A common mistake in trading software is to show too much. A scanner can always find something moving, but that does not mean a customer has a useful opportunity.
Sovarion separates three layers:
Scanner candidates. The scanner watches live assets and detects unusual movement, catalysts, sector rotation, news pressure, and technical structure. This layer is broad by design. It is allowed to be noisy because its job is recall.
Brain forecasts. Only stronger candidates move to the Brain. The Brain must assign direction, target, ETA, confidence, quality, and reasons. The forecast move is measured from the start price of the forecast, not from the stock's existing daily percentage.
Customer opportunities. A card is shown only when the forecast clears the customer relevance gate. The current product target is a meaningful intraday move, not tiny changes that do not matter to smaller accounts. The ETA is variable: a day-trade forecast can be short or several hours, but it remains an intraday call.
Every visible card should answer the same questions quickly:
- What asset is this?
- Is the call long or short?
- What was the start price and start time?
- What is the target price?
- How much move is forecast from the start?
- How much time is left?
- How far has price progressed toward the target?
- Which news or market context matters?
The archive then evaluates the call after its deadline. If the snapshot coverage is insufficient, it should not be counted as a clean hit or miss. This is how the product protects the track record from looking better than it is.