The skeptic asking this question is often right. Algo trading has attracted a genuine ecosystem of scams — guaranteed-return signal services, black-box systems with no explanation of how they work, and "AI-powered" strategies that can't describe what the AI does. Healthy suspicion is the correct starting position.

The answer is not that algorithmic trading is legitimate or illegitimate. It's that the space runs from fully defensible to outright fraud, and the markers that separate them are specific, checkable, and consistently ignored by people who want to believe the returns are real.

The Spectrum: Legitimate to Fraudulent

Understanding what legitimacy looks like requires mapping the whole range.

Tier 1 — Institutional systematic trading Quantitative hedge funds with disclosed strategies, audited track records, regulatory oversight, and multi-year out-of-sample performance histories. The Medallion Fund, Two Sigma, AQR. These exist. They are real. They are also inaccessible to most retail traders and tell you nothing about the platform you found via an Instagram ad.

Tier 2 — Retail systematic trading platforms Software tools, APIs, and strategy frameworks that let individual traders build and deploy rules-based strategies. No guaranteed returns. No black box. You write the logic (or choose from a documented catalog), you set the parameters, the platform executes. Risk is explicit. This is the legitimate retail tier.

Tier 3 — Signal services and copy trading Someone trades, you copy. Or someone sends signals and you act on them. These range from transparent (documented strategy, public performance, honest drawdown disclosure) to opaque (anonymous "trader", no verification, monthly fee for access). Red flags appear frequently in this tier.

Tier 4 — Fraud Guaranteed returns. Monthly yield products. "AI trading bots" that promise 5–15% per month with no drawdowns. Unregistered offshore funds. Performance charts that only go up. These are not edge cases in the industry — they are common and they cause real financial harm.

The red flags that mark Tiers 3 and 4:

  • Guaranteed returns — No legitimate trading operation guarantees returns. Markets are uncertain. Returns are probabilistic. Any platform that promises a specific monthly return is either lying about the guarantee or constructing a Ponzi structure where early investors are paid from new capital.
  • Black-box systems with no mechanism explanation — "Our proprietary algorithm" is not an explanation. Legitimate systematic strategies can describe what they do: the market condition they exploit, the entry and exit logic, the risk parameters. If a platform cannot or will not explain the mechanism, the mechanism does not exist or cannot survive scrutiny.
  • Backtests with no out-of-sample validation — A backtest that was never tested on data the strategy wasn't optimized against is a curve-fit, not a proof of concept. Ask: what is the out-of-sample performance? If the answer is "the backtest goes back 10 years," that's not an answer.
  • "AI" claims with no transparency — "Our AI identifies patterns" is meaningless without specifics. What model? Trained on what data? Predicting what output? With what validation methodology? "AI" as a marketing term without technical substance is a red flag, not a feature.
  • Unrealistically high Sharpe ratios in marketing materials — A Sharpe ratio above 2.0 sustained over multiple years is exceptional in institutional trading. Marketing materials claiming Sharpe ratios of 3, 5, or higher — especially without confidence intervals, drawdown data, or out-of-sample periods — are suspect. High Sharpe in a backtest is easy. High Sharpe in live trading is rare.

The 5 Questions to Ask Any Algo Platform

These questions have binary answers. Evasion is an answer.

1. What is the mechanism? Can you explain in plain language what market condition this strategy exploits and why that condition should produce positive returns? The explanation should be mechanically grounded — behavioral bias, structural constraint, information asymmetry — not circular ("it identifies profitable patterns"). If the platform cannot explain the mechanism, there is no mechanism.

2. What is the out-of-sample validation? Where was the strategy trained, and where was it tested? The testing period must be data the model never saw during optimization. Ideally: forward-tested in live markets, not a held-out historical window alone. If the "validation" is the same dataset as the optimization, there is no validation.

3. What does the worst drawdown look like? Any strategy with a real track record has losing periods. Ask for the maximum drawdown — the peak-to-trough decline — and the drawdown duration. If the answer is "we've never had a significant drawdown," that's a red flag, not a selling point. Every legitimate systematic strategy has had losing streaks.

4. Is the performance verifiable? Audited track records, third-party verification, live account statements — these exist for legitimate operations. Screenshots of P&L are not verification. If the claimed performance is not independently verifiable, treat it as unverified.

5. What is the incentive structure? Is the platform charging a flat software fee, a performance fee, or both? Performance fees align incentives, flat fees do not. But also ask: does the platform benefit from your trading volume? A platform that earns more when you trade more has an incentive to encourage overtrading, not optimal strategy execution. Understand how the platform makes money before you use it.

What Legitimate Algorithmic Trading Actually Looks Like

It is less exciting than the marketing suggests.

Legitimate systematic trading has drawdowns — sometimes prolonged ones. A strategy running at 15% annualized returns with a Sharpe of 1.2 will have months that are flat or down. It will have periods where the edge compresses because market conditions shift. The returns are not smooth. They are lumpy, occasionally frustrating, and the compound effect only becomes visible over years, not weeks.

Legitimate platforms disclose the mechanism. Not necessarily every parameter — there is reasonable IP protection in the specifics — but the category of inefficiency being exploited, the general entry and exit logic, and the risk management rules. You should be able to understand what the strategy does without being able to replicate it exactly.

Legitimate platforms are honest about edge decay. Market inefficiencies that become widely known and exploited compress and eventually disappear. A strategy that worked for five years may need to be retired or adapted. Platforms that claim permanently stable edges are either not tracking performance honestly or do not understand how markets evolve.

Legitimate platforms make the risk parameters explicit before you deploy capital. Maximum position size, stop-loss levels, maximum drawdown before shutdown — these are set by you, not determined by the platform's appetite. You own the risk decisions.

If a platform checks those boxes, it warrants evaluation. If it doesn't, the skeptic asking "is this legitimate?" was right to ask. See how systematic trading handles the edge-vs-gambling distinction for the underlying framework.

The Oyamori Approach

Transparency is not a feature — it's a precondition for trust. Every strategy in Oyamori's catalog includes a documented explanation of the market inefficiency it targets, the entry and exit conditions, and the risk parameters required for deployment. There are no black boxes.

Performance claims come with context. Drawdown periods are disclosed. Edge decay is monitored. When a strategy's behavior diverges from its historical distribution, the system flags it for review rather than continuing to execute on a dead assumption.

The risk structure is yours to set. Oyamori provides the execution infrastructure; you set the position sizing, the stop levels, and the portfolio-level limits. The platform's job is to execute your parameters consistently, not to manage risk on your behalf.

Algorithmic trading can be a legitimate, disciplined approach to markets. The five questions above will tell you whether the platform you're looking at is doing it the right way.

Next: The Retail Algo Trader Checklist →