Ask ten people what a trading bot is and five will say "a scam," four will say "a money printer," and one will shrug. All ten are wrong. A trading bot is something much more boring and much more useful: a program that follows a trading plan to the letter.
That's it. And that is exactly why a good one works.
A bot (or automated system, or trading robot, all the same thing) is software connected to a trading platform. It has rules written into it: "if A happens and B happens, buy; put the stop here and the target there; when either one is hit, close." The program watches the market, and when the conditions line up, it executes. It doesn't ask you first, it doesn't hesitate, and it doesn't stare at the account balance with a knot in its stomach.
The key word isn't "robot." It's rules. A bot doesn't think, doesn't improvise and has no intuition: it does exactly what its system says to do. If the system is bad, the bot loses money with admirable discipline. If the system is good, the bot executes it better than any human, because it never skips a rule and never skips a day.
A bot doesn't turn a bad plan into a good one. It turns a good plan into one that actually gets followed. Everything else (whether the plan is any good, whether the numbers are real) is something you verify, not something you believe.
Not because it's smarter. Because you have an amygdala and it doesn't. When there's money on the line, the human brain does strange things, all of them documented to death: holding losers hoping they come back, snatching winners early out of fear of giving them up, trading bigger after a bad day to "make it back," and abandoning the plan on precisely the day the plan mattered most.
The industry publishes the damage. Wherever regulators force brokers to publish the figure, the share of retail customers who lose money on leveraged products lands between roughly 70% and 85%. Those people do not lose for lack of information. Information is free and endless. They lose on execution: doing, with their nerves, the opposite of what they knew they should do.
A bot doesn't have a good day or a bad day. That's the entire product.
Take the Algomyth engine, the engine behind this site, as a concrete example, because it shows the anatomy of a bot with actual judgment behind it:
You can read the whole mechanism in how it works, and the reasoning behind that specific hour in the 8:30 data window and the 9:30 open.
When somebody says "trading bot" they could mean five completely different things, with risks that have nothing in common. A good chunk of this industry's bad reputation comes from throwing them all in one bucket.
| Type | What it does | What it leaves to you | Where it usually breaks |
|---|---|---|---|
| Rule-based futures system | Fixed logic in a set window, stop and target on entry | Turning the platform on | If it was fitted to the past, it unravels |
| Crypto grid bot | Buys and sells in fixed steps | Constant babysitting | One trend against you stacks losses with no ceiling |
| Signal service | Alerts you; you place the order | Being at the screen during market hours | Your execution ruins the good signal |
| Copy trading | Mirrors another person's trades | Choosing who to copy | You don't know their rules or when they change them |
| Machine-learning model | A statistical model decides | Maintaining the model | It's opaque: when it fails, nobody knows why |
None of them is a scam by definition, and none is a guarantee. What changes is what they demand from you and what you can check before you pay. That's what how to choose a trading bot is about, and the differences with copying a person are laid out in copy trading vs bots.
Numbers on their own say nothing. What matters is the relationship between them. If you risk one unit to make three, every win is worth triple every loss, and the consequence is counterintuitive: you don't need to be right most of the time. At three-to-one, breakeven sits at winning one trade in four, about 25%.
Run a hundred trades at a 40% hit rate, as a working hypothesis and not a promise. Forty winners at three units are 120 units; sixty losers at one unit are 60. Net: 60 units, minus commissions. Nothing spectacular. Consistent, which is the whole point.
This is also why bad stretches are unavoidable. If you're wrong 60% of the time, four or five losses in a row aren't a malfunction. They are arithmetic. The full calculation is in risk-reward and win rate, and the honest version of the money question is in how much a trading bot actually makes.
None of the four has anything to do with the system. All four are about what its owner does around it.
This is where serious bots separate from toys. An honest seller of an automated system can show you a reproducible backtest: a simulation of the system over historical data, run with a standard industry tool that you can run yourself, like the Strategy Analyzer inside NinjaTrader 8. You load the strategy, set the same period, hit Run, and check that the numbers are the ones you were shown.
A snake-oil seller shows you screenshots of profits in accounts you can't inspect, testimonials from people who don't exist, and monthly percentages no fund on the planet produces. We wrote a whole piece on how to spot a trading bot scam in five minutes, because the field is crowded and you should be able to defend yourself, from us or from anybody else.
The published Algomyth engine backtest runs from January 2022 to September 2026 on a mid-size account: $87,359 net across 1,164 trades, with the worst drawdown of the period at $5,273. That figure is already net: commissions of $1.90 per contract round turn and a tick of slippage (about $2.40 on a micro Nasdaq contract) are inside the number, not waiting to be subtracted from it. It is a simulation over historical data and past results do not guarantee future ones. The steps to reproduce it are on the performance page.
Hypothetical performance. Those figures come from a backtest, not from a live account. Simulated results are prepared with hindsight, carry no financial risk, and cannot fully reflect real execution, slippage or liquidity. Past performance is not indicative of future results.
The full report, the max drawdown, the losing months and the live signal log that archives every new signal with its outcome are all on one page.
A trading bot makes sense if you want exposure to trading without turning it into a second job: no candlestick courses, no early mornings in front of a chart, no wrestling with your own nerves. And it makes even more sense combined with a funded account, where the capital being traded isn't yours.
It doesn't make sense if you're after excitement (this is deliberately boring), if you expect to win every day, or if the money you'd put at risk is money you need. In that case it isn't your moment, and that's fine. If you want the whole picture, with costs, taxes and everything else around the decision, it's in the complete guide to trading bots.
Yes. Automating your own orders in your own account is legal in the US. What's regulated is managing other people's money or advising professionally for compensation, which requires registration with the CFTC and NFA. Buying software and running it on your own account is not that. Your results still have to be reported to the IRS, and futures get their own treatment under Section 1256.
It depends on when it trades. A system with a single daily window only needs the platform open during that window. If you'd rather not think about it, a small virtual private server keeps it running 24 hours a day for a few dollars a month, which is what most people end up doing.
There is no reliable answer, and anyone who gives you a specific one is inventing it. A statistical edge plays out over dozens or hundreds of trades and can easily start underwater. Judge it by the quarter, against its own historical behavior, not against your hopes.
The bot stops watching, but the stop and target orders it already sent are alive on the broker's servers. That's exactly why it matters that a system places real orders instead of holding mental stops: if your machine dies, the risk stays capped. Check the position as soon as you're back online.
Technically yes, and it's what most people do once a system has proven itself: the same trade gets copied across several funded accounts at once. Each account is its own evaluation you have to pay for and pass, and each prop firm sets its own limits on how many accounts one person can hold. The mechanics are in our piece on copying trades across accounts.
Everything we write comes back to one idea: verify, then decide. The live terminal exists for exactly that.
Unlock Algomyth Live