Observations
First release in preparation — no figures published yet.
Short answer
No observations have been published. This page shows the exact tables the first release will fill, with their column headers and no rows, so that the format can be checked in advance. Figures will appear here only after collection, double coding and the three reviews described in the protocol.
Release tables
Each table below is a template. The header row is final; the body is empty because there is no approved data. When the first release is published, each table will be filled with one row per indicator and will be dated with its collection window.
| Indicator | Base | Records in scope | Records with a value | Median | 25th percentile | 75th percentile | Not stated (%) | Unclear (%) |
|---|
| Indicator | Category | Records in scope | Records with a value | Share of records with a value (%) | Not stated (%) | Unclear (%) |
|---|
| Field | Records coded | Arbitrated (%) | Arbitrated: ambiguous wording (%) | Arbitrated: clause missed (%) | Arbitrated: unit error (%) | Coded as unclear (%) |
|---|
| Version | Change | Indicators affected | Reason |
|---|
How to read these tables once they are filled
- Read the base before the value. In Table 1, a wagering multiplier on a bonus-only base and one on a deposit-plus-bonus base describe different burdens. They appear on separate rows and must not be compared directly.
- Check the counts and the missing shares. A median will be printed only when the number of records with a value meets the minimum set in the protocol. The "not stated" and "unclear" shares are percentages of the records in scope, which exclude records where the clause cannot apply. A high "not stated" share means the median describes only the terms that mention the clause.
- Use the coding-quality table. Arbitration caused by ambiguous wording says the clause is hard to read; arbitration caused by a missed clause or a unit error says the coding needed correcting. A field with a high share of either deserves more care.
- Follow revisions. Corrections will never overwrite a figure silently. Table 4 will record what changed and why.
Reusable file
With the first release, the same figures will be offered as a CSV file whose columns are defined in the data dictionary, with the version and collection window on every row. The terms of reuse for that file will be stated when it is published. No file is available yet.
Publication steps
The steps below are in order. They carry no dates: a date will be added only once the step is complete.
- Protocol and data dictionary published. Done: they are on this site, open to comment before collection.
- Scope confirmed. Territorial scope, minimum sample and the regulatory question listed in the protocol are settled. Pending.
- Collection window. Terms are captured and saved for the window announced at the start of collection. Pending.
- Double coding and reconciliation. Two independent codings, arbitration and automated checks. Pending.
- Statistical, rights and regulatory reviews. The three checks described in the protocol. Pending.
- First release. Tables 1 to 4 filled, CSV published, collection window and version printed on every figure. Pending.
Until the last step is complete, any casino bonus statistic you see attributed to this site is not from this site.