Ratio builder (A÷B)Line = A÷B. Gray band = rolling ±2σ (break out = beyond ±2σ). Dark solid lines = SPX / NAS100 benchmarks (apart from the ratio = divergence). Dots = momentum turns (▲ peak / ▼ trough) and band breaks.
▾Context lanes — rolling corr vs benchmark · percentile · ±2σ/95th-pctile stateⓘ
Three ribbons aligned to the chart above (drag to pan, wheel to zoom, click for the synced date line). Corr: rolling correlation of the ratio's bar-to-bar moves vs the benchmark — blue +1 = moves WITH it, red −1 = moves AGAINST it, grey = |corr|<0.2 (moving on its own — the exact number is in the tooltip). Percentile: where each day's level sat in its trailing 252-bar range — red 100 = top of range, blue 0 = bottom (first ~40 bars hatched while history builds — ramping). State: red = beyond +2σ or the 95th pctile, blue = beyond −2σ or the 5th, grey = inside range. Hatched cells = no data.
▾Momentum — velocity & acceleration of the ratioⓘ
Velocity = smoothed bar-over-bar % change (is the ratio rising or falling, how fast). Acceleration (bars) = change in velocity — when it flips through zero the ratio is INFLECTING, i.e. a trend may be turning before price confirms.
▾Recent turns & breaks (newest first)
Analogs of NOW — the last times this ratio looked like this
A signature of the ratio's state RIGHT NOW is built from four computed reads — its level stretch
(z vs its own trailing window), velocity, acceleration, and its rolling correlation to the S&P — then history is
searched for the individual past bars whose signature sat within Tolerance (euclidean distance in z-space) of
today's. Each match is a real dated bar; the chart draws ITS forward path (the ratio re-based to 1.0 at the match date)
over the next N bars — one colored line per analog, no average, no median, no cone (individual paths only). The
S&P 500 and NAS100 panels below draw what the index did over those same dates — the cross-check
that says whether the ratio's move came with the market or against it. Honest n: 0–3 matches is a valid answer
(shown anyway, no inference). Click any line or any date chip to highlight that one analog across all three panels.
Recomputes live as you change the ratio, Tolerance or Forward bars — never a forward-return / hit-rate stat (banned).
This ratio — forward paths from each match
S&P 500 (^GSPC) — same dates (cross-check)
NAS100 (^NDX) — same dates (cross-check)
Stretch tape — every instrument vs its OWN historyCompanion heatmap to the ratio chart: one row per universe instrument (SPX / NAS100 pinned on top), each day colored by where the instrument's rolling z (Band-lookback window) ranked inside ITS OWN trailing 252 bars — deep red = 100th percentile (stretched high vs itself), deep blue = 0th (washed-out low), pale = mid-range, grey hatch = history still building (amber r = ramping). ONE shared 0–100 ramp for every row — the SAME percentile ramp as the lane under the chart — so a vertical scan of any date column shows what was rich vs cheap TOGETHER: cross-row co-moves and splits that overlapping lines can't show. Weekly/monthly macro series are carried forward between prints. Click a cell = page-wide synced dateline + all-rows tooltip; click a row LABEL = hide/show that row; the sort buttons rank rows by the dateline-selected column (none selected = latest visible bar; marker burned on the axis). Shares the View window, wheel-zoom and drag-pan with the charts above.
ColorSort
Multi-scale divergence — RSI price↔momentum across 1h · 4h · 1d
A divergence is when price and its momentum (RSI) disagree — price grinds to a new
high while momentum quietly rolls over (a top warning), or a new low while momentum turns up (a bottom warning).
It is a computed read you cannot see by eyeballing the line, so it is the one class of marker we label
(R-UI-23). This runs on each instrument's RAW price series (NOT the A÷B ratio — the builder above is untouched).
Each instrument gets 9 cells = 3 timeframes (1h / 4h / 1d) × 3 swing lookbacks (short 5 / normal 13 / wide 34).
Cell color = type (bear ▼, bull ▲,
hidden-bear ▿, hidden-bull ▵); size = strength;
it fades as it ages; blank = none; grey hatch = no intraday history for that instrument. The Confluence column
nets bull vs bear across all 9 — a divergence confirmed across many scales is far higher-conviction than one lone cell.
Pick a view mode for a different lens (each is explained on-screen). Click any row to load it in the price+RSI
chart, which draws the exact swing pair the read is computed from. Macro levels (M2/CPI/RRP/…) are excluded —
divergence is a price-action tool.
Retail P/C ref
loading…
loading divergence grid…
Macro levels — excluded on purposeNot price action, no intraday. Excluded from this grid: M2 (weekly & monthly), CPI, reverse repo (RRP), Fed balance sheet (WALCL), Treasury cash (TGA), and the FINRA margin statistics (margin debt, free credit, excess-leverage).
WhyThey are level/flow series with no real intraday cadence, so a price↔RSI divergence on them would fabricate a signal.
InsteadBuild ratios FROM them in the A÷B builder above.
Macro-ratio radar — common pairs ranked by extreme + momentumClick a row to load it into the builder. z = level extreme; vel = current momentum (▲rising/▼falling); corr near 0 = moving on its own vs the S&P. Spark = rolling z of the ratio, last ~120 bars (dashed guides = ±2σ; line red = currently stretched high, blue = low). Cell shade, shared scales: z / pctile / vel on the house diverging ramp — blue = low · white = mid · red = high (z clamped ±3σ, pctile 0–100 centered on 50, vel vs the column max); corr cells use the decoup-matrix ramp — red = −1 moves against S&P · blue = +1 moves with it, near-white ≈ 0 = decoupled.
What a ratio is (apps2/ratios)A ratio divides one instrument's price by another (gold÷oil, 10Y÷oil, VIX÷VVIX…) to expose a relationship raw prices hide.
Universerates · commodities · FX · vol · factors (macro).