HesaTrade scores markets by rules you can recompute yourself, sizes every position from your own risk limit, and measures its own accuracy out loud. No signals sold, no promises made.
No account needed to look around · runs in your browser · nothing to install
Every candidate is judged by seven separate methods, each from a documented school of thought. Each one may abstain when its precondition is missing, and each is weighted by its own measured track record — not by its reputation.
A divided council is a no. When the views disagree, that is not a hard decision — it is no decision at all.
Everything here is arithmetic you could redo on paper. There is no black box, because a box you cannot open is a box you cannot trust.
Stop, target and share count are fixed before you buy. Every position costs the same if it fails — that is the whole point of risking a fixed percentage.
Optimal size needs a known hit rate. Yours is an estimate, and overestimating it is the fastest route to zero. So a quarter of Kelly, capped again by your risk profile.
The rules run through history with spread and fees deducted, split into two halves. If the second half looks nothing like the first, the rules are worthless — and the app says so.
Hurst exponent, entropy and fat tails describe the character of a series. Near a Hurst of 0.5 the rules have no ground to stand on, and you are told before you act.
All stops trigger cleanly. Price gaps below your stop overnight. The whole market drops ten percent. A stop is an order, not a guarantee.
Headlines from eleven sources are archived and compared against their own previous week — which terms spiked, and which words keep appearing together.
Spot in euro per gram, the gold-silver ratio, distance from the 200-day average, and a premium calculator: with physical metal the dealer premium matters more than timing.
Positions open and close automatically against live prices, with a hard stop on new entries once the account is down fifteen percent. A risk limit does not argue.
Every probability is scored against the outcome. Above a Brier score of 0.25 it is worse than a coin toss, and you will read exactly that sentence.
Each stage has a condition the app can check. None of them can be skipped — that is the entire idea.
Six chapters and a test. Nine of eleven answers must be right.
Seven questions. Three answers lead to a clear no rather than a risk level.
Runs by itself in the background with every scan.
Brier score under 0.250 across at least twenty scored forecasts.
Both halves of the out-of-sample test positive, after costs.
Then, if ever: a quarter of the money you had in mind.
Twelve chapters on what is actually documented about markets — position sizing, fat tails, costs, the Kelly fraction, calibration, and the behavioural traps that cost more than any bad forecast. Written in German, about forty minutes, free and without a sign-up.
The base rate first: what the research on retail traders actually found, and why most of the damage comes from position size and costs rather than bad picks.
The mathematically optimal bet size — and why betting twice the correct amount does not double your growth but reduces it to nothing.
Interviewed traders contradict each other on almost everything. They agree on exactly three points, and none of them is about prediction.
The tool and the education are the product. Signals are never sold here — not for money, not for credits, not per download. That is a deliberate line, and the reason for it is in the terms.
Scanner, council, risk profile, paper trading, walk-forward test, the book, the briefing. No time limit, no card, no ads. If you never pay a cent, the tool still works completely.
Server-side watcher with push alerts, more markets and longer history, news and word analysis, sync across devices, data export. Priced as software, per month, cancellable at any time.
No affiliate links to brokers, no commission on where you trade, no paid signals, no "premium picks", no performance fee. Every one of those would put my interests against yours.