Me
Based on these criteria: Names the doc(s)you drew from and clearly cite them.
Preserves visuals: includes at least one relevant image/screenshot from the doc.
Preserves links: any in-page or cross-doc links are present and correctly formatted.
Lists key fields with short explanations, defaults, and any validation rules (required/optional).
Mentions limits (size, file type, counts, rate limits) if the feature has them.
State what’s not covered if the needed info isn’t in the KB page, without referencing “chunks”.
Formatting: clear headings/bullets/tables; no jargon unexplained; no hallucinations. Explain how A/B experiments in Hawk, GS and Harmony vary
Enki
# A/B Experiments Across Hawk, Game Server, and Harmony
Based on the available documentation, I can provide detailed information about A/B Experiments in **Game Server (GS)**, with references to **Hawk**. However, information about A/B Experiments specifically in **Harmony** is not currently available in the knowledge base.
---
## Game Server (GS) A/B Experiments
**Source**: [Setting up A/B Experiments](https://tripledotstudios.atlassian.net/wiki/spaces/KB/pages/2581889051/Setting+up+A+B+Experiments)
### Creating an Experiment
You can set up A/B Experiments in two ways:
1. Navigate to the A/B Experiments section and click the **green [+New] button**
2. Create directly from the entity you'd like to test using the **[New Experiment]** button
### Key Configuration Fields
| Field | Required | Options/Type | Description |
|-------|----------|--------------|-------------|
| **Name** | Yes | Text | The name of your experiment |
| **Status** | Yes | Active, Inactive, Allocation Paused, Archived | Controls experiment lifecycle (see Status Details below) |
| **Audience Type** | Yes | All, New, Old | **New**: Only users who install after experiment `start_date`<br>**Old**: Only users who installed before `start_date`<br>**All**: All users regardless of `install_date` |
| **Priority** | Yes | Numeric | Higher numbers = higher priority (e.g., 100 > 1). Determines which configuration is sent when a user matches multiple experiments |
| **Use Local Time** | Yes | Checkbox | Uses player device local time instead of UTC for all experiment dates |
| **Start At** | Yes | Date/Time | Experiment start date & time (UTC by default) |
| **Allocation Start** | Yes | Date/Time | When the experiment begins allocating users (same as or after Start At) |
| **End At** | No | Date/Time | Experiment end date & time |
| **Allocation End** | No | Date/Time | When allocation stops. System checks every 5 minutes and automatically changes Active experiments to "Allocation Paused" |
| **Maximum Audience Size** | No | Numeric | Approximate limit for allocated users. Leave empty for no limit |
| **Availability** | Yes | Test, Live | **Test**: Only affects registered test devices<br>**Live**: Affects live players |
| **Labels** | No | Dropdown | Categorization tags for organizing experiments |
| **Include in Lab Calculations** | No | Checkbox (default: checked) | Determines if experiment data is included in Lab Calculations. **Grace period**: 5 calendar days to enable if initially unchecked |
| **Notes** | No | Text | Comments, links, or relevant information |
### Status Options Explained
- **Active**: Experiment is live and affects users
- **Inactive**: Experiment is inactive and doesn't affect users. Participations still counted in "Experiment variant" rule
- **Allocation Paused**: Live for existing participants, but no new users allocated. Automatically set when Allocation End time is reached
- **Archived**: Experiment is finished and won't be referenced. Not editable. Can be reversed within **7-day grace period**
- **Permanently Archived**: After 7 days, status becomes irreversible. Participations cleared from GS database (available only in RedShift for analytics)
### Experiment Variants
Variants divide your audience into test groups with different configurations.
| Attribute | Description |
|-----------|-------------|
| **Name** | Variant name. First variant is always "Control" (unchangeable) |
| **Id** | Read-only, set by Game Server. Used in analytics to track specific variant participations |
| **Weight** | Defines user distribution. Example: Variant #1 (weight=1) and Variant #2 (weight=3) = 25%/75% split |
| **Skip Experiment for this variant** | When checked, users in this variant won't receive the A/B experiment |
| **Entities list** | Tested entities with values for the variant (only available after experiment creation) |
**Important**: Users are randomly assigned to variants within the specified weights.
### Priority System
A/B Experiments and overrides are **not exclusive**. Priority order:
1. **A/B Experiments** (highest priority)
2. **Overrides**
3. **Base Config** (Game Settings)
Users can receive settings from both experiments and overrides if they apply to different settings.
### Usage Tracking
The "Usage" counter displays how many times experiment data has been sent to users within a 24-hour period, resetting daily at midnight UTC.
### Limits & Validation
- **Maximum Audience Size**: Numbers are approximate. System stops allocation when limit reached OR when Allocation End time is reached (whichever comes first)
- **Grace Period for Archiving**: 7 calendar days to reverse Archived status
- **Grace Period for Lab Calculations**: 5 calendar days to enable after experiment creation
- **Allocation Check Frequency**: Every 5 minutes
- **Entities**: Can only be added after experiment creation
---
## Hawk A/B Experiments
**Source**: [Hawk Guides](https://tripledotstudios.atlassian.net/wiki/spaces/KB/pages/2916679818)
Hawk has its own A/B Experiments functionality documented separately. For detailed information about Hawk-specific A/B experiments:
**📚 See**: [A/B Experiments in Hawk](https://tripledotstudios.atlassian.net/wiki/spaces/KB/pages/3319988247/A+B+Experiments+in+Hawk)
---
## Harmony A/B Experiments
**What's Not Covered**: The knowledge base does not currently contain specific documentation about A/B Experiments in Harmony. If you need information about Harmony's experimentation capabilities, you may need to consult additional documentation or reach out to your team for guidance.
---
## Related Resources
- [Game Settings Configuration](https://tripledotstudios.atlassian.net/wiki/spaces/KB/pages/2580021328/Game+Settings+Configuration) - Understanding base configuration that experiments override
- [Hawk Guides Home](https://tripledotstudios.atlassian.net/wiki/spaces/KB/pages/2916679818) - Complete Hawk documentation
- [Knowledge Base Home](https://tripledotstudios.atlassian.net/wiki/spaces/KB/pages/2579759344) - Browse all topics
For external users without Tripledot email access, visit the [Public Knowledge Base](https://tripledotstudios.atlassian.net/wiki/x/hYEm3Q).