Risk disclosure

VibeTrading is an analysis tool. It can watch for conditions you define and tell you when they occur, and — if you set that up — it can trade them inside a contract you own and control. It is not a broker, not an adviser, and it chooses nothing for you: the strategy, the size and the limits are yours. Read this page before you act on anything the app shows you, and read the section on execution twice before you turn it on.

This is not financial advice

Nothing produced by VibeTrading — the levels, the W and M patterns, the head-and-shoulders and cup-and-handle shapes, the candle shapes, the confidence percentages, the measured-move targets, the alerts your rules fire, or anything the chat assistant writes — is investment advice, a recommendation, or a solicitation to buy or sell anything. It is arithmetic applied to public price history, presented for you to interpret. No one at VibeTrading knows your finances, your risk tolerance, or your goals, and the app does not take them into account.

If you want advice, speak to someone licensed to give it in your jurisdiction.

What the app actually does

It reads recent candles from a public market-data API, then computes things you could compute yourself with a spreadsheet and enough patience: which prices the market has revisited, and where two lows or two highs sit close enough together to form a double bottom or double top. Every threshold is a fixed rule expressed in ATR. There is no prediction model, no proprietary edge, and no crystal ball.

A confidence percentage measures how cleanly a shape matches its geometric definition — how close the two lows are, how deep the pattern is, how symmetric the legs are. It is not a probability that a trade will work. A 90% double bottom is a tidy-looking double bottom, nothing more.

Strategy alerts are notifications, not instructions

You can arm a rule — a pattern shape, or price meeting a support or resistance level — and the app will tell you when that condition is met. An alert means exactly one thing: the arithmetic described above matched on a candle that has closed. It is not a view on whether to trade, it carries no opinion about your position or your risk, and it is not a signal to act.

Alerts are deliberately late. A rule only fires on closed candles, and by default only once a pattern is confirmed and has survived a further candle — so by the time you hear from us, some of the move has already happened. That is the trade we chose: fewer alerts for setups that vanish, at the cost of a worse entry.

Do not use alerts as risk management. They are not a stop-loss and not a substitute for one. Delivery is best effort and is not guaranteed: our server may be down or restarting, market data may be cached or stale, an hourly rule may not notice for several minutes, and a rule may simply never fire. If you would be harmed by an alert arriving late or not at all, do not depend on it — put a real order on your exchange instead.

Rules are private to the wallet address that created them. Connecting a wallet and signing a message is how we establish that address, and for alerts that is all it does: no transaction, no token approval, and no ability for us to move anything you hold. Execution is a separate, explicit step — a contract you deploy and a permission you grant in their own transactions, which you can revoke at any time. Signing in has never authorised spending and still does not. If you lose access to that address, you lose access to the rules under it, and to any vault you deployed from it.

Chart patterns are not predictions

We have measured this and would rather tell you than let you assume otherwise: run the detector over a random walk with no structure in it at all, and it finds roughly as many patterns as it finds on real Bitcoin data. That is a property of chart patterns in general, not a defect in this implementation — random data genuinely contains W-shapes. It means a mark on your chart is evidence that a shape is present, not evidence that a move will follow.

Backtests are not forecasts

A backtest replays stored candles through the same code that fires your alerts, and models entries, exits, fees and slippage. It cannot model the order book you would really have traded into, funding, exchange outages, or your own behaviour on the day.

Costs are charged the way a decentralised exchange charges them: the pool's fee tier on every swap, so twice per trade, plus price impact on every fill and gas per swap. Gas is a cost in dollars, so the position size you enter decides what share of the trade it is; a small position pays a large share. Every report shows what the strategy earned before costs, what costs took, and what was left — and those three numbers, not the last one alone, are what tell you whether a losing result is a weak signal or an expensive round trip. The tiers, the impact and the gas are yours to set: we cannot know which pool you route through.

Results on the seen period are selected: the platform searches many settings there and keeps the best, which flatters that number by construction. The unseen period is measured once, with that single choice, and it is the number worth reading. A large drop from seen to unseen is ordinary, and we flag it — it means the settings fitted noise rather than anything that repeats. Every report says how many settings were tried. Fewer than thirty trades tells you very little either way, and we say so on the report rather than leaving you to notice.

