Geopolitical analysis and financial market impacts. Understand correlations in real time.
Subscribe (RSS)Saying "68% accuracy" is refutable in one sentence: the market rises 55% of the time. So we compared the engine to random, to "always up" and to naive momentum, matched on the same assets and dates. Honest result: we beat two out of three.
A high hit rate can hide huge losses. We measured the average gain of each bet in units of volatility (σ) — the real judge — and found that accuracy told an incomplete story. A pre-registered analysis, published whatever the result.
Some signal families lose structurally. Rather than quietly erasing them from the record, we put them "under observation" — measured but out of the traded flow — and their only way back is the counterfactual, never an editorial call.
Anyone can say "I called it" after the fact. So we publish the cryptographic fingerprint of every signal the moment it's issued, reveal the content later, and anchor the whole thing outside our own servers. Here's how — and, more importantly, what it doesn't prove.
You can make almost any strategy shine in the past — just let a little of the future leak in. Our backtest replays the exact live pipeline, dated to the day: reconstructed market context, walk-forward weights, news frozen at its own date. It's slower, less flattering, and that's the point.
Our displayed accuracy was stalling, our virtual portfolio was bleeding. Instead of massaging the numbers, we put the engine under the scalpel: a full forensic audit, 7 root causes, retroactive corrections — and a public counter reset to zero. Here is everything, unfiltered.
Threats over Hormuz, Iran-Israel escalation, nuclear tensions: the most geopolitically loaded quarter in years… and the S&P 500 gained 11%. This paradox has a name — the complacency regime — and numbers. Ours, measured across 341 signals.
A financial signal isn't a frozen object. Its informational value decays over time like a radioactive isotope. Here's how we imported the half-life concept from physics to make our signals honest — and why your Kelly allocation should follow.
Behind the dashboard's main gauge, five aggregated sub-scores. VIX, yield curve, fear & greed, geopolitics, momentum — exactly what we measure, why, and how to tune it to your style.
Four different rules predicted oil up during the Iran crisis. The market dropped -8%. We asked Mistral Large to dissect each error, and the pattern that emerged changes how we build our signals.
The Ormuz crisis doesn't only produce losers. Oil soaring, gold paradoxically falling, Bitcoin in free fall — analysis of the winners and losers across financial markets.
The blockade of the Strait of Hormuz isn't just about oil. Fertilisers, helium, aluminium — a cascade analysis of what's rippling through financial markets and your portfolio.