Privacy
Theory is an open, public forecasting service. This page explains what data we process and why — written to be read, not to obscure.
- Everything you submit as a forecaster — your predictions, reasoning, and the pseudonymous id they're attributed to — is public, permanent, and released as open data (CC0).
- We use no cookies, no analytics, and no trackers.
- We never publish your email address or real name.
Who we are
Theory (theory.encounter.bio) is operated by Raimo van der Klein (Encounter). For any privacy or data request, contact raimovanderklein@gmail.com.
What we process
Authentication. To submit forecasts you sign in through WorkOS AuthKit (via Google or a one-time email code). WorkOS handles the sign-in and holds your account details, such as your email. Theory receives only an opaque, stable identifier and uses it as your public forecaster id (for example workos:…). Your email and name are never shown publicly or stored in the public corpus.
Your contributions. The forecasts ("theories") you submit — the probability, mechanism, reasoning, evidence references, and any provenance you include — are stored and shown publicly, attributed to your pseudonymous id or to a display handle you choose. They are immutable: once submitted, a forecast is timestamped and is not edited or deleted, because the integrity of a forecasting record depends on it.
Server logs. Like any website, our host processes basic request metadata (IP address, user agent, timestamps) transiently to operate and secure the service.
Why — legal basis
We process this data to run the service and to maintain an attributable, tamper-evident public scientific record — our legitimate interest (GDPR Art. 6(1)(f)), and, for forecasters, to provide the service you signed up for. We do not sell data and do not use it for advertising.
Who processes data for us
- WorkOS — authentication and identity.
- Render — hosting and the database.
- Google — only when you choose "Sign in with Google".
- An RFC-3161 timestamping authority — receives only a hash to timestamp forecasts; no personal data.
- Revolut — only if you choose to donate; we never see your payment details.
Cookies & tracking
Theory sets no cookies and runs no analytics or trackers. The public site loads one web font (Silkscreen) from Google Fonts, which means Google receives your IP address when the font loads; no cookie is set.
Retention
Forecasts are permanent by design — they form an immutable public record. Server logs are short-lived. Your authentication account persists with WorkOS for as long as you use the service.
Your rights
You can request access to, correction of, or deletion of your personal data, and object to processing, by contacting us. One honest limit: because forecasts are immutable, timestamped, public-interest records released as open data, we may not be able to fully erase an individual submission (GDPR Art. 17(3) — archiving and scientific or public-interest processing). We can disassociate or rotate your public pseudonym and delete your WorkOS account.
A note to forecasters
Treat everything you submit as public and permanent. Do not put personal information — names, emails, internal hostnames — into free-text fields such as reasoning or provenance, as it becomes part of the open dataset.
Changes
We update this page as the service evolves; the date above reflects the latest version.
This statement describes Theory's current practices in good faith. It is not legal advice.