The platform
A measurement layer that turns public market data into a daily structural reading — the market organised as an atlas of economically coherent blocks, with a System reading across them and a written weekly analysis, alongside the risk tools you already run.
See it running on this week’s market.
A 30-minute demonstration on current data — the dashboard, the weekly reading, and where structural risk sits in your process.
or take the 90-second tour ↓
The dashboard
One screen, read in minutes, updated daily on the last settled session (D−1), the run day shown as preliminary until its inputs are published: the Topological Survival Score and its zone, the integrity of the manifold, and the contagion network showing where stress concentrates — epicentre, core, induced and periphery nodes.
The Atlas — the market in blocks
Large universes are not read as one flat panel. The flagship technology atlas covers 99 dimensions in nine economically coherent blocks — platforms, chip design, chip equipment, cybersecurity, software and data, Asia, AI infrastructure, Europe, and the macro context including the currencies of the chip supply chain. Each block is measured as a structure of its own. A System reading across blocks measures whether they still move independently — the number that collapses when the market becomes one trade. When alarms fire, an attribution module says where they concentrate: stress with an address. Underneath sits a market-state layer — thirteen macro conditions classified daily against thresholds frozen on the 2015-2019 reference years, filling the six-week latency of the official cycle phase.
The 90-second tour
Inside the Tech Manifold: a walkthrough of the console, its geometric signals and the daily reading — in a minute and a half.
From public data to a documented reading
Inputs
Public market prices for a defined universe of listed instruments — equities, indices, rates and commodities proxies. No client data, no positions, no proprietary feeds. The same inputs produce the same reading, every time.
Computation
The universe is mapped as a geometric structure and read at two levels: each block on its own recent window — its Topological Survival Score, its deformation, its curvature alarms — and the System across blocks. Alarm thresholds are causal and frozen; each asset is compared only with its own past. Attribution then shares the alarms out across blocks over the trailing month. The mathematics is published — see Research.
Outputs
A daily reading on the last settled session (D−1), with the run day flagged as preliminary until its inputs are published: the atlas mosaic with each block’s score and zone, the System number, the attribution shares, the contagion map with its epicentre. Once a week, a written analysis documents what moved, against which thresholds — and the limits of what the reading can tell you. Every screen has a manual chapter: what it answers, how to read it, the method in plain language. What “settled” and “preliminary” mean is stated once, in the glossary.
How it fits your weekly routine
- Each morning — a two-minute glance at the settled D−1 reading: the zone first, the stress location behind it.
- Each week — the written manifold analysis documents the week’s structural picture, in language a risk committee can minute and file.
- When the zone moves — the dashboard shows what changed, where, and against which threshold — a documented state, not a black-box alert. The reading sits next to VaR and volatility numbers as an independent second opinion on structure.
- Audit trail — every reading is reproducible from its inputs. Model validation teams can re-run any historical reading and obtain the same result.
Governance and security, by construction
No client data
The platform never receives portfolio holdings. Readings are computed on public data for agreed universes — there is nothing sensitive to store, and nothing to breach.
Deterministic and auditable
No machine-learning black box, no stochastic output. The computation is deterministic — same data, same numbers — every threshold is frozen and dated, and the methodology is published, so any reading can be independently verified.
A layer, not a signal
The platform measures structure. It does not produce trading recommendations, price predictions or portfolio advice — and it is not positioned to. Universes are governed by declared entry and exit criteria, reviewed quarterly — never by performance.
Start with the 30 minutes.
A 30-minute discovery session: a demonstration of the dashboard and the weekly reading on current market data, then a discussion of where structural risk hurts in your process.
Not ready for a call? Send us a question — methodology, scope, or how the reading would apply to your universe.