What is a research digital twin?
Here, a digital twin is a continuously updated analytical representation of the eligible Egyptian equity universe. After each completed session, the engine uses the information it can evaluate to update its learned representation of relative relationships among stocks, then publishes an inspectable public model state.
Why compare stocks with one another?
Cross-sectional analysis does not ask only whether one stock looks strong or weak in isolation. It asks where each stock stands relative to the other eligible securities under the same market conditions. Relative Rank therefore describes comparative structure, not a probability of profit.
Relative Rank
If a stock is #3 out of 91, only two eligible stocks occupy higher positions in that model ordering for the selected horizon. The rank does not imply a specified probability of profit, price target or expected absolute return.
Model Direction
Positive, Neutral and Negative are separate market-relative forward classifications. A highly ranked stock can still have Neutral or Negative Model Direction because ranking and classification answer different research questions.
Why four horizons?
EGX /Alpha publishes separate model states for 1D, 3D, 5D and 10D. A stock’s relative placement can differ across those windows because each horizon describes a different forward-time question.
Rank Move
Rank Move records how relative placement changed since the previous completed public state at the same horizon. It is movement in the model representation, not price movement.
Market Structure and breadth
Direction breadth shows how many eligible stocks are classified Positive, Neutral or Negative, while rank movers show changing relative relationships inside the universe. They are different dimensions of the same model state.
How forecasts become evidence
A forecast is recorded first. Only after the corresponding horizon completes can its outcome be evaluated. The site therefore preserves dated public states and reports evidence after maturity rather than selecting examples with hindsight.
How to use the site for learning
- Choose the research horizon you want to understand.
- Read Relative Rank as placement inside the eligible universe.
- Read Model Direction as a separate classification.
- Compare Rank Move and market breadth to see how the model state is changing.
- Use the archive to inspect what the model published before later outcomes were known.
What the model does not know about a visitor
The outputs are general and non-personalised. The model does not know a visitor’s objectives, holdings, financial circumstances, risk tolerance or whether any security is suitable for that person.
A transparent learning record
Public Model Memory preserves the chronology of published states. The public record lets readers study how the representation changed and compare earlier observations with later market outcomes.
Relative Rank and Model Direction are general research outputs, not Buy, Hold or Sell recommendations.