A model you are meant to argue with.
Signalvexa exists because most crypto tools give you a verdict and hide the reasoning. A number with no derivation is not analysis — it is a mood. So this platform inverts the usual arrangement: the derivation is the product, and the headline probability is just its summary.
Every report shows all seven factors, their fixed weights, their individual scores and the evidence behind each one. If you disagree with the conclusion, you can find the exact factor you disagree with and discount it yourself.
The seven factors and what they weigh.
Each factor resolves to a score between −1 and +1. Weights are fixed in advance and never tuned per asset, so the same evidence always produces the same number.
Trend structure
weight 0.20Relationship between the 20- and 50-day moving averages, price displacement from the 20-day mean, and a least-squares slope over the last thirty closes.
Momentum
weight 0.16MACD(12,26,9) histogram normalised against price, combined with realised 7-day and 24-hour returns.
Mean reversion
weight 0.14Wilder-smoothed RSI(14). Extremes are faded — deeply oversold reads score constructively, overbought reads score as a headwind.
Volume confirmation
weight 0.12Recent 7-day average volume against its 30-day baseline, signed by the prevailing price direction, plus turnover as a share of market cap.
Price structure
weight 0.12Position within the 90-day range and the asymmetry between distance-to-support and distance-to-resistance.
News & sentiment
weight 0.14Public newsroom headlines filtered to the asset and scored for tone, damped by how much coverage actually exists.
Market regime
weight 0.12Total crypto market-cap change and the asset's relative strength against that tape over 24 hours.
From composite to probability
The weighted composite is mapped onto a buy probability bounded between 8% and 92%. The bound is deliberate: a model that outputs 0% or 100% is claiming certainty it cannot possibly have. Confidence is computed from how tightly the factors agree with the composite, plus how many data sources were actually reachable. Risk comes from annualised realised volatility, positional stretch within the 90-day range, and any RSI extreme — so a strong-looking signal in a violent tape still reads as risky.
What this will not do for you.
It cannot see the future
Every factor is backward-looking. A probability is a statement about how conditions have historically resolved, not a forecast of what will happen next.
It has no position context
Signalvexa does not know your entry, your size, your time horizon or your tolerance for drawdown — all of which matter more than the direction of a signal.
News scoring is shallow
Headline tone is scored lexically, not semantically. Sarcasm, nuance and unpublished information are invisible to it.
Sources can fail
When a feed is unreachable, that factor is dropped from the weighting rather than guessed at, and the report says so explicitly on its face.