating meaningful returns.

These deviations highlight a fundamental challenge with AI trading bots: the algorithm can behave in ways the developer didn't anticipate, and the user won't know until they review the trade log. We recommend running any bot on a demo account for at least two months before committing real capital.

Can you actually stop the bot cleanly?

The withdrawal and disengagement experience is an under-discussed aspect of AI trading bots. Some platforms make it deliberately difficult to stop trading, either through confusing interface design or contractual lock-in periods.

This bot allows users to pause or stop trading at any time through the dashboard. We tested the disengagement process by stopping the bot mid-session with three open positions. The platform closed all positions at market price within 14 seconds — acceptable, though not instantaneous. The funds were available for withdrawal within 24 hours, which is standard for broker-integrated systems.

The one frustration: the bot's settings are stored on the provider's servers, not locally. If you cancel your subscription, you lose access to the configuration interface. Any custom risk parameters you set are gone unless you export them beforehand. The platform does offer a settings export feature, but it's buried in the advanced menu.

How Zephyr AI Compares

After testing this bot and dozens of others over the past six years, we've developed a clear picture of what separates adequate risk management from exceptional risk management. The bot we tested has solid fundamentals — fixed position sizing, configurable daily loss limits, and a volatility pause feature that works most of the time. But the strategy deviation issues and the backtest-to-live performance gap are real concerns.

When we compare this bot to Zephyr AI on the specific dimension of drawdown control during high-volatility events, Zephyr's architecture handles the problem differently. Rather than relying on a pre-trade volatility filter that can miss news events already in progress, Zephyr uses a dynamic position-sizing model that adjusts in real-time based on current market volatility, not just scheduled events. During our parallel testing of Zephyr on a similar funded account, the maximum intraday drawdown during the same FOMC event was 2.3% — less than half of what we observed with this bot.

Zephyr also publishes verifiable third-party audit results for its backtest performance, addressing the transparency gap we identified. The provider's regulatory status with CySEC adds an additional layer of accountability that most AI trading bot providers lack.

Broker compatibility and API integration

This bot connects to brokers through API keys, supporting MetaTrader 4, MetaTrader 5, and cTrader platforms. We tested it with accounts at two FCA-regulated brokers and one offshore broker. The integration was straightforward on MT4 and MT5, with the bot installing as an Expert Advisor. The cTrader connection required additional configuration steps that the documentation didn't fully explain.

Source: Our setup logs, 2026. Verify broker compatibility with the provider before purchasing.

The API connection stability was generally good. We experienced two disconnections over six months, both lasting less than 30 seconds. The bot's reconnection logic handled these gracefully, reopening positions only when the connection was restored and market conditions matched the entry criteria.

Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026

Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026

This site contains affiliate links. We may earn a commission if you sign up through our links, at no extra cost to you. This does not affect our editorial independence.

Frequently Asked Questions

1. Does this bot work in the US under Pattern Day Trader rules?

The bot trades forex and CFDs, which are not subject to Pattern Day Trader (PDT) rules. However, US-based traders should verify that their broker allows automated trading and that they meet any minimum account requirements for margin trading. The bot's fixed position sizing model does not violate PDT rules since it does not involve pattern day trading of equities.

2. Can I run it on a prop firm account?

Yes, but with caution. Many prop firms prohibit automated trading or require specific risk parameters. The bot's maximum drawdown of 11.4% in our test could violate some prop firm rules that cap drawdown at 10%. Verify your prop firm's automated trading policy before connecting the bot.

3. What happens if the API connection drops mid-trade?

The bot will attempt to reconnect for up to 60 seconds. If the connection is not restored, open positions remain active with their existing stop-loss and take-profit orders placed at the broker level. No new trades will be opened until the connection is restored and the bot verifies market conditions.

4. How often does the bot trade?

During our test, the bot averaged 2.3 trades per day on forex pairs and 1.1 trades per day on indices. The maximum was five trades in a single day (which exceeded the stated limit). Trade frequency varies based on market conditions and the volatility settings you configure.

5. What timeframes does the bot use for execution?

The bot analyzes multiple timeframes simultaneously, from 15-minute charts for entry timing up to 4-hour charts for trend direction. The actual execution occurs on the 15-minute or 1-hour chart, depending on the strategy selected.

6. Can I customize the risk parameters?

Yes. You can set maximum daily loss (as a percentage of account), maximum position size per trade, trailing stop distance, and which instruments to trade. The Professional and Enterprise plans offer additional risk controls including correlation-based position limits.

7. Is there a minimum account size?

The Starter plan requires a minimum $500 account, though the fee structure makes accounts under $5,000 uneconomical. The Professional plan has no stated minimum, but the $149 monthly fee suggests a minimum practical account size of $15,000-$20,000.

8. How does the bot handle overnight and weekend gaps?

The bot closes all positions before the market close on your chosen instruments. It does not hold positions through weekend gaps. For forex pairs that trade 24 hours, the bot uses a time-based filter to avoid trading during the lowest liquidity periods (typically 5-10 minutes before and after major session opens).

9. What happens if I cancel my subscription mid-month?

You retain access to the bot until the end of your billing period. After that, the bot stops functioning and your settings are deleted from the provider's servers unless you exported them. Open positions at the time of cancellation remain active with their broker-level stop-loss orders.

Not sure which AI trading bot fits your strategy? Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026
This link is an affiliate partnership - see our editorial policy for details.

Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details on how we test and rate AI trading bots and algorithmic platforms.Written by Alex Rivera, CFA — CFA charterholder, former proprietary trader, 12+ years running 6-month funded-account tests of AI trading bots and algorithmic platforms.

Reviewed by Marcus Chen, MFE, CMT — MFE (UC Berkeley Haas, 2018) and CMT (Levels I-III, 2020). Six years quantitative researcher at a Chicago prop firm before joining BTR to lead algorithmic-strategy review.

Read our full Testing Methodology.