Most scoring tools ask you to trust a number they will not explain. This page is the opposite: the entire KairosBias model, layer by layer, with the exact weights and thresholds the live site uses. If a score ever looks wrong, this is the reference to check it against, and the methodology page goes one level deeper.
The five layers
Each layer produces a sub-score from -100 to +100 per currency, and the composite is the weighted blend: 0.32 x EMA + 0.27 x COT + 0.18 x momentum + 0.13 x rates + 0.10 x seasonality. The weights are deliberate: trend gets the most because fighting it is the most expensive habit in forex; positioning is next because it is the strongest independent confirmation; momentum and rates refine rather than dominate; seasonality carries the least weight and its own consistency-based guardrail, because a historical calendar pattern is the easiest of the five to mistake for a stronger signal than it is. Pro users can change the weights if their process disagrees with ours.
| Layer | Weight | What it measures | Updates |
|---|---|---|---|
| EMA trend structure | 32% | Where price sits relative to its long moving averages, per currency, aggregated from its pairs | Every trading day |
| COT positioning | 27% | Large speculators' net futures position and its week-over-week shift, from CFTC data | Weekly (Fridays) |
| Momentum | 18% | Whether recent movement is accelerating or fading | Every trading day |
| Rate differentials | 13% | Each central bank's policy rate relative to the other majors | On central bank changes |
| Seasonality | 10% | Historical calendar-month tendency for the currency, weighted by how consistent that tendency has been across roughly 27 years | Monthly |
From scores to labels and pairs
All 8 currencies are free forever, with exact scores and sub-scores. What's metered is the pair board: six pairs (built from USD, NZD, CAD, CHF) are free, the other 22 of 28 are Pro. Nothing about the formula is gated: what you pay for is coverage, not the secret.
- ·Bias labels: strong bull at +60 and above, bull at +20, neutral between -20 and +20, bear at -20, strong bear at -60.
- ·Pair divergence: every one of the 28 pairs is scored as base currency minus quote currency. USD at +55 and JPY at -40 gives USD/JPY a divergence of +95.
- ·Confidence: not the size of the divergence, but how many of the five layers agree. All five pointing the same way is high; four agreeing with one dissenter is medium; a real split is low, which in practice means skip.
What the model does not do
It does not read news, predict data releases, or know about tomorrow's central bank meeting. It does not produce entries, targets or stops, and it never will, because it is an analytical tool, not a signal service. It tells you where the persistent macro pressure is, from primary sources: CFTC filings, central bank rates, official ECB exchange rate fixings. The trade is yours.
Common questions
Yes. The five layers, their exact weights (EMA 32%, COT 27%, momentum 18%, rates 13%, seasonality 10%), the bias thresholds and the layer-agreement rule behind each confidence label are all published, and this article states them. The free tier exposes exact scores and sub-scores for all 8 currencies so the model can be audited against the market.
Once per trading day, in practice. The price-derived layers (EMA structure, momentum) are built on the ECB's euro reference fixing, which is published once per working day, so that is the fastest anything here can genuinely move. COT positioning updates weekly when the CFTC publishes. Rate differentials update when central banks move. The scoring job itself runs every 4 hours, but running it more often than the data changes does not make the number newer, so we describe the cadence by the data rather than by the job.
See it live, free.
All 8 major currencies scored from EMA structure, COT positioning, momentum, rate differentials and seasonality, refreshed every trading day. No card, no expiry.
Open the free meterAnalytical tool, not financial advice