7vest
Systematic investing research

Invest smarter.Automatically.

7vest is an algorithmic investing platform being built on an unfashionable premise: most trading strategies don't work. The ones that appear to usually haven't been tested hard enough.

So we built the testing first.

Strategies evaluated · verdicts to date
Momentum crossover · infrastructure REITs
Rejected
Looked profitable in-sample. The edge vanished on data the model had never seen.
All-weather macro defense rotation
Rejected
Did not hold across independent time periods.
Precious-metals ratio trend
Did not clear
Doubled the headline return of simply holding gold — at identical risk-adjusted efficiency, and only at one narrow parameter setting. Not an edge.
Calendar & seasonality patterns
Rejected
Statistically suggestive, economically meaningless. No mechanism, no promotion.
Mean reversion · large-cap biotech
In observation
Cleared out-of-sample validation, a full-cycle re-test, and a survivability study for overnight gap risk. It is now being watched forward, capped, with no money at stake — because clearing the tests is not the same as being right.

Four rejections for every candidate still standing. That ratio is the product.

The premise

A backtest is a hypothesis, not a result.

Any strategy can be made to look brilliant against the history it was designed on. That is not a discovery; it is a description of the past written in advance. The interesting question is what survives contact with data the model has never seen.

Most things don't. 7vest is built to find that out early, cheaply, and before capital is committed — rather than late, expensively, and in a live account.

The gates

How a strategy earns its place.

Every candidate moves through the same sequence. Each gate can end it. Nothing skips ahead, and passing is never assumed to be the same as working.

  1. A mechanism, stated first
    Why should this relationship exist? A pattern with no economic reason to exist is a coincidence waiting to be discovered.
  2. Out-of-sample validation
    The strategy is tested repeatedly on stretches of history it was never fitted to. An edge that only appears in-sample is not an edge.
  3. Full-cycle re-testing
    Does it survive a different rate environment, a different regime, a different decade? Short histories flatter strategies.
  4. Survivability under stress
    What happens when a position gaps overnight and no signal can react? Position limits are derived from that answer, not from optimism.
  5. Forward observation, unfunded
    The strategy runs in the open with nothing at stake, and is judged on what it does next — not on what it would have done.
  6. Human review
    Clearing every gate makes a strategy eligible for a decision. It does not make the decision.

Where things stand

7vest is in research.

We would rather say this plainly than imply otherwise. The platform is under active development. Strategies are being tested and, mostly, discarded. Nothing is live.

Stage
Research & validation
Live trading
Not enabled
Outside capital
Not accepted
Published returns
None — and none until a strategy has earned them