Intellectia.AI Review: A 200% Backtested Return Is Not a 200% Return

A backtest applies a set of rules to historical price data and reports what would have happened. It costs nothing, takes minutes, and is the easiest number in finance to make look extraordinary. Here is why backtested results systematically overstate what you would actually get. Overfitting. If you test enough rule combinations against the same history, some will fit it beautifully by chance. With modern computing you can test millions. The rule that best explains the past is frequently the one that has learned the past's noise rather than any durable pattern — and noise does not repeat. This is the central problem and it is not a minor one. Look-ahead bias. Using information in the simulation that would not have been available at the time — a restated earnings figure, an index constituent list as it stands now, a delayed data release treated as instant. Easy to introduce accidentally and it produces spectacular results. Survivorship bias. Testing on today's listed companies excludes everything that went bankrupt, delisted or was acquired. The historical universe contained failures; a modern database frequently does not. Transaction costs, slippage and market impact. Backtests commonly assume you trade at the closing price with no spread, no commission and no effect on the market. Real trading has all three, and a high-turnover strategy — which "updated every Monday" implies — is exactly where these costs bite hardest. Selection of the reporting window. A strategy tested from a market bottom looks very different from one tested from a peak. Now put the number in context. An over-200% annualized return, sustained, would be one of the greatest performance records in financial history. For comparison, the most celebrated quantitative fund in existence is generally reported around 66% gross annually before fees, and it closed to outside money decades ago. The base rate for a retail subscription product achieving triple that is very low, and that judgment does not require any view about this specific company's competence or honesty — it is what the distribution of outcomes in this field looks like. What to ask for instead: Live, out-of-sample, verified track record. Real money, real fills, over a period after the strategy was fixed — ideally third-party verified. This is the only performance evidence that means anything, and any provider with one will lead with it. The full rule set and the exact test period, including whether costs and slippage were modelled. Drawdown, not just return. A strategy returning 200% with a 70% drawdown is one almost nobody could actually hold through, and returns you abandon are returns you do not get.

The platform offers automated execution, one-click portfolio copying and 24/7 autonomous trading. These features change your risk in ways worth spelling out. Copy trading. Mirroring another account's positions automatically. What it does not give you: their capital base, their risk tolerance, their time horizon, their tax position, or their reason for a trade. A position that is 2% of a large account may be 40% of yours. Many platforms scale proportionally — confirm how sizing works before enabling anything. Whose trades are you copying? "Copy Trade Whales" typically means following large positions inferred from public disclosures. For US institutional managers, 13F filings are quarterly and filed up to 45 days after quarter end — meaning a position you copy may have been established four and a half months earlier and may already have been sold. They also exclude short positions, derivatives and non-US holdings, so you see a partial picture. This is a genuinely important limitation and it is rarely disclosed prominently. Autonomous trading. Software placing orders without you approving each one. Confirm the failure modes before enabling it. What happens on a data outage, a bad tick, a halted stock, an API failure or a gap open? Ask whether there are hard position limits, a daily loss limit and a kill switch you control. Understand your broker connection. What permissions does the platform hold — read-only, trade, or withdrawal? Never grant withdrawal permissions to a third-party tool. Automation removes the pause. The moment where you look at an order and think "actually, no" is a genuine risk control, and automation deletes it. A regulatory question worth asking directly. In the US, providing personalised investment advice for compensation generally requires registration as an investment adviser, with the fiduciary duties that brings. Publishers of impersonal, general market commentary may fall under a statutory exclusion. Ask which the company considers itself, and whether it or any affiliate is registered. A registered adviser owes you a duty of care and loyalty; a publisher does not. That is a material difference and you are entitled to a clear answer. And check the site's own disclaimer, which for products in this category typically states that nothing is investment advice and that you are solely responsible for your decisions. That disclaimer and a 200% headline are difficult to hold in the same hand, and where they conflict, the disclaimer is the operative one.

It would be unbalanced to write only about the performance claim. Setting that aside, parts of this are legitimately useful. Screening and data aggregation. Filtering thousands of securities on defined criteria is a real, unglamorous, valuable task, and software does it far better than a person. If the screener is good, that is worth paying for. Summarising filings and news. Reading and condensing earnings reports, filings and coverage is genuinely time-consuming, and language models are reasonably good at it. Verify anything material against the primary document — models still misstate numbers, and a misread earnings figure is expensive. A question-answering agent for finance concepts is a decent educational tool, with the same verification caveat. Idea generation. A screener output is a research starting list, not a buy list. Used that way, it saves time. The free tier is the right way to evaluate all of this. "$0 forever, no credit card" means you can assess whether the screener and summaries are good before paying anything — and the $1 trial similarly. Use them. On the trial specifically, check what it converts to, on what date, and at what price before entering card details, and set a reminder. On the ratings. Product Hunt 5.0, G2 5.0, App Store 4.8, Google Play 4.7 are real platforms. App-store ratings measure whether an app is pleasant to use, not whether its recommendations make money — and the two are unrelated. Nobody rates a trading app after a losing year. The things that actually determine investment outcomes, and none is on this site: Costs. Fees compound against you exactly as returns compound for you. A subscription at $799.99 a year is a fixed drag that a small portfolio may never overcome. Diversification and position sizing. Time in the market, and not selling at the bottom. Tax treatment. High-turnover strategies generate short-term capital gains, taxed as ordinary income in the US — which can quietly consume a large share of any edge. For most people, the boring answer remains the best-evidenced one: broad low-cost index funds, held, with regular contributions. That is not an exciting recommendation and it is the one supported by the weight of evidence.

What does Intellectia.AI cost? There is a free tier described as $0 forever with no credit card, a $1 trial, and paid plans with figures of $199.95, $300, $499.99 and $799.99 appearing on the site. Confirm current pricing and what the trial converts to before entering card details. Is the 200% annualized return real? The site states it comes from backtesting. A backtest is a simulation over historical data, not money that was made, and backtests systematically overstate results through overfitting, survivorship and look-ahead bias and the exclusion of trading costs. What should I ask for instead? A live, out-of-sample, verified track record with real fills over a period after the strategy was fixed — plus maximum drawdown, not just return. Is copy trading safe? It carries the copied trader's risk without their capital base, horizon or tax position. If it tracks institutional filings, note that US 13F filings are quarterly and filed up to 45 days after quarter end, exclude shorts and derivatives, and may describe positions already sold. Should I enable autonomous trading? Only after confirming hard position limits, a daily loss limit and a kill switch you control, and only with trade — never withdrawal — permissions on your brokerage account. Is the company a registered investment adviser? Ask directly. Personalised investment advice for compensation generally requires registration in the US; impersonal publishing may fall under an exclusion. A registered adviser owes you fiduciary duties; a publisher does not. Do the app-store ratings mean the picks work? No. They measure whether the app is pleasant to use, which is unrelated to whether its recommendations make money. What about taxes? High-turnover strategies generate short-term capital gains, taxed as ordinary income in the US, which can consume a large share of any edge. Which markets does it cover? US markets currently, with stated plans to expand. What is the better-evidenced approach? Broad, low-cost, diversified index funds, held long term with regular contributions. Unexciting, and supported by the weight of evidence.