Turn your market data into tested trading signals
AlphaGen helps commodities professionals turn proprietary and external data into systematic market indicators, through a structured research process grounded in their economic priors.
One month of full access · one named user · guided onboarding.
Illustrative interface. Historical backtest results are evidence about a hypothesis, not a guarantee of future performance.
The problem
Your best data is doing the least work.
The information that could sharpen your edge — physical prices, flows, inventories, proprietary series — rarely makes it out of a spreadsheet. Every new idea starts from zero, and most of them never get properly tested.
More data. Limited use.
Fundamentals, physical prices, flows and stocks, operational data, alternative sets, proprietary series — most of it is reviewed manually, held in spreadsheets, and interpreted differently by each user.
Every hypothesis restarts from scratch.
A spreadsheet for one idea. A Python notebook for another. A request to the quant team for a third. Each hypothesis becomes a new research project, with its own methodology and its own maintenance.
Market expertise stays upstream.
The trader knows which relationships are economically plausible: the direction, the lag, the seasonality. But turning that judgement into a systematic indicator usually requires programming, statistical work and infrastructure they don't own.
The gap isn't access to data. It's the absence of a consistent process to go from “this dataset looks relevant” to “we've tested the relationship in a repeatable way and we know whether it's worth deeper use”.
The key idea
Your market expertise guides the research.
AlphaGen doesn't search blindly across every possible relationship. You bring your economic reasoning up front — as priors — and the platform runs a consistent process to test, validate and monitor an indicator built on those priors.
A prior is a market hypothesis you supply before the analysis begins: the expected direction of a relationship, its seasonality, whether it carries a trend.
Priors are the language AlphaGen uses to turn your market knowledge into an explicit, testable input — instead of leaving it implicit in a spreadsheet or a notebook.
- Market expertise
- Economic rationale
- Data selection
- The priors
- Interpretation of the output
- A consistent construction process
- Standardised validation
- Statistically robust testing to avoid overfitting
- Performance reporting
- Ongoing updates and monitoring
- Export by file or API — Excel plugin in development
Market expertise, structured by software. AlphaGen doesn't replace the trader, the analyst or the quant. It removes the need to rebuild the same data-to-signal research workflow every time a hypothesis changes.
How it works
Data. Priors. Signal.
A structured workflow that replaces the recurring spreadsheet, notebook or quant request behind each new hypothesis.
Load your price data
Import the target market series — a front-month contract, a spread, a physical benchmark — that the analysis should try to explain.
Add your variables
Bring in the numerical series you think carry market information: fundamentals, physical prices, flows, operational data, alternative sets, internal indicators.
Set your priors
For each variable, declare the expected impact direction, whether the relationship is seasonal, whether it carries a trend, and how observations should be aggregated.
Calibrate on history
AlphaGen fits an aggregated intensity signal on historical data and produces a NAV-vs-Benchmark backtest with a full set of performance statistics.
Predict on new data
Once calibrated, generate the latest signal — a directional intensity on a 0–100 scale, labelled Bullish, Neutral or Bearish, with a per-variable decomposition.
Iterate and monitor
Adjust variables and priors, re-calibrate, and check whether the relationship keeps behaving as expected as new observations arrive.
Export or integrate
Move kept signals into your existing workflow via Excel/CSV export, the AlphaGen Excel plugin, or the API — whichever fits your team.
The benefits
A more consistent path from hypothesis to research output.
Built for commodities markets — and applicable to other data-driven markets — AlphaGen productises the recurring parts of the workflow so your team focuses on judgement, not plumbing.
Test more hypotheses, faster.
Cover the recurring part of the research workflow once, and reuse it across each new market idea instead of rebuilding a spreadsheet or notebook per hypothesis.
Reduce discretionary and methodological bias.
A single, consistent construction and validation process across users and hypotheses. Explicit assumptions in, statistically robust evaluation out.
Make your assumptions explicit.
Priors turn implicit trader judgement into an auditable input: direction of impact, seasonality, trend, target series. What was in the analyst's head becomes a documented research asset.
Get more out of your proprietary data.
Test whether the datasets your organisation already owns carry useful market information. The advantage of your data comes from your ability to systematically evaluate it.
Check whether the relationship still holds.
Once calibrated, indicators are updated as new observations arrive — so you can tell active relationships from decayed ones and decide when to recalibrate.
Live inside Excel, or connect via API.
Kept indicators can be exported to CSV or accessed via API today — and pulled straight into a spreadsheet through the AlphaGen Excel plugin, currently in development.
Bring AlphaGen signals into your spreadsheet.
Most traders already live in Excel. The AlphaGen Excel plugin will pull the latest intensity signal, its Bullish / Neutral / Bearish label and its per-variable breakdown straight into a cell — so a fresh calibration doesn't mean rebuilding your existing workbook.
- Reference any project or calibration by name
- Refresh on open, on demand, or on a schedule
- Compose with your existing formulas, pivots and charts
- Same access controls as the web app — no shadow copies
Preview only. The AlphaGen Excel plugin is currently under development — the illustration below is indicative and does not reflect the final interface, function names or feature set.
