Cycles Frontier Technology Part of Cycles IQ · in build
Know which cycles to trust.
Every scan today is computed, looked at and thrown away. The Cycles Knowledge Graph keeps them. It follows every cycle through its life, connects cycles across markets, and scores every projected turn against the turn that came. The scanner becomes a memory.
After the scan
The scanner shows which cycles are there. The graph tells you what they are worth.
Can I trust this cycle?
Today you judge a cycle by eye. The graph gives every cycle a life story: how old it is, whether its length drifts, whether it has reset. From thousands of completed cycle lives it tells you how likely a cycle like this one is to survive one more swing.
Is it part of something bigger?
A projected low in the S&P 500 means more when the same cycles in most global indices bottom in the same weeks. The graph measures that confluence across markets and shows which markets tend to move first.
Has this kind of call worked before?
Every projected high and low is written down before it happens and checked after, including the misses. A projected turn stops being a date and becomes a window, as wide as the record says it should be.
Why it is new
Rich in assertion. Poor in evidence. Until now.
The cycles field has catalogued thousands of cycles and hardly ever checked what they did next. That is why academic finance dismisses it, and it is not entirely wrong to do so. The graph produces what the field does not have today.
| What the graph measures | The field today |
|---|---|
| How long cycles live, how they drift and when they reset, across a large universe of markets | Not measured |
| A scored record of cycle projections, made before the fact and checked after it | Not available |
| A test of Hurst's nominal model and its harmonic ratios over thousands of instruments | Asserted from case studies |
| Cycle breadth: how many markets turn together | Illustrated, not measured |
| Which cycles are tied to the calendar, and which are not | Assumed by seasonal charts |
| Which settings of the method actually work better | Anecdotal |
Dewey's catalogue, turned into a living instrument. Cycle analysis gets a way to be wrong in public, and to get better because of it.
What you get
Nine answers. Each one replaces a judgement with a measurement.
- 01
Cycle biography
Age, length drift and resets for every cycle. "Looks reliable" becomes a survival probability.
- 02
Calibrated turn windows
How far off past projections were, by market and cycle length. You see where the method works and where it does not.
- 03
Cycle breadth
The share of markets in a cycle band that are rising, topping, falling or bottoming. Market breadth, for cycles.
- 04
Stand-down warning
When many cycles in a sector die at once, projections are least reliable. The graph tells you when to stand down.
- 05
Tested lead and lag
Which markets move first, and whether the liquidity cycle really leads equities. Most links fail the test. The few that hold are worth a great deal.
- 06
Analogues
The past dates whose cycle picture looked most like today, and what followed.
- 07
A week that starts from what changed
New cycles, dying cycles, building confluence. No walk through fifty charts.
- 08
Calendar-driven or not
Which cycles are really tied to the calendar, from the annual cycle down to month-end and option expiry.
- 09
What works
Every call is made by a known version of the method and checked afterwards. The graph learns which settings have forecasting value.
Seasonality, measured
We do not do seasonals. We measure which cycles are calendar-driven.
A seasonal chart averages decades of returns by calendar date and presents the average as a forecast. A few outlier years dominate it, and it changes shape with every sample.
The graph checks every cycle near one year, or a fraction of it, against the calendar itself. If it stays locked for years, it is calendar-driven, usually for a real reason, as in gas, power and crops. If it drifts, it is a market cycle that happens to be about a year long, and trading it by the calendar would be a mistake.
The same check runs at the short end. A 21-day cycle locked to month-end is a flow effect, and you want to know that before you trade it.
Ask what-if
Ask what-if. Get data, not opinion.
Whether a blend of the three strongest cycles beats the dominant cycle alone, whether detrending helps, whether a longer look-back window helps: today these are matters of belief.
As a Cycles IQ member you put such a question to the graph. It is tested over years of history across a panel of markets, and a few weeks later you get the answer as a finding, with the evidence behind it. What the graph has not measured, it answers with "not measured", never with a guess.
The first three questions of this kind are already running. Findings Report 01 is the graph's first result.
How you use it
Scan the lenses. Ask the agent. Build on the tools.
See it at a glance
Cycle map, market card, confluence calendar, births and deaths, findings. Instant, and ready to paste into a client note.
Ask what nobody predefined
The model picks the tools and tells the story. Every number and every chart comes from the graph, and every sentence points to its source.
Build your own screens
Query the graph from Claude, ChatGPT or Cursor and build your own screens on it, with the Cycles AI Skill as the guide.
Honest by design
A measuring instrument pointed at the method itself.
The graph does not make cycles more predictive than they are. Across thousands of markets, coincidences are everywhere, and a well-built graph can make nonsense look authoritative. These rules are part of the product.
- No hindsight. Every result uses only what was known on that date.
- Persistence first. A cycle counts only when it holds over several scans, a link between markets only when it survives.
- Tested against chance. Every relationship has to beat random coincidence before it is shown.
- Misses included. Every projection is recorded before the fact and published with its outcome.
- Methods are scored, never people. No member forecasts, no trades, no personal track records.
The graph watches the markets, watches itself, and changes only on its own evidence. Some of what it reports will contradict what the cycles community has believed for decades. That is the point: Market Zeitgeist becomes the place that has the data.
How to get it
Part of Cycles IQ. From the first finding.
The Knowledge Graph is part of the Cycles IQ membership, not a separate tier. It is built in stages, evidence first: Findings Report 01, then the graph tools for MCP and API, then the lenses and the built-in agent. Members get each stage as it lands.
-
Membership
Request an invitation
Cycles IQ
The Knowledge Graph with the Cycles API, the MCP server and the Cycles AI Skill. Members put their own what-if questions to the graph.
-
Research
Subscribe
Newsletter
The findings reports and a weekly cycle map generated from the graph.
-
Institutions
Get in touch
Institutional access
Snapshot exports or read access to the graph, on request.