A NinjaTrader strategy sells for $39, lifetime updates included. You buy it, load it, and pay $100 for a 50k evaluation to trade it on. Eleven weeks later the account is closed on a trailing drawdown rule, you have bought a second evaluation, and you still cannot tell whether the strategy is broken or just having an ordinary bad run, because nobody ever showed you what an ordinary bad run looks like.
Spent: $239 and three months. The $39 was the smallest number in that sentence and the only one you compared.
Running an automated system has a cost stack, and the software sits near the bottom of it. Here is the stack for one person trading one funded account, in round US numbers.
| Line | Rough cost | How often you pay it |
|---|---|---|
| 50k evaluation fee | ≈ $100 | Every time an account closes and you replace it |
| Commissions | $1.90 a round turn, plus a tick of slippage | Every trade, about 250 a year at one trade a day |
| Real-time futures data | Low tens of dollars | Monthly, for as long as you trade |
| VPS, if you use one | Low double digits | Monthly |
| The strategy license | $0 to four figures | Once, or monthly forever |
| Months spent finding out | Never invoiced, still spent | Once per dead system |
Exactly one row is "the price of the bot", and it is the row a cheap system competes on. Every other row is charged by the market, the broker and the prop firm at the same rate whether the logic is excellent or worthless. That is the whole arithmetic: a bad system does not save you the other rows, it makes you pay them for nothing.
You are not really buying software. You are buying the right to spend evaluation fees, commissions and months of your life executing somebody else's logic. The license price decides almost none of that. What the logic does with your money decides all of it.
There is no shortage of free strategy code: forums, GitHub, video descriptions, a friend's ZIP file. Almost all of it shares three properties. No track record over a period long enough to mean anything, no stated risk model, and nobody responsible for it when the platform updates and it stops compiling.
That combination is not neutral. The testing cost lands entirely on you, and you pay it in the two currencies that hurt: evaluation fees and calendar time. A free strategy that takes four months and two blown accounts to disprove cost $200 and a third of a year. It was never free. It billed you late, in a different unit. The longer version, including the handful of genuinely useful free tools, is in free trading bots.
The subscription case hides best. A monthly fee moves the seller's incentive away from the system working and toward you staying subscribed, which is not the same goal. The trade-offs are in subscription versus license.
Ordinary numbers, not a worst case: six systems over two years, each abandoned after a couple of months, with fees and data running the whole time.
| Line | Arithmetic | Total |
|---|---|---|
| Six cheap systems | 6 × ≈$45 | ≈ $270 |
| Evaluations consumed | 9 × ≈$100 | ≈ $900 |
| Market data | 24 months × ≈$15 | ≈ $360 |
| Cash spent | ≈ $1,530 | |
| Time spent | 24 months | Not recoverable |
| Track record owned at the end | None |
Fifteen hundred dollars and two years, and the thing you were buying all along (evidence that a set of rules works) is the one thing you still do not have. It happens one $39 decision at a time, because each purchase is too small to think hard about.
Cheap is a price. Expensive is a thing you find out afterward.
Every figure behind the engines on this site is published with its costs stated in the open, and the live signal log keeps growing on top of it.
A paid system earns its price only if it hands you things a free one structurally cannot.
To make that concrete: the Algomyth engine takes one trade a day in the window that opens at 9:30 AM ET, when the US cash market opens and the day's volume arrives. The stop and the target go into the market at the same instant as the entry, the target is 3× the risk, and nothing is ever held overnight. The backtest runs day by day from January 2022 to September 2026 on a mid-size account: 1,164 trades, a 1.46 profit factor, a worst drawdown of $5,273, and $87,359 net, with the commission and a tick of slippage already taken out of every trade. With the target at three times the stop, the system does not have to be right most of the time: break-even sits at a 25% win rate, so being right one time in four is enough to hold the line.
Hypothetical performance. The Algomyth engine figures above come from a backtest over historical data, 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, real or simulated, does not guarantee future results.
You do not need to be a programmer to filter most of this out. Four questions, asked before money moves.
Anything that fails those four is not a bargain at any price. The warning signs, from doctored equity curves to guaranteed monthly percentages, are catalogued in trading bot scams and red flags, and the full evaluation checklist is in how to choose a trading bot. To test a track record rather than trust it, start with how to spot a fake track record.
None of which means expensive is better. Price says nothing about quality in either direction, and expensive and worthless is a common combination. The narrower claim is this: the price tag is the smallest variable in the decision, so it deserves the least weight. Rank on evidence, then published downside, then support, and look at price last, once the first three have eliminated most of the field. Where the paid options differ is compared in the best trading bots guide, and a full year of running costs is broken out in the cost to run an automated system.
No, and price on its own is uninformative. What correlates with quality is evidence: a long backtest you can reproduce, a stated risk model, a published drawdown, and somebody maintaining the code. A cheap system with all four is a good deal, and it is rare mainly because producing those things takes work a $39 price cannot fund.
Time, followed by evaluation fees. Every system you disprove by trading it consumes months you cannot get back, and on funded capital each closed account is roughly $100 for a 50k evaluation paid in advance. Two years of cheap systems can total well over a thousand dollars in cash and leave you with no usable track record. That is the bill nobody quotes at the point of sale.
Because the drawdown decides whether you are still running the system when the profit arrives. On funded capital it also decides whether the account survives at all, since prop firm rules close accounts on the low point and do not wait for a recovery. A seller who publishes profit and omits drawdown has answered the easy question and skipped the one that determines your outcome. More on that on our risk page.
Less than most people do, if you insist on reproducible evidence before funding anything. Rerunning a published backtest costs an evening, and simulation costs patience. Money should move only once the report has reproduced on your machine and you have decided you could sit through the worst drawdown it shows. Those two steps turn an open-ended search into a bounded one.
Everything we write comes back to one idea: verify, then decide. The live terminal exists for exactly that.
Unlock Algomyth Live