One Read Instead of Fifty Tabs
The problem ODIN exists to solve is noise. On any given day the market throws off dozens of conflicting signals — one indicator says one thing, another says the opposite, and a wall of headlines says everything at once. Most people respond by opening more dashboards, which makes the noise louder, not clearer.
ODIN does the opposite. It watches sixteen independent areas of the market at once, weighs what each one is saying, and resolves all of it into a single composite read:
- BULL, BEAR, or NEUTRAL — the overall direction the data is leaning.
- A conviction score — a number that tells you how strongly the data agrees with itself.
That second number is the point. A direction on its own is just an opinion. A direction plus a measure of how much the underlying signals actually agree is something you can think with.
The Sixteen Pillars
ODIN reads the market through sixteen independent "pillars." Each pillar measures a different, well-established part of the market, so no single data source can dominate the read. They fall into a few broad families:
Each pillar is computed from real market data, independently. The skill is not in any one pillar — those are well understood. The skill is in combining them, which is the next section.
The Conviction Score: Measuring Agreement
Once every pillar has produced its own reading, ODIN measures how much they agree with each other. That agreement is what the conviction score represents.
Think of the sixteen pillars as sixteen independent analysts in a room. If three of them lean one way and the rest are split, that's a low-conviction situation — the score will be low, and the honest message is "the market doesn't agree with itself right now." If most of them are pointing the same direction at once — technical, macro, on-chain, and sentiment all aligned — that's a high-conviction situation, and the score will be high.
So a conviction score is not a promise about what will happen. It's a precise description of how unified the current evidence is. A low score is genuinely useful information: it tells you this is a moment of disagreement, not clarity.
Adaptive Weighting
Not every signal matters equally in every kind of market. Funding rates carry more meaning in a derivatives-driven rally; on-chain flows matter more in an accumulation phase. A tool that weighs all sixteen pillars the same in every market is using a blunt instrument.
ODIN's pillar weights are not hand-set and frozen. They are recalculated by a machine-learning step (logistic regression) that adjusts how much each pillar contributes as market conditions shift. In plain terms: when the market changes character, the way ODIN listens to its sixteen pillars changes with it, automatically.
Keeping the Model Honest
There's a classic trap in any model that learns from data: it can become very good at "explaining" the past while being useless about anything it hasn't seen. ODIN is built to avoid that trap.
It uses walk-forward validation — the weighting is derived on one window of data and then only applied to the next, unseen window, never the same data it learned from. This is a standard discipline in serious quantitative work, and it exists for one reason: to stop a model from fooling itself.
To be clear about what this is and isn't: this is a description of how the model is built and kept disciplined. It is not a performance claim, a prediction of future results, or a promise of accuracy. No model can promise the future. What this method does is keep ODIN's reasoning grounded in how signals have actually behaved, rather than in wishful curve-fitting.
Where the Data Comes From
ODIN runs on live market data, refreshed continuously:
- Price, volume, and derivatives data from major exchanges and venues.
- On-chain data from the Bitcoin network itself.
- Options data (including Deribit) for positioning and implied volatility.
- Macro and volatility references from broader financial markets.
- News sentiment read automatically from major crypto and financial outlets.
- Prediction-market data from Polymarket.
The signals update around the clock, which is why ODIN can produce a current read at any hour rather than a stale once-a-day snapshot.
What ODIN Is, and What It Is Not
This is the most important section on the page, so it's the plainest.
- A market-intelligence tool that synthesizes sixteen independent signals into one read.
- Identical for every user — a non-tailored, general-information product.
- A way to measure the market and see how much the evidence agrees.
- Not advice. ODIN never tells you to buy, sell, or hold anything. It does not know your finances, your goals, or your risk tolerance, and it does not try to.
- Not personalized. The read you see is the same read everyone sees. Nothing about it is tailored to you.
- Not a prediction or a guarantee. A conviction score describes how much the current data agrees — not what the price will do next.
- Not a broker or a manager. ODIN does not trade for you and does not manage money.
What you do with ODIN's read is entirely your decision. That's not a disclaimer bolted on at the bottom — it's the whole philosophy. We inform. You decide.
How to Read a Signal Well
A few honest pointers for getting value from the read:
- Treat a low conviction score as a real answer. "The market doesn't agree with itself" is useful to know, not a failure of the tool.
- The direction and the conviction matter together. A BULL read at low conviction and a BULL read at high conviction are very different situations.
- It's one input, not the only one. ODIN is built to be the clearest single read you have — it is not built to replace your own judgment, your own research, or professional advice.
See It in Action
That's the method, start to finish. No black box, no mystique — just sixteen signals, measured honestly and combined into one read you can actually use. The daily composite read is free.
Start Free → See Plans & Pricing →