markitel
Ten weighted evidence factors, regime-aware scoring, timestamped publication, and a public record that keeps the losses. This page is the due-diligence version — the same story we tell partners.
These are the exact factors and weights running in production — this table is generated from the same constants the engine uses, so it cannot silently drift from reality.
Higher-highs/higher-lows analysis — is there a trend to trade at all?
Direction of the EMA ribbon and where price sits relative to it.
Overbought/oversold readings plus divergence against price.
Crossover state and histogram direction.
Trend-strength confirmation — filters chop from conviction.
Whether participation actually confirms the move.
Volatility context and percent-B positioning.
Momentum oscillator for timing within the setup.
Proximity to key levels that could cap or fuel the move.
Candlestick confirmation — deliberately the smallest voice.
Weights shown are the production configuration. Factor weighting is an evidence model, not a performance promise.
Each factor contributes weighted evidence toward a single verdict. A signal only publishes when the combined evidence clears the configured threshold — most candidate setups never make it out.
Trending, mean-reverting, gappy — the engine scores the same factors differently depending on the kind of tape it is looking at, instead of pretending one playbook fits all markets.
Every published signal carries its factor breakdown, confidence, and a plain-language rationale. You can inspect the chain — nothing is a black box you are asked to trust.
Entry, stop, and targets are published at issue time and are not edited after the fact. What you see in the feed is what was called, when it was called.
The public feed keeps the full history — wins and losses. Nothing is quietly removed when it goes wrong.
Practice-mode results are always marked as simulated and are not indicative of future performance. We never dress simulated results up as live execution.
Not because ours would look bad — because the number is too easy to game and too easy to misread. A win rate depends entirely on how you count: which trades enter the denominator, how expired setups are treated, whether results come from paper or live execution, and how position management is accounted. Two honest teams can compute wildly different figures from the same history.
Our position: performance statistics belong behind independent verification, or nowhere. Until then we publish the thing a statistic can’t fake — the complete, timestamped record — and let you compute whatever measure you trust.
Real-time feeds. One verdict. Zero guesswork.
Start on the Free $0 plan — then connect supported venues as your workflow moves from signal review to execution.