How Much Can a Trading Bot Lose?

A trading bot can lose a large part of the money it trades, and with leveraged funds it can lose nearly all of it. In AlgolaneBots's own historical backtests, the worst falls ranged from about 20% to 53% over the main test periods, and one strategy fell close to 100% in a simulated version of the 2000-2002 crash. How much you could lose depends on what the bot trades, how much leverage it uses, how it sizes positions, and what you do when a fall starts.

The short answer in numbers

The figures below come from historical backtests, which are computer tests of each strategy's rules on past prices. Nobody earned these results, and future results will differ. Each strategy is shown next to its benchmark over the same period, so the fall can be judged in context rather than on its own.

Over the same periods, the yearly results were +26.5% for Steady Growth, +37.3% for Big Tech Trend, +15.1% for Gold Breakout at 2% risk (gold held: +11.0%), +71.9% for Market Switch and +129.1% for Leveraged Growth, against roughly +14% to +15% a year for the S&P 500 and +18% to +19% for the Nasdaq 100. Those returns came with the falls above; they are not separable.

Each strategy has a caveat that matters when reading these numbers:

• Steady Growth: its seven names were picked because they already did well, which flatters the test.
• Big Tech Trend: its stock list was chosen with hindsight, and three stocks made 81% of the profit.
• Gold Breakout: doubling trading costs cuts its yearly result roughly in half.
• Market Switch: nearly half of its tested profit came from UVXY, and it failed outside the US.
• Leveraged Growth: its rules may be fitted to the years after 2010, and leveraged funds can lose value quickly.

These caveats point the same way. A backtest that looks back at winners with knowledge of who won tends to understate how much a strategy could lose in a future nobody has seen.

Why the worst fall in a backtest is not the worst possible fall

The "worst fall" figure, often called max drawdown, is the largest drop from a peak to a low during the test. It is the most useful single number for judging risk, but it has a ceiling built into it: it only knows about the years tested.

Two examples from our own data show why that matters.

Market Switch fell 42% at its worst between 2011 and 2026, in March 2020. In a simulated three-fund version run over the 2000-2002 dot-com crash, it fell 100%, against 47% for the S&P 500. It took 259 months to get back to its pre-crash value in that simulation, compared with 78 months for the S&P 500. In the 2007-2009 financial crisis simulation, it fell 96% against 55% for the S&P 500.

Leveraged Growth fell 53% at its worst from 2011 onwards. In the simulated dot-com version it fell 91%, against 47% for the S&P 500, and took 115 months to recover compared with 78.

Those older simulations use a simplified three-fund version, because many of the funds these strategies trade did not exist then. They are rougher than the main backtests. But they make one point clearly: a strategy that looks manageable over 15 good years can behave very differently in a market it was not built around. For more on this, see how strategies held up in 2008, 2020 and 2022.

What drives a bot's losses

Leverage

Leveraged funds such as TQQQ aim for three times the daily move of the index. On a day the Nasdaq 100 falls 10%, TQQQ is designed to fall around 30%. Because these funds reset daily, a choppy market can also wear their value down even if the index ends roughly flat, an effect often called volatility decay. Our post on leveraged ETF decay explains the arithmetic.

Two of AlgolaneBots's strategies, Leveraged Growth and Market Switch, use leveraged funds. They had the deepest falls in the table, which is what you would expect.

Volatility funds

UVXY tracks short-term volatility futures with leverage. It tends to jump when markets panic and has lost most of its value over long periods. A strategy that holds it at the right moment can gain a lot; one that holds it at the wrong moment, or for too long, can lose quickly. Market Switch's caveat that nearly half its tested profit came from UVXY is a warning about concentration in one unusual fund.

Concentration

A strategy that can put most of its money in one or two names, as Steady Growth can, has fewer places for losses to be spread. If one name has a very bad stretch, the account feels it directly.

Position size per trade

For a strategy like Gold Breakout, the risk setting is the main dial. In the backtest from June 2004 to November 2026, 1% risk per trade had a worst fall of 23%, while 5% risk had a worst fall of 74%. Gold held fell as much as 45% over the same period. The risk per trade post covers this in depth.

Speed of the fall

Some falls are fast. Leveraged Growth's 53% fall happened between 20 February and 18 March 2020, under a month. A bot that checks once a day or once a month cannot react faster than its schedule. That is a design choice, not a flaw, but it means losses can build between checks.

Losses that are not about the strategy

Not every loss comes from the market. Some come from the plumbing.

AlgolaneBots runs these checks on every order run. They do not stop market losses, and they cannot rule out every possible failure, but they cut down on losses caused by software mistakes. Our post on the kill switch covers the last item.

Can a bot lose more than you put in?

For most retail setups that buy funds or stocks without borrowing, the most you can lose is what you put in. A fund can fall to near zero, but you do not owe more than the cost.

The picture changes with margin, short selling and leveraged products traded on margin, such as many forex and CFD accounts used with MetaTrader 5. Gold Breakout can sell short, and in MT5 accounts the broker's own margin rules apply. Many brokers offer some form of negative balance protection, but the rules vary by broker and country, so check yours directly. AlgolaneBots is software, not a broker, and does not hold customer money or control margin rules.

What you can control

You cannot control the market. You can control several things that decide how much a bad stretch costs you.

1. Which strategy you run. The table above shows a wide range of historical falls. A strategy with a 20% worst fall and one with a 53% worst fall are different experiences.
2. How much you allocate. Losing 50% of a small slice of your savings is very different from losing 50% of all of it.
3. The risk setting, where one exists, such as Gold Breakout's 1%, 2%, 3% or 5%.
4. Whether you mix strategies. In the backtest, a 50/50 example mix of Leveraged Growth and Gold Breakout, rebalanced monthly from 2011, had a worst fall of 39% against 53% for Leveraged Growth alone; its yearly result was +63.9% against +129.1% alone. This is an illustration from a historical backtest, not advice.
5. Your own reaction. Selling at the bottom of a fall turns a temporary drop into a permanent one. That is not an argument to hold anything forever; it is an argument to decide your limits before the fall, not during it.

How to use loss figures well

A worst-fall figure is most useful as a question to yourself: "If my account dropped this much, and stayed down for this long, would I stick with it?" Gold Breakout at 2% risk fell 41% and took 41 months after the bottom to recover. That is more than three years of waiting.

If the honest answer is no, a lower-risk choice or a smaller amount is the sensible response. Our post on how long strategies take to recover looks at that waiting time in detail.

Key takeaways

• Bots can lose a large share of an account; in AlgolaneBots's historical backtests, worst falls ranged from 20% to 53% from 2011, and up to 100% in a simulated dot-com crash.
• Always compare a bot's fall with its benchmark over the same period, and check how long recovery took.
• Leverage, volatility funds, concentration and position size are the main drivers of deep losses.
• Pre-trade checks and a kill switch reduce losses from software mistakes, not from the market.
• Backtest worst falls are a floor for planning, not a ceiling; the future can be worse.

See the risk before you commit

The safest place to learn how a strategy loses money is on paper. Start the free 90-day trial with a $10,000 paper account and watch each of the five strategies through good and bad weeks before deciding anything. Sign up here; no card is needed. This article is general information, not personal advice; see the risk notice.