One map instead of 31 charts · Lines = fronts · Tracers = money flowing to the stronger · Circle size = strength · Green/red arcs = winning/losing vs that enemy · Click “?” for how to use it
The free page is the live battlefield; Pro is your always-on field intelligence officer. The founding-member program is open:
A monitoring & alerting tool — not investment advice, no performance promises. Features ship per roadmap. Payments handled securely by Stripe.
In one line: we don’t compare “how much it moved” — we compare “how unusual the move is relative to that pair’s own normal volatility.”
Pairs have very different natural ranges: a 0.05% five-minute move is routine for GBP/JPY but a rare event for EUR/CHF. Ranking by raw % change lets naturally-volatile pairs dominate the board while genuine anomalies in quiet pairs drown.
① Take each pair’s temperature: the standard deviation σ of its 1-minute returns over the last 6 hours — “how much it normally moves per minute” (refreshed every 30s).
② Divide the current window’s return by that normal volatility scaled to the same window:
Statistically this is a z-score. The unit is “σ” (standard deviations).
| σ-score | Meaning |
|---|---|
| 0 ~ ±1σ | Normal fluctuation — mostly noise |
| ±1 ~ 2σ | Directional push |
| ±2 ~ 3σ | Notable anomaly — pay attention |
| beyond ±3σ | Violent move, statistically rare (<0.3% of the time) |
Each currency’s strength = the signed average σ-score across its fronts against every other nation. EUR at +1.5σ means “EUR’s average push against the field is 1.5× the normal volatility of those fronts.” Circle sizes, tracer flow, the ranking and the Main event are all driven by this score.
The header shows the Main event: the direct cross of the current strongest × weakest currency. The map separately labels the three hottest fronts by per-pair σ-score — different measures, often different pairs.
Reversals usually start on lower timeframes: when the clearest higher-TF trend cracks on the lower TFs, that is often the first sign of exhaustion. Radar rule: |higher-TF σ-score| ≥ 1.2 and a counter-move on a lower TF (fast ≥ 0.6σ or slow ≥ 0.5σ). Both lower TFs counter at once = “2-TF confirmed” (🔄 on the map + a feed alert, throttled to one per pair per 30 min); one TF = “1-TF watch”. In each radar row, a green value passed its threshold; grey = right direction but below threshold.
The TF combo is selectable next to the radar title. The ★ preset “1d ← 4h+1h” is the official signal preset, joined by the experimental 🧪 “1w ← 1d+4h” — Pro Telegram alerts push these two combos (experimental messages are tagged 🧪 and tracked separately in the ledger); both are 🔒 Pro-only on the free page. Intraday combos (4h/1h base) are for your own screen-time exploration and are never pushed. σ-normalization keeps thresholds meaningful across timeframes.
| Event | Trigger |
|---|---|
| ⚔️ storms | a pair moves > 3σ of its own norm within 1 minute (throttled per pair, 2 min) |
| 🔥 ROUTS | same but > 5σ; the losing base emits a shockwave |
| 👑 throne / 💀 weakest | #1 / last place changes hands (follows your selected window & mode) |
Combat events always use the 1-minute window with σ thresholds, so alert density stays roughly uniform between quiet and busy sessions. The bottom INTENSITY meter is fixed at 5 minutes.
· For ~1 hour after major releases (NFP, CPI…) the whole board’s σ is elevated, so subsequent moves read “calmer” — an inherent property of normalization.
· Prefer raw % view? Toggle “σ-Norm” off in the header.
Hover any front to see raw return, σ-score and the pair’s 1-min σ.
Built by 𝕏 @jesselaunz
Axis = your local time; bars = each market's local 09:00–17:00 (DST-aware), bright = open now, faded = weekend; gold line = now. The London × New York overlap usually carries the highest liquidity and volatility — a hotter INTENSITY meter there is normal.
| Currency | Policy rate |
|---|
Largest positive carry (long side earns overnight interest): —
Rates as of —, policy rates only; actual swap depends on your broker (admin fees can flip small differentials). Positions held over Wednesday close usually incur triple swap.
| Market | Spec net | Weekly Δ | Net/OI | Idx52w | 2y curve | Report date |
|---|
How to read the three columns:
· Spec net = speculators' longs − shorts (contracts). Positive = more money betting up, negative = betting down — direction comes from this sign only;
· Net/OI = net position ÷ that market's total open interest. Removes market-size differences — the "concentration".
Example: —;
· Idx52w = where today's net sits between its 52-week high and low (0 = most short of the year, 100 = most long, 50 = midpoint).
It measures position relative to its own year — not direction; ≥90 / ≤10 gold ⚠ = crowded zone.
Read all three together: net = size, Net/OI = concentration, Idx = abnormality. The more crowded the positioning, the harder the reversal tends to hit when the radar fires.
· DX (US Dollar Index futures) reads the OPPOSITE way: it is the only contract that bets on the dollar directly (positive = net long USD).
Our 22-year study found its extremes are a momentum signal, not reversal fuel — after positioning enters the most-long-dollar bucket,
the dollar index gained +1.07% over the next 8 weeks on average (t=5.18); the most-short bucket averaged −0.58%. That is why DX extremes are
marked ⇈/⇊ (with the trend) instead of the ⚠ used for single currencies. A likely reason: the dollar's counterparties are spread across seven
currencies with no single squeeze point, and dollar extremes tend to coincide with macro regime shifts (2008 crisis, 2014 hiking cycle, 2022 tightening) that persist. Published by the CFTC every Friday, auto-refreshed here every Saturday — all four columns free; history and two-year charts on the COT report page. Click a market code ↗ for CME's official daily volume & open-interest page — price up with OI falling = old positions covering, price up with OI rising = fresh money; cross-check it against the weekly COT to track unwind progress.
Economic calendar (ForexFactory) · CFTC COT reports · Central banks
Hover for weekly values · Data: CFTC non-commercial net positioning, last 104 weeks