What a "signal" usually means, and what it should
Most paid signal services sell a complete trade: ticker, entry, target, stop, go. That packaging is the product, and it is also the problem — it hands you a decision rather than the evidence for one, which means you cannot tell a good call from a lucky one until the money has already moved.
Worth saying plainly what TradePilot's signals are, because they are deliberately not that shape and arriving expecting entry and stop levels will disappoint you.
What a recommendation actually contains
Four things, and no more: a direction (buy, sell or hold), a conviction score, a risk score, and the reasoning that produced them. Some also carry a backtest summary.
There is no entry price, no target and no stop loss. Those depend on your position size, your timeframe and what else you hold — none of which a sentiment engine knows. A tool that supplies them anyway is guessing at your circumstances and presenting the guess as precision.
The risk score is not what you would assume
This is the field worth understanding, because it reads backwards until you know how it is built. Risk is calculated from the amount of evidence, not the volatility of the asset: it starts at 5 and falls as the number of analysed events rises.
So a risk of 5 means "we barely have anything on this" and a risk of 1 means "this is built on a lot of material". It is a measure of how thin the ground is, not how dangerous the trade is. A stable blue chip nobody wrote about this week scores riskier than a volatile name in the news — which is correct for what the number describes, and confusing if you assume it means something else.
Conviction and risk are a pair
Neither is useful alone. High conviction with high risk means a strong reading drawn from very little — the configuration most likely to be a coincidence. High conviction with low risk means a strong reading drawn from a lot of material, which is the only combination worth acting on quickly.
Low conviction with low risk is the quietly valuable one: plenty of evidence, and it says nothing is happening. That is a real finding and most signal services have no way to express it, because "no trade" does not sell subscriptions.
Nothing fires automatically
Every recommendation arrives pending and stays there until you approve or reject it, or it expires. There is no auto-execution and no broker connection placing orders on your behalf.
That is deliberate rather than a missing feature. A signal you had to actively accept is one you had to read, and the approve-or-reject step is where the reasoning gets examined rather than skimmed. It also means the record of what you declined exists — which is usually more instructive than the record of what you took.
How to use free signals without being used by them
- Read the reasoning before the direction. If the explanation does not survive contact with what you already know about the ticker, the direction is irrelevant.
- Check the risk score to see how much material is behind it. A confident call on thin evidence deserves scepticism, not speed.
- Decide your own entry, size and stop. These are functions of your account, not of the signal.
- Reject freely. A rejection is a data point about your own judgement, and the pending queue is not a to-do list.
What free genuinely costs here
The honest answer: these signals are a reading of public discussion, so they are late to anything that was never discussed and loud about anything heavily promoted. They work best as a prompt to investigate a ticker you already understand, and worst as a way to find tickers you have never heard of.
Judge them the way we would want to be judged — predictions resolve to correct or incorrect once their timeframe closes, so there is a record to read rather than a claim to accept.
See current recommendations and their reasoning on the recommendations page, or create a free account.
Nothing here is financial advice. A recommendation is an analysis of public discussion, not a solicitation to trade.