MARKET RISK — you can lose the whole position
Equities can fall to a fraction of their value or to nothing, for reasons no model saw and no data panel contained. A forecast range does not put a floor under anything.
This is research.
It is not advice, and it never becomes advice.
GioModel publishes forecasts, the evidence behind them, and the failures that did not survive that evidence. What it does not publish is a recommendation. This page states the boundary precisely, because a boundary that is only implied is a boundary that gets crossed. Last updated: 7 August 2026.
Nine statements, in full
Each one is meant literally. None of them is softened elsewhere on the site.
- 01This is research and education
Everything published here exists to show how a forecast is built, validated and graded. It is written to be checked, argued with and reproduced. It is not written to tell you what to buy.
- 02It is not investment advice
Nothing on this site is investment advice, a recommendation to buy, sell or hold any security, an offer, or a solicitation. Giovanni Alexander is not a licensed investment adviser, broker or dealer, and GioModel is not a registered advisory service in any jurisdiction.
- 03It is not personalised to you
The research knows nothing about your income, your obligations, your tax position, your time horizon or your tolerance for loss. Even the Portfolio plan, which computes risk on positions you enter yourself, is arithmetic applied to numbers you typed — not an assessment of whether those positions suit you.
- 04There is no advisory or fiduciary relationship
Reading this site, subscribing to it, or emailing the owner does not make him your adviser and creates no fiduciary duty. A subscription buys earlier access to frozen numbers. It does not buy a duty of care over your money.
- 05GioModel does not trade or hold anything for you
No trades are executed here. No client assets, cash or securities are ever held, custodied or managed. There is no brokerage account, no discretionary mandate and no way for GioModel to move a position on your behalf, by design.
- 06Models publish which gates they failed
Every model states how many of the nine evidence gates it passed and exactly which ones it did not. A model that has not cleared its gates is not approved for directional deployment, and that is published rather than buried. No subscription moves a gate, and paying more does not upgrade the evidence.
- 07Forecast ranges are frozen simulations
A published range is the output of a model run on a point-in-time data panel that was frozen before the forecast was made. It is a simulation, not a prediction with a guarantee attached. Ranges are calibrated to cover roughly four outcomes in five, which means they can and do finish outside their bands — that is the intended behaviour of an honest interval, not a malfunction.
- 08Past results do not predict future results
The public track record exists so the work can be judged, not so it can be extrapolated. A model that graded well on resolved forecasts may fail on the next one. Market regimes change, data providers revise, and a validated method can still be wrong repeatedly in a row.
- 09The decision is yours, and only yours
You are solely responsible for what you do with anything published here, and for any profit or loss that follows. Consider speaking to a licensed adviser who is regulated where you live and who is able to take your circumstances into account. Never risk money you cannot afford to lose.
A gate result is a statement about proof, not a statement about profit. A model that passes eight of nine gates has demonstrated eight things and failed to demonstrate one, and the one it failed is named. That disclosure is the product. It is not a signal, a score to trade on, or a promise that the ninth gate would have passed given more data. If you find yourself reading a gate count as permission, the research has been misread. Read the method for what each gate tests and the track record for how past forecasts actually resolved.
Risks worth naming out loud
MODEL RISK — the method can be wrong
Validation reduces the chance of fooling yourself; it does not remove it. A specification that survived walk-forward testing can still be a poor description of the next regime.
DATA RISK — inputs can be wrong or late
Market data comes from a third-party provider and may be delayed, revised or in error. Provenance is recorded and quarantined rows are removed rather than repaired, but no provider is perfect.
STALENESS RISK — a frozen number ages
Every published range is tied to the moment its panel was frozen. It does not update itself as the market moves, and it should never be treated as a live view of anything.
If you take one thing from this page
- Builds forecasts on frozen, provenance-recorded data
- Publishes the validation rules and the gate results in full
- Grades every past forecast on its exact resolution session
- Keeps its failures online instead of deleting them
- Recommend a security, a size, an entry or an exit
- Assess whether anything is suitable for your circumstances
- Execute trades, hold assets, or manage money
- Guarantee accuracy, availability, or any outcome at all
- Check the method and the track record before believing a number
- Treat every range as a simulation that can finish outside its band
- Speak to a licensed adviser regulated where you live before acting
- Assume full responsibility for your own decisions