Pre-registration — the cornerstone of verifiable science

A pre-registration is a written commitment — locked in BEFORE you collect data — that says: this is my hypothesis, this is exactly how I'll test it, and this is what would count as a "positive" or "negative" outcome. Discovery makes this a tamper-proof PDF anyone can verify.

**Why this matters:**
Without pre-registration, you can analyze the data 17 different ways, find the one that crosses p<0.05, and report that one. That's p-hacking. Pre-registration prevents it by recording your analysis plan before the data exists.

**The five mandatory sections:**
1. **Hypothesis** — the specific, falsifiable claim. "If X, then Y, because Z."
2. **Background** — what prior work informs this. Cite negative results too — they're often the most useful prior knowledge.
3. **Methodology** — materials, sample size, procedure. Specific enough that someone else could repeat it.
4. **Analysis Plan** — exactly which statistical test on which data, with what success threshold. Locking this is what prevents p-hacking later.
5. **Success Criteria** — what outcome would you call positive vs negative? Drift from this at result time is auto-flagged.

You'll also pick:
- **Predicted Outcome** — honest expectation, not a goal (positive / negative / inconclusive).
- **Estimated Completion** — optional informational date.
- **Sensor Thresholds** — optional min/max envelopes for any bound sensor. The anomaly watcher pings you if a reading drifts past these.

**How locking works:**
1. You click Lock Pre-Registration. (At least 20 chars per section is required — this is the only friction we add, because skipping is a footgun.)
2. Turtini renders a portrait-letter PDF with all the above + your org / researcher attribution.
3. The PDF gets stamped with a Verified-by-Turtini QR code that resolves to /verify/{id}.
4. The PDF's SHA-256 is registered in pdfVerifications and becomes the **root** of the experiment's hash chain.
5. The experiment status flips from draft → active. A permanent DOI is minted. The pre-reg becomes public at turtini.com/discovery/exp/{doi}.

**After locking — what you can and can't do:**
- ✓ Append daily sessions (hash-chained, ordered, signed by the chain head).
- ✓ Bind sensors. Upload datasets. Publish a Result. Run reproductions. Invite peer reviews.
- ✗ Edit the locked pre-reg. That's intentional. If your plan changed substantively, that drift will be auto-shown next to your published Result; you provide a written rationale, reviewers see both versions.
- ✗ Delete the locked PDF or its registry entry. The chain has to survive — that's what makes it verifiable.

**Authority — who can lock:**
- The experiment owner always can.
- In org context, an org owner / admin / editor can lock on the researcher's behalf (e.g. a PI overseeing a grad student).

This is intentionally a careful, one-way door. Take the time to draft it well.