Methodology & limitations

Last updated: 26 October 2026

This page documents what TradingQuant AI does today and where its limits are. It is not a performance report and it is not personalized investment advice.

1. Research universe

The current research universe covers approximately 300 US equities. Only a small subset becomes an executed change in the operator's model portfolio. Coverage is systematic and general; it is not selected for any subscriber's objectives, horizon or financial situation.

2. Decision process

• Five AI analysts evaluate fundamentals, momentum, sentiment, risk and macro context.
• A decision agent synthesizes the committee's consensus into BUY, SELL, SHORT or HOLD.
• The internal 0–100 model score is a research input, not a calibrated probability of profit, a success rate or a guarantee.
• The reasoning is part of the research record. AI-generated or AI-assisted explanations can be wrong, incomplete or inconsistent.

3. Risk and execution

• A deterministic risk layer controls position sizing, exposure limits and stops, and can reject a proposed trade.
• Orders are placed only in the operator's own Alpaca model account. The system does not connect to subscribers' brokers or execute in their accounts.
• The current model portfolio is simulated (paper trading). The plan is to move the operator's portfolio live before public launch.
• No client funds are managed, custodied or traded.

4. What the current offer includes

• Email updates when an entry, exit or other relevant change is reflected in the model portfolio.
• Ticker, action, reference price, model size, stop and committee reasoning, where available.
• Updates sent after the model-portfolio change, not before it.
• A searchable email record from the day the subscriber joins. There is no separate dashboard or historical online archive in the initial MVP.

HOLD and rejected decisions can be shown as research cases; they are not sold as default real-time trade alerts. Additional daily or weekly premium recaps and a historical archive will only be advertised when they are operational.

5. Recorded examples used on the landing

LEVI — 23 October 2026, simulated model portfolio.

• Recorded BUY decision with an internal score of 77/100.
• 33 shares entered at $20.33, approximately 6.9% of the model balance at entry.
• Stop level shown on the landing: $19.46. Position data was checked on 26 October 2026 and may have changed since.
• Reasoning included Piotroski 8/9, lower leverage than the sector and a P/E of 14.4 versus 25.3 for the sector at the time of analysis.

CLS — 26 October 2026, simulated model portfolio.

• Recorded HOLD with no new capital deployed and an internal score of 59/100.
• Strong quality metrics were outweighed by a confirmed downtrend, negative momentum and insider selling.
• The example shows risk discipline; it is not presented as a real-time HOLD alert.

These are static, recorded examples from the simulated model. They are not a live feed, client results or recommendations to act. Data and position details were checked on 26 October 2026.

6. Track record and performance disclosure

• Current track-record status: simulated model portfolio (paper trading).
• If the portfolio moves live, the label will be operator's live model portfolio, never client funds.
• Simulated results can differ from live execution because of slippage, commissions, liquidity and market impact.
• When results are published, they should include the period, benchmark, relevant costs and limitations, and enough context to avoid cherry-picking.
• No performance curve is currently shown on the landing. Past performance does not guarantee future results.

7. Data and AI limitations

• Market, fundamental, news and sentiment data can be delayed, incomplete, revised or wrong.
• AI outputs can misread data, produce flawed reasoning or disagree with later evidence.
• The service does not redistribute a real-time market-data feed.
• Always verify important information with independent sources before making your own decision.

8. Product boundaries

• No broker connection, auto-trading or execution in a subscriber's account.
• No personalized advice, portfolio management, custody or tax advice.
• No guaranteed returns or assurance that the model will be profitable.
• No app or private dashboard in the initial MVP.
• Capital at risk. Do not invest more than you can afford to lose.

9. Operator and revisions

TradingQuant AI is operated by Diasoluciones. Material changes to the research process, delivery scope or track-record status should be reflected on this page with a new revision date.

Questions about the methodology: quanttrading@diasoluciones.com.