Accuracy/Counts
For every snapshot in the archive, we pair each preference's score with the actual price move 1, 7, and 30 trading days later. A score ≥ 50 expects an upward move; a score below 50 expects a flat or downward move. Two ways to look at the data: the Aggregate tab sums hits, misses, and neutrals across all tickers in your selected range; the By ticker tab shows one symbol's full archived history day by day. These are raw counts from your own archive. No claims, no predictions. Hit rate is computed only over the decided subset (hits + misses); neutrals are excluded because they aren't wrong, they're unknown.
Why we show counts, not just rates: A 100% hit rate over 3 decided rows is meaningless. The total tells you whether the rate is informative or noise.
Why we exclude neutrals: Neutrals are rows where we don't yet have the future price (e.g., today's 30-day return won't exist for another 30 trading days), or rows where the score was ineligible. They aren't "wrong" — they're "no data." Including them would muddy the rate.
Aggregate vs. By ticker: Aggregate is good for comparing preferences against each other across many tickers. By ticker is good for asking "did this score work for AAPL specifically?" — useful when one stock's behavior surprises you.
What this is not: Not a performance claim, not a prediction of future accuracy, not investment advice. Past hits don't guarantee future hits. We show this so you can interrogate your own data, not so we can sell you a number.