Quantitative signal feed
The scores and directional biases of our three macro models, delivered every Sunday, so you can arbitrate allocations without emotional bias.
Switch between M1 Surprises, M3 Scoring and Fair Value to read the score matrix, the conviction thresholds and the factor breakdown.
Inflation · PCE · Emploi · Taux · Croissance · Contrarian
| Zone | Inflation | PCE | Emploi | Taux | Croissance | Contrarian |
|---|---|---|---|---|---|---|
| CAD | -0.230 | — | 2.690 | 0.000 | 0.470 | -0.057 |
| GBP | -0.260 | — | 0.160 | 0.000 | 1.500 | -1.308 |
| CHF | -0.800 | — | — | -0.570 | 0.290 | -1.676 |
| AUD | -0.830 | — | 0.620 | 0.060 | -0.850 | 0.344 |
| USD | -0.600 | -0.260 | 0.320 | 0.000 | -0.040 | 0.000 |
| JPY | -0.280 | — | — | 0.000 | -0.160 | -0.350 |
| EUR | -0.460 | — | 0.000 | 0.000 | 0.770 | 0.036 |
| NZD | 0.200 | — | 0.200 | -0.180 | -0.110 | -1.356 |
Tendance -16 pts
| Zone | Score | Rank | Trend | Δ 1M | Δ 3M |
|---|---|---|---|---|---|
| CAD | 2.920 | #1 | -0.521 | -1.115 | |
| GBP | 1.400 | #2 | 0.709 | 1.235 | |
| EUR | 0.310 | #3 | 0.403 | 0.694 | |
| NZD | 0.120 | #4 | -0.057 | 0.384 | |
| JPY | -0.430 | #5 | 0.965 | 1.723 | |
| USD | -0.580 | #6 | -0.692 | -0.078 | |
| AUD | -1.010 | #7 | -0.026 | 1.746 | |
| CHF | -1.080 | #8 | 1.152 | 0.970 |
BEER Panel · Macro-Momentum · Consensus (moy.)
| Zone | BEER Panel | Macro-Mom | Consensus | Blend | Bias |
|---|---|---|---|---|---|
| JPY | 0.879 | -0.948 | -0.034 | -0.399 | |
| AUD | -0.732 | 0.835 | 0.052 | 0.365 | |
| GBP | -0.361 | 0.474 | 0.056 | 0.223 | |
| CAD | 0.819 | -0.057 | 0.381 | 0.206 | |
| EUR | -0.620 | 0.075 | -0.273 | -0.134 | |
| NZD | -0.455 | 0.069 | -0.193 | -0.088 | |
Base - Quote
| Pair | L1 (Valeur) | L2 (Cyclique) | L3 (Régime) | Total |
|---|---|---|---|---|
| AUDJPY | -1.611 | 1.783 | 0.086 | 0.765 |
| GBPJPY | -1.241 | 1.421 | 0.090 | 0.623 |
| CADJPY | -0.060 | 0.890 | 0.415 | 0.605 |
| EURAUD | 0.111 | -0.761 | -0.325 | -0.499 |
| AUDNZD | -0.277 | 0.766 | 0.245 | 0.454 |
| USDJPY | -0.879 | 0.948 | 0.034 | 0.399 |
| AUDUSD | -0.732 | 0.835 | 0.052 | 0.365 |
| EURGBP | -0.259 | -0.399 | -0.329 | -0.357 |
| EURCAD | -1.440 | 0.132 | -0.654 | -0.339 |
| GBPNZD | 0.094 | 0.405 | 0.249 | 0.311 |
| NZDJPY | -1.335 | 1.017 | -0.159 | 0.311 |
Panel BEER (pooled DOLS BNP-style) · barre = décote (vert) / surcote (rouge) · trié par signal
| Devise | Spot | Fair Value | Décote ↔ Surcote | Écart | Verdict | Coint. | Signal |
|---|---|---|---|---|---|---|---|
| AUD /USD | 0.717 | 0.746 | | -3.9% | SOUS-ÉVALUÉ | +0.365 | |
| GBP /USD | 1.352 | 1.471 | | -8% | SOUS-ÉVALUÉ | +0.223 | |
| CAD /USD | 0.721 | 0.828 | | -12.9% ⚠ | NON FIABLE | +0.206 | |
| NZD /USD | 0.581 | 0.621 | | -6.4% | SOUS-ÉVALUÉ | -0.088 | |
| EUR /USD | 1.160 | 1.212 | | -4.3% | SOUS-ÉVALUÉ | -0.134 | |
| JPY /USD | 0.007 | 0.021 | | -68.6% ⚠ | NON FIABLE | -0.399 |
Fair value = équilibre Panel BEER — pooled DOLS G-6 avec effets fixes pays sur taux réel 10A + terms of trade (indice matières premières). Refit causal trimestriel. Écart < 0 = devise sous-évaluée (achat) ; > 0 = sur-évaluée (vente). Signal tradable = blend Panel BEER (30%) + Macro-Momentum walk-forward Ridge (70%) avec gate de cohérence : si le timing tactique contredit la valorisation, pas de trade (signal 0).
