Define
the target.
We are not asking whether ASTS is “a good stock.” We are freezing a smaller question about raw closing prices, with a timestamp, a horizon, and one answer.
These readings update independently. They are not model inputs, deployment evidence, or a trading recommendation.
- Latest price
- $62.31 · fallback
- 20D momentum
- -13.4% Close-to-close · 20 completed sessions
- RSI 14
- 47.2 neutral · Wilder method
- Relative volume
- 1.0× Versus the prior 20 completed sessions
TECHNICALS THROUGH · TIMESTAMPED FALLBACK
Will ASTS close higher five trading days from now?
At the close of each trading day, the model will estimate whether the raw closing price five trading sessions later will be above today’s raw close. A tie belongs to class 0.
if raw close at t + 5 is greater than raw close at t
otherwise
Specific enough to test.
A fixed horizon turns vague “up or down” language into labels we can reproduce. Using the post-close decision time also gives us a clean information boundary.
Direction is not a trade.
The label ignores position size, transaction costs, risk, and the magnitude of a move. Model 1 stops at probability evidence and makes no trading-performance claim.
A target contract we can audit.
- One asset and one raw-close outcome
- A decision timestamp that blocks future data
- A fixed five-trading-day horizon
- A fold-local probability benchmark the model must beat
The primary score is Brier improvement against a smoothed fold-local training prior. A development model must improve probability error—not merely look more complicated.
Available now in Model 1.