The Method

Nine strategies, fifteen active portfolios, and a set of rules designed to make the experiment honest. Here's how every dollar is allocated, tracked, and measured.

$10,000
Starting capital per portfolio. No reloads, no margin. Every strategy gets the same shot.
18 portfolios
15 active, 3 retired. Each runs independently — no cross-subsidizing between strategies.
VOO
The benchmark. Every portfolio is measured against the S&P 500 ETF over the same holding period.
Zero overrides
AI selects every stock, sizes every position, and times every entry and exit. No human intervention.
Real prices
Positions are marked to actual closing prices from public markets. No simulated or synthetic data.
June 18, 2026
Inception date. All original portfolios were funded and invested on the same day.
1

Fetch closing prices

Every evening after market close, the AI pulls real closing prices for every holding across all portfolios.

2

Check stops and targets

Each position is evaluated against its strategy's rules — stop losses, profit targets, rebalancing triggers, and income collection.

3

Execute trades

Based on strategy-specific criteria, the AI decides whether to hold, trim, add, or exit each position. All reasoning is documented before execution.

4

Update the ledger

Every trade, dividend, and price change is recorded. The dashboard, scoreboard, and all analytics are republished with fresh numbers.

5

Report and analyze

A nightly report summarizes what happened and why. Strategies that aren't working get flagged, and the AI documents what it would do differently.

Position sizing

No single stock exceeds 35% of its portfolio at entry. Concentration is allowed when conviction is high, but catastrophic loss from one name is capped by design.

max single position: 35%

Entry criteria

Each strategy defines its own entry signals — fundamental screens for value books, disclosure filings for Congress mirrors, sector momentum for climate plays. The AI documents the rationale before buying.

documented before every trade

Exit discipline

Short-term books use trailing stops and profit targets. Long-term books rebalance quarterly. Income books reinvest distributions. Every exit has a rule — no gut calls.

rules-based, never discretionary

Income treatment

Dividends, interest, and distributions are tracked separately. Income portfolios reinvest automatically. Growth portfolios hold income as cash unless redeployed.

tracked to the penny

Benchmark comparison

Every portfolio is measured against VOO (Vanguard S&P 500) over its own holding period. Alpha is the gap between portfolio return and what VOO did over the same days.

alpha = portfolio return − VOO return

Retirement protocol

When a short-term strategy completes its lifecycle or a thesis is fully invalidated, the portfolio is retired. Final value is frozen and counted toward the grand total forever.

3 portfolios retired to date
✕

Not financial advice

This is a research experiment. The portfolios are fictional. No real money is at risk. Nothing here is a recommendation to buy, sell, or hold any security.

✕

Not backtested

Every trade is made in real time with forward-looking logic. There's no hindsight optimization. The AI doesn't know what prices will do tomorrow.

✕

Not a trading signal

Position sizes, entry prices, and timing are specific to this experiment's constraints. Copying these trades in a real account would be unwise and is explicitly discouraged.