How Orion reads the market: conditions, not calls
Orion is a weather report for the market, not a tip service. It does not tell you what to do. It reads the conditions, the pressure systems moving under price, and writes them down one reading at a time. This is a short tour of how that works, so you can judge what the readings are and, just as importantly, what they are not.
Every reading is a condition
A reading is a description of one force in the market at a moment in time: leverage building, coins leaving exchanges, price stretched far above its long-term mean, the crowd leaning heavily one way. It is never a directive to act. Two readings can point in different directions at the same time, because the market is more than one thing at once, and that tension is information, not a contradiction to be smoothed over.
Six lenses
Each reading belongs to a lens, the kind of force it measures. Grouping them this way keeps the picture honest: you can see that a reading is one input among many, not a verdict.
- Structure. Where price sits against its long-term moving averages, momentum across timeframes, and classic cycle models.
- On-chain. What the blockchain itself shows: coins moving to and from exchanges, cost-basis models, network activity, holder profit and loss.
- Leverage. Funding, open interest, the long/short balance, and where liquidation liquidity is stacked.
- Sentiment. Fear and greed, breadth, and how the crowd feels versus how it is positioned.
- Macro. The dollar, rates, liquidity and money supply, the tide the whole risk market floats on.
- Cycle. The cumulative read that weighs everything above into where we sit in the broader cycle.
Why we wait 90 days
Subscribers see readings live. The public record fills in behind, on a 90-day embargo. That delay is deliberate: it keeps the free public archive as a clean historical record rather than a real-time feed, and it means the track record you can browse has already had time to be judged by the market. Nothing gets quietly deleted. Every reading is written down with its date and the price at the moment it fired.
A reading counts only once its 30-day window has closed, never on the day it fires. We score it against what price actually did, and we keep every reading in the record. A track record that only shows its wins is not a track record; it is marketing.
How we score it
For each reading with a direction, we ask a simple question 30 days later: did price move the way the reading leaned? Bullish readings are scored against an up-move, bearish against a down-move. Readings that describe a neutral condition are not scored as right or wrong, because they were never a directional claim. The result is a plain, backward-looking accuracy figure: the share of scored readings the market went on to confirm. It is a description of logged data, not a promise about the future.
The method keeps improving, in public
Orion is a platform we are actively building, and the method gets sharper over time. When we find a better way to read something, we ship the upgrade and write up what changed, the way we did when we added a structure and liquidity cross-check to the leverage read. We version those changes in the open: earlier readings stay in the record, scored as they landed, and the improved version starts a clean series alongside, so you can watch the upgrade prove itself. You can read that story in Reading leverage in context.
That is the whole idea: read the conditions, write them down, wait for the market to judge them, and keep the full record in public. See it live on the analytics page, browse every scored entry on the track record, or subscribe to receive readings as they fire.