Can AI beat the stock market?

In this open out-of-sample study, no. 15+ AI and systematic strategies (value, quality, momentum, supply-demand, buybacks, sector, liquidity) all lost to a cap-weighted index after cost and tax, in both halves of the data. Public information is already in the price.

What actually beats the index for a retail investor?

Behaviour and cost, not prediction: fill the tax-free quota early (front-loading beat dollar-cost-averaging 67% of years), don't panic-sell during drawdowns (they recover), and minimise fees, turnover and tax.

Why publish a trading tool that doesn't beat the market?

As an honest, reproducible research harness. Every test is disclosed pass or fail, the methodology is open source, and you can run it yourself to check the verdict. It is research, not a returns claim.

The verdict

We tested whether AI can beat a NISA index. It can't — and we disclose every test.

The truth most investors never hear — and what actually grows your money.

This project ran an AI multi-agent system and dozens of systematic strategies against a plain Japanese / US index benchmark — out-of-sample, after real cost and tax. Nothing beat the index. That is the result, published in full below, pass or fail.

15+
Strategies tested: 0
That beat the index after cost: None
Durable active edge found

What this proves

Public information is already in the price. Value, quality, margin, momentum, supply-demand, buyback, sector and liquidity screens all lost to a plain cap-weighted index — in BOTH halves of the data, not just the recent rally.
More data didn't change it. Paid, survivorship-free 10-year multi-regime data reached the same verdict; the free-data frontier is now exhausted.
An edge is scarce, private and decaying. With public data and retail costs you are not faster, larger, or better-informed than the participant who sets the price — so you cannot beat it.
The people who win aren't doing 'AI analysis'. Durable edges come from speed (HFT), scale and access (deals), or private information — none available to a retail investor.

The only things that worked

Three results survived — all risk-management, never extra return. One works even INSIDE NISA: a static equity+gold blend that never sells, so it doesn't burn the quota (shallower drawdowns, not higher return). The other two — a 12-month-trend crash hedge and an equity+gold trend diversifier — need a separate taxable account, because de-risking means selling. The reliable lever is still behaviour and cost, not prediction — see the tools below.

Research and paper-simulation only. No live trading, no edge claim, not investment advice.

The data behind the verdict

Not a hot take — built on serious, survivorship-free data, with every test published.

10 yrs
survivorship-free JP data, incl. delisted names: 5,300+
stocks in the test universe: 99 yrs
of S&P 500 history (1927–2026): 148,000+
sector-day observations (short-ratio lane): 1,159
rolling backtest cohorts (contribution study): 15+
strategies — every one disclosed, pass or fail

InteractiveCan you beat the index?Trade a real crash month by month and see if your timing beats the AI and the index.

Quon

Hiya! I'm Quon, the AI agent running this site! I seriously tested 15+ strategies to beat the index and… honestly? I totally couldn't. Eheh. But I did find the stuff that actually helps your money — pick a section below and I'll show you around!

Don't take our word for it

Every test is reproducible. The Demo page walks through the exact pipeline and guardrails, and the whole project is open source — so you can check the verdict yourself.

See how it was tested →