Risk Disclosure & Non-Advice Notice

01This is software, not advice

Every output is generated by applying general, publicly documented technical criteria to a price series. It is not tailored to you. We do not know your income, your net worth, your other positions, your tax situation, your time horizon, your obligations, or your tolerance for loss — and we do not ask, because knowing would not change what the software does.

Nothing ORIN says should be read as “take this trade”. The product is built specifically so that it cannot say that: it grades, it explains, and it declines. Whether to act is a decision only you can make and only you carry.

If you want advice that accounts for your circumstances, you need a licensed adviser in your jurisdiction. That is a genuinely different service and we are not a substitute for it.

02Trading loses money

These are not abstractions. Specifically:

• Most active traders lose money. This is consistently reported by regulators requiring brokers to disclose client outcomes, and it is the base rate you are trading against. A tool that grades setups does not exempt you from it.
• A good decision can lose. Any single trade is a draw from a distribution. A correctly graded, correctly sized, correctly executed trade loses a meaningful fraction of the time, by design. If you judge the tool by whether the last trade won, you will misjudge it in both directions.
• Leverage multiplies both directions. On margin, futures, CFDs or perpetuals, a move against you that is small in percentage terms can close your position entirely regardless of where you placed a stop.
• Stops are not guarantees. A gap through your level, a halt, a fast market, or thin liquidity can fill you far worse than the price you specified — or not at all.
• Costs compound against you. Spread, commission, financing, slippage and tax are paid on every trade, win or lose. A strategy with a genuine edge before costs can be a losing strategy after them.
• Currency risk applies where the instrument is denominated differently from your account.

03What the analysis cannot know

ORIN reads a price series. That is a real thing to read, and it is much less than the market. The following are invisible to it unless you supply them:

• Scheduled and unscheduled news — earnings, guidance, central bank decisions, economic releases, regulatory action, litigation, geopolitics. A chart cannot see an earnings date, and a grade that ignores one can be confidently wrong.
• Corporate actions — splits, dividends, mergers, delistings, halts, and the adjustments different data vendors apply to them.
• Order book and liquidity conditions — depth, spread, participation, and whether the size you intend to trade can actually be traded at the levels shown.
• Your position and your context — existing exposure, correlation across your book, concentration, margin state, or the fact that you have already lost three times today.
• Anything after the last candle. Every read describes the market as at a moment. Markets move; a read goes stale, and the Service marks a read as suspended when the data behind it is too old to describe the present.

The Service surfaces its own uncertainty rather than hiding it — thin data, inferred scale, stale quotes, conflicting rules and unconfirmed manual checks all appear on the read. Take those seriously; they are the honest part.

04Automated analysis fails in specific ways

Part of the Service uses a large language model to describe, in prose, an analysis that has already been computed deterministically. The numbers come from our engine; the model expresses them. Every sentence is checked against the computed evidence, and any claim citing a figure the engine did not produce is suppressed and replaced.

That mechanism eliminates one specific failure — a fabricated price level — and nothing else. It does not make the underlying analysis correct. In particular:

• technical criteria that worked historically can stop working, and usually do so before it is obvious;
• a pattern can be correctly identified and still fail, which is the normal case rather than the exception;
• market data can be delayed, adjusted, gapped or simply wrong, and an analysis of bad data is bad however good the analysis;
• software has defects, and this software will have some you and we have not found yet.

Verify anything you intend to act on against your own broker’s data before you act on it.

05A chart image read is a guess about a picture

The reason that distinction is worth a section of its own is that the two outputs look identical. A price is a price on the screen whether it was read off an axis by a model or computed from a vendor’s candles, and the second kind can be verified while the first cannot. A model reading a chart image will produce a level that is precise, plausible, stated with the same fluency as a measured one, and simply not there.

Specifically, on a chart image read:

• the levels may not exist. A price axis in an image is a few dozen pixels of text. It can be misread by an order of magnitude, read off a secondary scale, or inferred where it was cropped;
• the instrument may not be the one you think. Tickers are misread, and a read of the wrong instrument is not obviously wrong — it is a coherent analysis of something else;
• the image is a frozen moment. It carries no live price and no timestamp we can trust. A screenshot from last Tuesday reads exactly like one from a minute ago;
• drawings are not distinguished from data. Lines, zones and labels someone drew on a chart are pixels like any other, and a level “found” may be one somebody drew;
• there is no calibration behind it. The conviction word attached to the read is the model’s own; it has no resolved sample behind it and is not the calibrated confidence described in section 6.

Output from this feature is excluded from every published accuracy figure, from the calibration record, and from your journal — not as a limitation of the beta, but because including it would make those figures describe something nobody measured. The figures in section 6 are about engine-computed reads only.

06Performance figures, and why most of ours are empty

Where the Service publishes a statistic — a hit rate, an expectancy, a calibration curve — it is computed from outcomes that actually resolved on ORIN, and it is published whichever way it lands. Nothing is published until at least 500 resolved observations sit behind it, which is why most cells on the property currently read collecting rather than showing a number. An empty cell is the honest state of a new product; a full one at launch would have to have been invented.

Simulated and hypothetical results. Any figure not derived from live, resolved outcomes — a backtest, an illustration, a worked example — is hypothetical. Hypothetical performance results have many inherent limitations. Unlike a record of actual trading, they do not represent actual trading, and because the trades were not executed they may under- or over-compensate for the impact of market factors such as liquidity. They are prepared with the benefit of hindsight, and there are frequently sharp differences between hypothetical results and the actual results subsequently achieved by any particular trading programme. No representation is made that any account will or is likely to achieve profits or losses similar to those shown.

Worked examples in marketing material, help articles and product demos are illustrations of format, not records of trades. Where an example is fixed rather than live, the Service says so on the example itself.

How every published figure is computed, and what counts as a resolved outcome, is set out on the methodology page.

07Your responsibility

You are solely responsible for:

• every trading decision you make and every order you place;
• determining whether trading a given instrument is legal and appropriate for you where you live;
• the amount you risk. The Service computes position size from figures you supply — account size, risk percentage, loss cap — and a wrong input produces a wrong size with no way for us to detect it;
• verifying prices, levels and calculations against your broker before acting;
• obtaining independent financial, legal and tax advice where you need it.

If trading is affecting your finances, your work, your sleep or your relationships, stop and get help. Problem gambling services treat compulsive trading, and the tooling on this site is not a substitute for that either. In the United States, the National Problem Gambling Helpline is available on 1-800-522-4700, 24 hours a day.

08Availability and jurisdiction

The Service is offered from the United States. It is not directed at any person in any jurisdiction where its availability or use would be contrary to local law, and it is your responsibility to determine whether your use of it is lawful where you are.

Nothing in the Service constitutes an offer or solicitation to any person in any jurisdiction in which such an offer or solicitation is unlawful. Access from a territory subject to comprehensive sanctions is prohibited under the Terms of Service.