Trading can lose you money

Trading cryptocurrency carries substantial risk, including the total loss of the money you put in. Crypto markets run continuously, move violently, and are lightly regulated compared with equities. Leverage multiplies losses as readily as gains. Past price behaviour does not predict future price behaviour. Only risk money you can afford to lose entirely.

What we never do

• We do not hold your funds. There is no deposit to us and no balance with us. If you use execution, your money sits in a contract you deployed and own, and the permission you give us cannot withdraw from it — there is no function in that contract by which we could, whatever we wanted or were asked to do.
• We never ask for exchange API keys, seed phrases or private keys. There is nowhere to enter them and no feature that would use them. If anything ever asks you for them in our name, it is not us.
• We cannot move a stop once it is set. Your stop, your target and your deadline are written into the contract when a position opens, and nothing changes them afterwards — not us, and not you.
• We do not choose what to trade. Every rule, every limit and every size is yours. We do not recommend strategies and we do not accept discretion to invent one.
• We do not sell signals, promise returns, or publish track records. Any claim of guaranteed profit attributed to VibeTrading is fraudulent.

If you switch execution on

Execution is off. It is off for everyone, on our side as well as yours, and turning it on takes three separate things: a contract you deploy yourself, a permission you grant on-chain, and a backtest that passes. Arming a rule for alerts does none of them. What follows describes what happens once you have done all three.

Your money stays in your contract. You deploy a vault, you fund it, you own it. We hold a permission that can do exactly two things inside it: open a position and close one. It cannot transfer, it cannot approve anyone else, it cannot change where a trade routes, and it expires by itself. You can revoke it in one transaction, without our cooperation and without telling us.

Your stop does not depend on us. It is stored in the contract and checked against a Chainlink price. Anyone at all can send the transaction that closes a stopped position, and the contract pays them a small fee from the vault for doing it — so if our servers are down, a stranger has a reason to close your position for you. That is deliberate, and it is why a stop here is not best-effort in the way an alert is.

A stop is not a fill price. A stop at $100 means the contract will allow an exit once the price is at or below $100. It does not mean you get $100. If the market gaps you get whatever the pool pays when the transaction lands, which can be much worse. Our backtests model the same thing, so a report and a real trade speak the same language — but neither is a promise.

Three different prices are involved, and they disagree. Your rule fires on candles from a centralised exchange. Your stop is authorised by a Chainlink feed. Your trade fills at whatever a Uniswap pool quotes at that moment. We record the gap on every fill, in basis points, and show it to you — because it is real, and you should see it rather than discover it.

Costs can exceed the edge. A swap pays the pool's fee going in and coming out, plus gas. Measured against a 1.5-ATR stop, that is roughly 0.03R a trade on a daily chart and more than 2R on a five-minute one. That is why the app refuses to arm a strategy whose own backtest does not clear its costs, and why intraday trading on a decentralised exchange is arithmetic rather than opinion.

A backtest is required, and it is still not a forecast. A rule cannot be armed for execution unless a backtest of that exact rule, on data the tuning never saw, clears a bar you set yourself. That check exists to stop the obvious mistakes. It cannot tell you the future, and a strategy that passed it can lose money immediately and continuously.

Start in shadow mode. Shadow records every trade the rule would have made and sends none of them. A week of it costs nothing and tells you what your strategy really pays in fees, gas and slippage against what the backtest assumed. Skipping it is the most expensive mistake available here.

The contract is new and unaudited. It carries a hard cap of a few hundred dollars per vault, written into the code so that we cannot raise it either, and it will not hold more until someone who is not us has audited it. Smart contracts lose money to bugs. This one could.

Software fails, and this software is young. Our executor can be down, slow or wrong. It can miss an entry entirely — which costs you an opportunity rather than capital, because entries depend on us and exits do not. Read that sentence again before you rely on either.

The data may be wrong or late

Market data comes from a third-party public API and is provided as-is. It may be delayed, incomplete, or unavailable. Prices shown may differ from those on your exchange. Do not rely on this app as your source of truth for a live position.

The software is young

VibeTrading is in active development and is offered free. It has bugs. Several detection rules on this site were corrected in the past week after users reported patterns being missed. Treat its output with the scepticism you would apply to any early tool.

Talk to us

If something the app shows looks wrong, tell us at dev@vibetrading.club. Reports like that are how the detector gets better.