Learn / Feature intuition
Answer page / feature intuition
Topic cluster / Signal and portfolio designA binary signal outputs a discrete action or state like trade or do not trade. A continuous signal outputs a score, rank, probability, or target exposure. That difference is about signal representation, not about whether the instrument itself is a binary option or event market.
What to remember
A binary signal says something like on or off, long or flat, enter or do not enter. A continuous signal says how much, how strong, how attractive, or how risky, usually with a score or target size.
That distinction lives at the signal layer. It is different from the instrument layer. You can run a binary signal on a continuous market like perps, and you can run a continuous score on an event-style market if the research object supports it.
Binary signals are often useful when the edge is sparse, threshold-based, or easier to express as a decision rule than as a smooth score.
Continuous signals are often better when the edge changes by degree rather than by pure state. They give you more flexibility for ranking, sizing, blending, and portfolio optimization.
The common mistake is treating binary signal as if it meant binary option or event market. It does not. One idea describes the output of the model. The other describes the payoff or contract structure of the instrument being traded.
Most of Alphora's launch surfaces lean continuous: scores, ranks, carry layers, and portfolio context for crypto perps. But even there, many practical decisions are binary at the execution layer, like whether a regime gate turns a sleeve on or off.