Category and positioning
Research and validation, before execution. The gap between a notebook and an enterprise risk system.
Execution platforms are built to trade. Enterprise risk systems are built for firms. Notebooks are built for anything, by you, again each time. Quant Compose sits deliberately in the gap: rigorous enough for an institution's methodology, accessible to an individual or a class, and finished before an order exists.
Approaches, compared fairly
Categories rather than names, including the row that says when not to use us.
| Criterion | Notebooks and scripts | Spreadsheets | Execution platforms | Enterprise risk systems | Quant Compose |
|---|---|---|---|---|---|
| Best at | Open-ended exploration and any method you can code | Quick arithmetic, ad hoc allocation, sharing a table | Turning a rule into orders, live and paper, with a strategy marketplace | Firm-wide risk at institutional scale and price | Research and validation before execution, for individuals, students and small teams |
| Validation by default | Whatever you wrote this time | None | Backtests, usually in-sample and rarely purged | Extensive, configured by a team | Chronological split, train-only scaler, purge gap, walk-forward folds, worst fold reported, trials counted |
| Reproducibility | Depends on discipline: environments drift, cells run out of order | Fragile: formulas and pasted values | Strategy formats are often undisclosed | Strong, inside the vendor | Projects isolate state; workflows are versioned specifications; runs are records; the experiment record hashes its inputs |
| Coding required | Yes | No | Sometimes | No, but heavy configuration | No. Configuration, not scripting; a custom-script space is labelled future |
| Method library | Everything in the ecosystem, uncurated | Built-in functions | Indicators and strategy templates | Vendor models | 119 catalogued methods with typed parameters and citations; 72 checked against a published reference |
| Execution | If you wire it | No | Yes, that is the point | Through order management systems | No. Nothing places or routes an order |
| Cost and access | Free, plus your time | Cheap | Free tiers to subscription | Institutional contracts | Free during private beta; planned plans published as planned |
| When not to use it | When results have to be defended by someone who did not write the code | When the method is more than a formula | When the question is whether an idea is real, not whether it can trade | When you are one person, a class, or a small team | When you need to execute, when you need intraday data, or when your method is not in the catalogue yet |
The comparison describes categories of tool as they are commonly used. It is not a claim about any named product, and the Quant Compose column states only what the workspace does today.
Six things the gap needs
The pillars, in priority order.
- 01
All in one, no fragmentation
Data, methods, pipeline, analysis, saved results and report in one project. Nothing is re-exported between tools to keep going.
- 02
No coding required
Configuration, not scripting. The methodology is built in and explained inline as you use it.
- 03
Honest by construction
Purged, embargoed, walk-forward validation on by default. Forecasts scored against a naive baseline even when the baseline wins. Weak results stay visible.
- 04
Reproducible
Projects are isolated. Saved results keep their parameters. The experiment record hashes its inputs and the trial ledger is append-only. Anyone on the team can re-run a result.
- 05
Composable
A pipeline you build once and re-run on new data, not a drawer of disconnected calculators.
- 06
Minutes, not weeks
What used to be a week of notebook plumbing becomes an afternoon of configuration, with the record kept for you.
What this is not
Stated plainly, because it matters as much as what is included.
- Not a live trading or order-routing system. The health endpoint reports live_order_routing: false.
- Not a guarantee that a validated strategy will perform. Validation reduces self-deception; it does not remove risk.
- Not a substitute for your own judgement on methodology. The workspace refuses and explains; it does not decide for you.
- Not investment advice. Nothing it computes is a recommendation to buy or sell any security.
If the gap is where you work, this is for you.
Bring one real question to the private beta. Applications are read by a person.