Signature Pivots: Conviction & Divergence

Signature Pivots: Conviction & Divergence

Every chart prints dozens of swings and most of them are noise. This
tool has one job: help you pick the pivots worth anchoring analysis
on. It finds confirmed swing highs and lows, filters out the minor
wiggles, stamps each survivor with a time signature, grades it by
volume conviction, and warns when a new extreme was made on hollow
volume. It draws and grades structure - it never tells you to buy or
sell.

How pivots are found

A swing high is a bar whose high is the highest of N bars on either
side (N is the "Pivot strength" input; swing lows mirror it). A pivot
can therefore only confirm N bars after the fact - that is detection
lag, not repainting, and once a marker prints it stays. On top of that
sits a size filter: a new pivot must be at least "Min swing size" ATRs
away from the previous kept pivot, so a daily chart shows the swings
that matter instead of every five-bar fractal. Set it to 0 if you want
them all.

The signature

This is the unusual part. Each pivot's UTC timestamp is converted,
through sidereal time at the market center you choose, into the degree
of the zodiac rising on the eastern horizon at that exact moment - the
Ascendant. That gives a number from 0 to 360 which sweeps the full
circle once every 23 hours 56 minutes. It depends only on the clock
and the location, never on price, so the same moment always maps to
the same number on every chart and timeframe. In plain terms, it tags
each pivot with where in the daily rotation cycle it formed. The old
observation being tested here is that turns tend to recur at the same
phase of that cycle.

Matching

The last 24 pivots (a setting) are kept in memory. Each new pivot's
signature is compared against stored pivots of the same kind only -
highs against earlier highs, lows against earlier lows - measured
correctly around the 360 ring, so 359 and 1 are two degrees apart, not
358. If the nearest same-kind signature sits within the match
tolerance (1 degree by default), the pivot is flagged: its label
highlights with a star, a dotted line links it back to the pivot it
repeats, and an alert fires. A match is a candidate worth studying,
not proof of anything - the defaults are deliberately strict so that a
star stays rare.

Conviction and divergence

Volume flow is measured with on-balance volume: each bar's volume
added on an up close, subtracted on a down close. A pivot's conviction
is simply whether OBV stood above or below its own moving average at
the pivot bar - was volume flow supporting the move that made the
swing, or fighting it. Divergence looks at consecutive extremes: a new
swing high above the previous one, with OBV lower than it was at the
previous high, prints a bearish divergence mark (price stretched,
volume flow did not follow); new lows mirror that for bullish
divergence.
Why these three live in one tool: they are three questions about the
same object. Did this pivot form at a recurring phase of the rotation
cycle? Did volume back the swing? Was the extreme hollow? Read
together on one label, they turn a plain zigzag into a graded list of
anchor candidates.

Reading the chart

H and L mark confirmed swing highs and lows. The caret after the
letter is conviction at that pivot (^ above the OBV average, v below).
The number is the signature in degrees. A star and highlighted color
mean a signature match, with a dotted line back to the earlier pivot
it repeats. "Div" marks print at divergent extremes. The zigzag
connects consecutive pivots so the structure stays visible. The HUD
shows the live conviction state, the last signature, the last
high-value pivot, and running pivot and match counts - event counts,
not performance figures.

Settings worth knowing

Pivot strength and Min swing size control how selective the swing
detection is. Match tolerance and Match memory control how strict a
"repeat" must be - loosening either will flag more pivots, and past a
point most of those extra stars are coincidence, which is why the
defaults are tight. The Market Center sets the geographic reference
for the signature; keep one consistent choice across your charts so
signatures stay comparable (for 24/7 crypto any consistent center
works). "Only show high-value pivots" hides everything except matches
if you want the quietest possible chart.

Using it

It runs on any timeframe, with one honest caveat: the matching layer
is most alive on intraday charts. On daily bars every pivot stamps at
the same time of day, so signatures drift about a degree per day and
repeats are naturally rare and far apart - on a daily chart this tool
is mainly a swing, conviction and divergence grader. Use it as a
filter, not a trigger: when you need a swing to anchor other work on,
prefer the starred pivots, and among those prefer the ones whose
conviction agreed with the swing and that carry no divergence warning.
Alerts are provided for a high-value pivot, bullish and bearish
divergence, and each new confirmed swing.


What is original here

Swing detection, OBV and divergence are standard public techniques.
The original work is the signature layer - the rising-degree
computation runs entirely on the chart from the timestamp, with no
external data - together with the ring-correct same-kind matching and
its pivot-to-pivot links, and the fusion of the three readings into a
single graded label. Written from scratch in Pine v6; no open-source
code was reused.


Honest notes

Pivots confirm late by construction. Some matches will always be
coincidence; the tolerance and memory settings exist to keep that in
check, and a star is an invitation to look closer, nothing more.
Different market centers give different signature numbers, so keep
yours consistent. Educational study tool; the HUD counts events, not
wins; no performance claims; not financial advice.