01
Every figure has a source
Every country field carries a provenance mark, every refreshed value its source and year, every live source its name and the time it answered, and every analyst row the file, feed or equation it came from.
For organisations
Built to be trusted with a decision
A decision made on a model has to survive the question of where the model's numbers came from. This one is built so that question always has an answer on the page: every figure sourced, every answer worked, every run repeatable, and the organisation's own data kept where it can see it.
What an organisation can rely on
02
Every answer has its working
The analyst lists its plan, its evidence and its conclusions with their citations and confidence, and the written up version is checked sentence by sentence against the evidence on the server.
03
Every run can be repeated
A scenario is saved with its sentences, its seed, its number of futures and its focus. The same seed plays the same futures, so a result shown in a meeting can be replayed exactly afterwards.
04
An audit trail you can keep
Analyses kept against the account, each with its question, plan, evidence and conclusions as they stood when asked; every analysis and the company reading as a report for print or PDF; the evidence table as CSV, the whole analysis as JSON, every table on every tab as CSV, and the briefing with the source of every sentence in the Working tab.
05
Ranges, not a single number
The company is priced in every future the run played, so a board sees the spread, the share of futures that lose and the result in each way a decision could go, with what was kept at the median named.
06
Tested against the record
The model is started from a past year with that year's figures and scored against what the World Bank recorded after it, and the result is printed whichever way it comes out, with what flatters it named.
07
Stress tested on your own files
The shocks your suppliers, routes and markets are most exposed to, each run alone and priced line by line, ranked by the worst year.
08
One world for the team
Analyses shared with every member, saved scenarios shared with a switch, the organisation's company files published by an administrator, and one set of world figures for everyone.
09
Signs to watch
Every way a scenario could go comes with the signs in the world that would show it happening, and what the live desk reads for the parties today, so a plan can be revisited when the signs appear rather than on a calendar.
10
Your data stays yours
Company files are kept against the account that uploaded them or in that browser only, and are sent nowhere else. When a written answer is asked for, the model provider is sent the question and the evidence rows it is written from, which can include lines about the company; never the files themselves.
11
Access set by role
Whether visitors must sign in to run it, who may save, and who may spend the model key (administrators, members or everyone) are settings. World figures from the app are for administrators only.
12
Spend has a ceiling
The model key lives on the server and is never sent to a browser. A monthly allowance caps what it can spend, and the count is shown to administrators.
13
It runs where people are
The simulation runs in the browser: no queue, no credits, no upload of what anyone is thinking about. When a host blocks the WordPress data routes, the app takes a second road through admin-ajax on its own.
14
Readable by everyone
Controls carry visible labels tied to them, the text in every theme clears 4.5 to 1 against what it sits on, and every motion stops under reduced motion. The accessibility page says what is done and what is not yet.
Where each thing is kept
| What | Where | Who can see it |
|---|---|---|
| The world record | The WordPress database, over the curated dataset shipped with the plugin | Everyone who may run the simulator |
| World figures you supply | Beside the record, marked as yours, removable in one step | Everyone who may run it; only administrators may change them |
| Company files | The uploading account, or that browser only | That account, or that browser |
| Kept analyses | The account that kept them (fifty at most, a megabyte in all), or that browser only | That account, or that browser |
| Analyses shared with the team | The site, a hundred at most | Everyone allowed to save; taken down by whoever shared it or an administrator |
| The organisation's company files | The site, published by an administrator | Everyone allowed to save; changed only by administrators |
| The record's history, for the backtest | The site, fetched from the World Bank and kept thirty days | Everyone who may run the simulator |
| Saved scenarios | The site, against the account that saved them | That account and administrators, unless it is marked public |
| The model key | The site's options, server side only; the settings screen shows its last four characters | Nobody, in full |
| A run | The browser that ran it; a saved run keeps a summary only | The person who ran it |
What this page does not claim
No certification
Nothing here is certified by anybody, and nothing on this site says otherwise. What it offers instead is the ability to check every number yourself.
No customer list
There are no logos, quotes or case studies on this site, because none would be honest to print yet.
No prediction
A run is a distribution with a stated reliability and the coefficients it is most sensitive to. The analyst says how sure it is and what would change its answer.
Talk to us about your deployment
Who should run it, who should save, where the model key should sit, which world figures you want to supply, and what the data room should hold for your organisation.