Performance
The reporting infrastructure is built. Every trade this system takes will be published on this page, winners and losers alike. It is empty because we have not started.
| Date | Player | Market | Side | Stake | Entry | Exit | P&L | Result |
|---|
How this is reported
- Every trade published, winners and losers, with no selection.
- Stake, entry and exit shown for each position.
- Updated automatically from the system's own records, not hand-entered.
- Paper-traded results are clearly marked as simulated until trading is live.
Almost everyone is playing the wrong game
The crowded question
Nearly every model built for sports markets asks who wins. It is the most researched question in the field, contested by well-resourced firms, and the remaining edge is thin and expensive to hold.
Winning that argument also requires waiting for the result. Capital sits exposed to everything that can happen in ninety minutes of play.
The question we ask
When a recognised starter is ruled out, the market reprices. That move is a crowd reacting to an announcement — its size is a property of attention and perception, not of the game.
FREDO forecasts the direction and magnitude of that repricing, enters ahead of the crowd, and closes once the move is in. The result of the match is irrelevant to the position.
We are not trying to be right about the game. We are trying to be early to the reaction.
Four stages, end to end
Each stage is automated. Nothing is discretionary once the system is running.
Watch
Continuous monitoring of public disclosures and reporting, drawn from a vetted source list.
Value
The proprietary score estimates how far this piece of news moves this particular market.
Size
Position size is set by that score, then tested against live spreads and fees before anything is sent.
Exit
The position is closed at a target set by the same score, or cut at a fixed loss.
The Player Value rating
One number, one question
Every offensive player carries a rating from 1 to 100, rebuilt automatically as the season runs. It answers a single question: if this man is ruled out, how far does this market move?
Not how good he is. Not whether his team wins. How loudly the market reacts to his absence.
What it controls
That one number drives both sides of the trade. It sets the size of the position, and it sets the price at which the position is closed.
Because a single rating governs both, the system's behaviour is consistent and auditable: any position can be traced back to the number that produced it.
Built to refuse
An automated system that trades everything is a liability. The value is in what it declines.
Hard capital limits
Per-position, per-day and total exposure ceilings are enforced in code. The system physically cannot exceed them — there is no override, and no discretionary path around them.
Cost gate
Every candidate is priced against live spreads and fees before it is sent. If the expected move does not clear the cost of trading, the trade is declined.
Data integrity gate
If any critical feed goes stale, all trading halts automatically. The system stops rather than acting on old information.
Evidence threshold
Players without a sufficient record are never traded, however attractive the headline. An unmeasured player is not an opportunity.
What already exists
The data pipeline, the scoring layer and the capital controls are built and tested. Execution is the remaining work.
Players tracked and refreshed daily, with no manual data entry
Seasons of play-by-play, participation and injury history
Games with closing market prices, used to validate the ratings
Fully automated scheduled refresh that checks its own work
How we know if it works
In sequence. Each stage has to hold before the next one is funded.
Measure the ratings against history
Player Value is tested against real historical market movement — did markets actually move as far as the rating said they would, on games already priced and settled.
Paper-trade a full season
A complete season is traded on paper under the same capital limits, cost gate and data gates that govern live trading. Simulated results are marked as simulated.
Go live at small size
Live execution begins at deliberately small size, with the same ceilings enforced in code.
Scale only on measured results
Size increases on evidence from live trading, published on this page, and on nothing else.