When our AI publishes a score of 75 next to a stock, it's fair to ask: 75 out of what, and how? A number is only as useful as the reasoning behind it, so here's exactly how the AlphaScore is built — no black box.
Every day, for each stock in the universe, the AI produces four separate reads, each on a 1-to-10 scale:
- Technical — the trend and price setup. Is the stock riding above its longer-term average, or falling through support? Is momentum with it or against it?
- Sentiment — the mood around the name. What are the headlines and the market's tone saying, filtered from noise into a single read.
- Macro — how the stock fits the broader regime. A name can be strong on its own but swimming against a risk-off tide; this score captures that fit (or friction) with the wider market.
- Catalyst — the near-term trigger. Earnings, a product event, a sector rotation — whatever might actually move the stock in the sessions just ahead.
The AlphaScore is simply the average of those four, scaled to 100. A stock that reads 7, 8, 6, and 7 lands at (7+8+6+7) ÷ 4 × 10 = 70. Deliberately, no single read dominates — a great chart with no catalyst and poor macro fit won't run away with the number, and neither will a hot headline on a technically broken stock. The four have to broadly agree for a score to be high.
Two things the AlphaScore is not. First, it isn't a "buy signal" — it's a read on the setup, one input among several, and it says nothing about how much (if any) of the stock the AI actually holds. Position size is a separate calculation driven by conviction and volatility, which is why a name can carry a big weight with a middling score, or a small weight with a high one. Second, it isn't a prediction of the price. It's a structured snapshot of how four dimensions line up today.
Why we show the math at all. Most scoring models ask you to trust a number you can't inspect. We'd rather you see the recipe: four honest reads, equally weighted, averaged, published. You can watch each name's score against what actually happens over time and judge for yourself whether the reasoning holds. That's the whole point of scoring in public.
This is educational and informational content about how the Alphixir engine works — not investment advice.