Can an AI model beat the Nasdaq-100?

RLTW is a public experiment: we build a Nasdaq-100 portfolio with machine learning, momentum, fundamental quality and risk rules, then compare it over time with QQQ and an active momentum benchmark.

No promises. Public rules, stated costs, measurable results.

Understand the experiment Theoretical results Go to portfolio

Start with the theoretical test

Walk-forward backtest: €10,000 from 2 January 2025 becomes...

Same model

QQQ

~€13,890

+39%

Passive benchmark

Buy and hold the Nasdaq-100 over the same period.

~€41,720

+317%

Max DD −40%

Top 10. Friday close → Monday open. Official rules, 0.19% fees.

EMC

~€51,120

+411%

Max DD −37%

Tighter MC: Top 8, 60% Stage 2 + 40% last 4 weeks. Same calendar as MC.

RLTW v3

~€33,300

+233%

Max DD −23%

Top 8 ML. Trade when the Wizard / check-up says so.

Simulation from €10,000 on 2 January 2025, same 0.19% costs. The live book is a separate account: it starts later, from 10,000 USD, and can be up or down.See live ->

Step 2 · Who is performing now

Live track record, separate from the backtest

Each live book starts at $10,000. RLTW, MC and QQQ from 25 August 2026; EMC from 31 August 2026. A weak live period is part of the test: decisions are recorded before outcomes.

Live details ->

RLTW v3 live

—

Experimental AI model, 0.19%/side costs included.

EMC live

—

Started 31 August 2026 from $10,000

Our momentum book, Top 8.

Corrected MomentumCommand live

—

Active momentum benchmark, rerun with the same costs.

QQQ live