=ALPHAGEN.SIGNAL("brent-fundamentals")| A · Project | B · Signal | C · Label | D · Updated | |
|---|---|---|---|---|
| 2 | brent-fundamentals | 72.4 | Bullish | 09:42 |
| 3 | wheat-basis | 48.1 | Neutral | 09:42 |
| 4 | copper-ndvi | 31.6 | Bearish | 09:42 |
| 5 | gas-inventories | 63.9 | Bullish | 09:42 |
Illustrative mockup of the Excel add-in (in development). Final function names, interface and features may differ.
The output
From hypothesis to a monitored market indicator.
Every run produces a documented artefact you can review, share and challenge. The output is a research indicator — not a trade recommendation.
A continuous, standardised score from 0 to 100 that summarises where the market sits given your data and your priors.
The most recent intensity, labelled Bullish, Neutral or Bearish, with a per-variable decomposition so you can see what is driving it.
How the aggregated strategy has behaved historically against the underlying, with the full set of performance statistics — Sharpe, Hit Ratio, Max Drawdown, Annualized Return and more.
Once calibrated, the signal is refreshed as new observations arrive — so you can see whether the relationship is still active or has decayed and needs a recalibration.
inventories68.0Bullishphysical_prem61.0Bullishflows49.0Neutralndvi_proxy34.0Bearish
Illustrative values shown for interface reference only. Historical backtest results are evidence about a hypothesis; they are not a forecast or a guarantee of future performance.
One month of full platform access on hypotheses defined by your team.
Proprietary data
Built for data your organisation doesn't want to expose.
A large part of the market information that matters — physical prices, operational data, internal series — is proprietary. AlphaGen is designed so you can put that data to work while keeping control over the raw series.
During the Guided Pilot we'll walk your team through our current data-handling controls and share the anonymisation script so you can start on your own terms.
Client-side anonymisation script
AlphaGen ships a Python script that runs on your machine and anonymises your time series before they leave your environment. Column names, identifiers and scale can be obfuscated locally — only the transformed series is uploaded to the platform.
- Runs entirely client-side, on data you control
- Preserves the statistical structure the model needs
- You keep the mapping — we never see it
End-to-end data encryption
We are implementing full encryption in transit and at rest, with client-controlled keys and a documented data-handling standard. This will complement — not replace — the client-side anonymisation option, which some teams prefer to keep as a first line of defence.
Scope of the tool
Research infrastructure — not an investment recommendation.
AlphaGen structures the research. You keep control of interpretation, portfolio construction, risk management and execution.
What AlphaGen is
- Generates directional market indicators
- Applies a consistent research methodology
- Uses statistically robust testing to guard against overfitting
- Reduces the risk of discretionary and methodological bias
- Monitors accepted indicators as new data arrives
- Supports proprietary data with client-controlled handling
What AlphaGen is not
- Executes trades or connects to brokers
- Replaces the trader, the analyst or the quant team
- Provides investment advice or trade recommendations
- Guarantees returns or future performance
- Produces a production-ready trading strategy
- Removes the need for portfolio, risk and execution processes
A signal is a research output — a score or a directional probability. A strategy additionally requires entry and exit rules, position sizing, portfolio constraints, risk limits, execution and governance. AlphaGen produces the former and supports the latter, but never conflates the two.
The pilot
Start with a hypothesis.
The best next step isn't another generic demo — it's a full month of platform access on your own data, with an AlphaGen engineer at your side.
One month. Full access. Real hypotheses.
One named user on your team gets unrestricted access to the platform for a full month, starting with a live onboarding session. Learning the tool takes time — the pilot gives your team enough of it, with an AlphaGen engineer assisting throughout.
Onboarding session
A live session with an AlphaGen engineer to set up your workspace, review your first hypothesis and walk through the platform on your own data.
Full platform access
Unrestricted use of the product for a full month — enough time to move beyond the first hypothesis and really put the workflow through your own market ideas.
Direct assistance
We stay reachable during the pilot to help you frame priors, set up variables and interpret results — via chat, email or short working calls.
Your decision
At the end of the month, you decide whether to move to an annual license. No auto-conversion, no pressure.
At the end of the month, your user has calibrated and reviewed real signals on your own markets — enough hands-on experience to decide, on the merits, whether an annual license makes sense.
Start with a hypothesis.
The AlphaGen Guided Pilot runs the full research workflow on a market hypothesis defined by your team. One month. One user. As many indicators as you want to build yourself.
Success criterion: your user creates and validates a signal on your own hypothesis.
Contact
Bring a hypothesis. We'll take it from there.
Tell us which dataset you'd test first and which market it should inform. We'll come back with a short intro call to see whether a Guided Pilot fits — or a longer conversation if you have team, licensing or API questions.
Quick response
We typically reply within 24 hours on business days.
Book a working session
A 15-minute intro call or a full 30-minute product walkthrough on your data.
Team & enterprise
Multi-user workspaces, API access and dedicated support for larger accounts.