✦ AI portfolio
A virtual model portfolio run by our analysis bot itself: up to 20 stocks, reviewed every week after the weekly close – with reasons for every purchase and sale.
How the AI portfolio works – simply explained
1. The starting capital
The AI manages a virtual portfolio of 100,000 – no real money. At most 20 stocks, each starting with about 5 % of the portfolio (around 5,000 at the start).
2. Every Saturday: the decision
After the weekly close the AI receives the weekly trends, the market lights, the 60 strongest stocks in an uptrend and its own results so far. It decides what to buy and sell – always with a reason.
3. The purchase
Only stocks in an uptrend are bought. For every purchase the AI states a holding horizon (short, medium or long term) and a target with a date, e.g. “target €240 by December”. Execution is at the opening price of the next trading day – a price nobody knew when the decision was made.
4. Holding: let profits run
A target that has been reached is not a reason to sell. It counts as a hit in the record, and the AI sets a new, higher target. Every position is protected by its stop: it sits at the trend line and moves up as prices rise – the further the stock climbs, the more profit is locked in.
5. Selling – three ways
- Automatic stop: If a week closes below the stop, the position is sold at the open of the next trading day. This rule always applies – the AI cannot override it.
- 10 % cap: If a stock grows beyond 10 % of the portfolio, the part above that is sold. The rest keeps running. That way no single winner dominates the whole portfolio.
- AI decision: It may sell earlier, but only with a reason from the data – e.g. if the home market's light turns red or a clearly stronger candidate needs the slot.
An example
The AI buys Siemens at €200 for €5,000 (25 shares), target €240 by December, stop €180. In November Siemens is at €240: target reached – the AI holds and sets a new target. The trend line has meanwhile risen to €215, so the stop is above the purchase price: a profit is locked in. If Siemens climbs further to €300, the stop rises with it. Only when a week closes below the stop is the position sold.
6. Learning and transparency
The AI learns from its results – through an open memory rather than through new training: every week, it reviews its completed trades, its success rate and the lessons it has learnt, and adjusts its approach accordingly. This ensures that it is clear what it has learnt.
Every strategy is assessed: with every purchase, the AI specifies which strategies it is based on. After 5 completed trades, a review takes place – if these trades perform worse than the others, the strategy is automatically removed.
Everything remains visible: every position, every trade – whether it made a profit or a loss – how often targets were met, and a comparison with the S&P 500 and the DAX. A fee of 0.1 per cent is deducted for each buy and sell transaction.
The official portfolio value is based on the weekly close. The positions table also shows the current price, delayed by about 15 minutes.
Performance since start
🔒 Positions, buys and sells – exclusively for Premium members
WeekTraders’ Premium service includes details of the shares held by the AI, along with their purchase price, target price and stop-loss level; a rationale for every buy and sell transaction; the weekly commentary; and the AI’s verified insights.
Until the launch, you can follow the AI portfolio free of charge in the Weekly Market Brief – featuring all the week’s buys and sells.
Sign up freeVirtual model portfolio run by an AI – no real money, no investment advice and no solicitation to buy or sell. Execution at the opening price of the week after the decision, 0.1 % costs per trade, prices in trading currency without currency effects, dividends and taxes. Past results are not a reliable indicator of future performance.