Most trading ideas lose money.
Know before yours does.

Describe any strategy in plain English. Cosmu tests it on real prices, with your fees, against pure luck, and gives you a straight answer.

· Reading your idea· Scanning 6 years of news· Pulling real prices· Backtesting with your fees· Luck test · 10,000 random dates
News spikes found0
Average return, 2 weeks after each news spike
Beats random
—
After the news+2.1%
Random 2 weeks+0.7%
made money lost moneyEach band is one 2-week trade
202020212022202320242025
Made money
20 of 34
Beats random dates
97%
$1,000 each time
+$700
over 34 trades

Why most traders lose.

01

You buy the headline.

Cosmu shows what happened every time before, and whether the news or the market did the work.

Big AI launch today. Buy Nvidia?
After the last 8 launches+8.9%
Any random month, same years+7.9%
No edge
02

Fees eat the edge.

Every result is net of your broker's real commission, currency fee and spread. Before you trade, not after.

Before costs+89%
After costs−60%
Our own crypto momentum research · 6,036 trades · fees and slippage included
03

Luck looks like skill.

Every idea is checked against luck: random dates, overfitting tests, a period it never saw. Most fail. You only act on the few that don't.

Strategies tested3,661
Survived the first screen335
Earned paper trading11
Our own engine, Jun–Aug 2026 · 146,059 backtests · 2.3M simulated trades

From idea to trade, safely.

The pipeline our first engine ran, June to August 2026.

  1. 1

    Describe it

    In plain words. AI turns it into precise rules, with no peeking at the future.

  2. 2

    Test it

    On years of real prices, news, sentiment and on-chain data, with your broker's fees.

  3. 3

    Judge it

    Against luck: random dates, overfitting checks, an untouched test period. Most ideas die here.

  4. 4

    Paper-trade it

    Survivors trade live prices with virtual money, each in its own account.

  5. 5

    Go live, if you choose

    Real money only when you say so, behind five safety locks.

Any idea. Numbers or news.

Quantitative
Buy Bitcoin after a 10% daily crashBuy the Nasdaq when the Fed cuts ratesSell when a stock runs 20% above its 200-day average
Qualitative
Buy defense stocks when a war breaks outBuy the Nasdaq when layoff news spikesBuy Bitcoin when a bank collapses

Stop guessing. Test it first.