6 majeures + 15 croisées · fair = spot × exp(-[misval_base - misval_quote]) · trié par l'écart
| Paire | Spot | Fair Value | Décote ↔ Surcote | Écart | Verdict | Fiab. | Signal |
|---|---|---|---|---|---|---|---|
| USDJPY | 153.5 | 48.187 | | +218.6% | SUR-ÉVALUÉE | +0.000 | |
| USDCAD | 1.387 | 1.208 | | +14.8% | SUR-ÉVALUÉE | -0.206 | |
| GBPUSD | 1.352 | 1.471 | | -8% | SOUS-ÉVALUÉE | +0.223 | |
| NZDUSD | 0.581 | 0.621 | | -6.4% | SOUS-ÉVALUÉE | +0.000 | |
| EURUSD | 1.160 | 1.212 | | -4.3% | SOUS-ÉVALUÉE | +0.000 | |
| AUDUSD | 0.717 | 0.746 | | -3.9% | SOUS-ÉVALUÉE | +0.365 |
Fair value d'une paire = ratio des fair values de chaque zone (Panel BEER) — la paire est sur-évaluée si le spot est au-dessus, sous-évaluée en dessous. Signal = différentiel du blend BEER+Macro-Momentum de la base moins celui de la quote, avec gate de cohérence.
Preview on sample data, not updated.
Each model isolates one problem: the shock of a release, the ranking of G7 economies, or the gap to fair value.
Measures the gap between releases and consensus across 20 sub-indicators (inflation, jobs, policy rates, GDP, PMI), with an exponential decay of the information edge.
Role: Captures the repricing after a release, before consensus realigns.
Cross-sectional ranking of G7 currencies combining macro value, activity momentum and a monetary regime filter, with weights calibrated per currency.
Role: Builds a long/short basket from the fundamental ranking of the G7.
Equilibrium model (BEER): deviation of the real exchange rate from its long-run mean, completed by the Taylor-rule differential.
Role: Spots lasting valuation gaps and the pull back to equilibrium.
What our subscribers actually do with the biases they receive each week.
Sunday evening
The G7 bias matrix gives the direction of each currency before the open: you start Monday with a plan.
Before a release
The surprise score shows which currencies would react most if the print misses consensus.
At month end
Fair value gaps flag currencies that are persistently rich or cheap, useful when rebalancing.
01
Training windows are sealed: the model never sees information dated after its computation date.
02
The signal must stay positive across several cycles (inflation shocks, hikes, recessions), not a single favourable regime.
03
When statistical certainty is too low, no bias is published. Staying neutral protects capital.
One plan, no commitment: our model results by email, every Sunday.
A systematic frame at the start of each week, without intraday noise.
€29/ month
No commitment · or €290 per year (2 months free)
Try it free for 2 weeks, no commitment.
Every Sunday at 8 pm (Paris time), by email, so you start the week with a clear frame.
No. They are statistical model outputs, published for information. They do not account for your situation and are not investment advice.
No. Each send explains in a few lines what the bias means and when it applies. Technical vocabulary is kept to a minimum.
Yes. There is no commitment: the subscription ends at the end of the current month, on request.
Horacle Hub opens the Edge Finder, COT positioning and market regime detection.
Informational and educational content. The systematic signals shown are neither a solicitation to buy or sell nor personalised investment advice. Past performance and out-of-sample simulations do not guarantee future results. Any investment carries a risk of capital loss.