Trust score formula

Visual block constructor for the platform trust score used in join restrictions and trust signals.

← Settings

Basics

  • The formula is a tree of blocks. Child blocks are evaluated first; their results flow up to the root.
  • Change a block's behavior with the type dropdown at the top of each card. Switching type resets that block's children to sensible defaults.
  • Nested blocks appear indented inside their parent. Use Remove on a child to delete it (not available on the root).

Color guide

Each block card has a colored left stripe: sky = player data, amber = score rules, violet = math, emerald = final step. Rating sources use orange (FACEIT), indigo (Premier), and green (Leetify).

Input parameters

Add a Parameter block and pick one of the runtime inputs:

FACEIT ELO
FACEIT skill rating. Included in weighted sources when value ≥ 0.
Premier rating
CS2 Premier from Leetify. Included when value ≥ 0.
Leetify score
Leetify composite rating. Included when present (any value).
CS2 hours
Steam playtime. Usually fed into an Hours multiplier block.

Block types

  • Math (Add, Subtract, Multiply, Divide, Min, Max, Clamp, Round) — combine numeric child blocks.
  • Linear normalize — scale an input to 0–output max using an input cap (e.g. ELO / 4000 × 1000). Used for FACEIT and Premier.
  • Leetify normalize — offset + range mapping (default offset 2, range 5).
  • Weighted sources — adaptive average of rating sources. Weights renormalize when a source is missing. Add/remove entries with + Add source.
  • Hours multiplier — gates the final score by playtime: 0 if hours unknown, linear ramp up to threshold, then capped at 1.

Default v1 layout

Round
└── Multiply
    ├── Weighted sources
    │   ├── FACEIT  (55%) → Linear normalize (max 4000)
    │   ├── Premier (30%) → Linear normalize (max 30000)
    │   └── Leetify (15%) → Leetify normalize
    └── Hours multiplier (threshold 400, input CS2 hours)

Use Reset to default v1 to restore this structure.

Recommended workflow

  1. Edit metadata (version, color bands) if you are shipping a new formula version.
  2. Adjust the block tree — start from the default unless you need a fully custom flow.
  3. Use the Preview panel: pick a preset or enter sample values and confirm performance score, raw score, and hours multiplier.
  4. Fix any validation errors shown in preview before saving.
  5. Click Save formula. Cached performance scores are cleared platform-wide; players get recomputed scores on next request.

Tips

  • Bump Version (e.g. trust-score-v2) when changing weights or structure so analytics can distinguish snapshots.
  • Hours threshold in metadata only affects confidence notes text; the multiplier threshold is set on the Hours multiplier block.
  • Per-user admin overrides still bypass this formula entirely.

Formula metadata

General settings — most admins only need to adjust color bands or version when publishing changes.

Version name
Display label
Hours note threshold
Cache time (seconds)
Score color bands

High score: 700 and above

Formula tree

Follow the colored cards from the inside out. Start with the default unless you know you need changes.

Color guide

Player dataMathScore rulesFinal step
FACEIT ELOPremier ratingLeetify scoreCS2 hoursDiscord connected

Each card's left stripe matches its category. Rating sources in the mix block use orange / indigo / green.

RootFinal

Rounds to a whole number — usually the last step

Value to round
Math

Multiplies two numbers

First value
Rules

Combines FACEIT / Premier / Leetify with weights

Mixes available ratings — missing sources are skipped automatically
FACEIT55% weight
Rating source
Weight (0–1)

≈ 55% when available

Count this source when
How to scale this rating
Rules

Scales a rating to 0–1000 (like FACEIT or Premier)

Rating input
Data

Reads live data about the player

FACEIT ELO

Skill rating from FACEIT

Top of scale (input)

e.g. 4000 for FACEIT

Performance score max

Usually 1000

Premier30% weight
Rating source
Weight (0–1)

≈ 30% when available

Count this source when
How to scale this rating
Rules

Scales a rating to 0–1000 (like FACEIT or Premier)

Rating input
Data

Reads live data about the player

Premier rating

CS2 Premier from Leetify

Top of scale (input)

e.g. 4000 for FACEIT

Performance score max

Usually 1000

Leetify15% weight
Rating source
Weight (0–1)

≈ 15% when available

Count this source when
How to scale this rating
Rules

Converts Leetify score to 0–1000

Leetify input
Data

Reads live data about the player

Leetify score

Leetify composite rating

Offset
Range
Performance score max
Second value
Rules

Reduces score when playtime is low or hidden

Playtime gate — score is reduced when hours are low or hidden
Playtime source
Data

Reads live data about the player

CS2 hours

Steam playtime in hours

Full trust at (hours)

Try it out

Pick a sample player profile and see the trust score your formula produces.

Trust score

1000

out of 1000

High score
Before playtime gate
1000
Playtime multiplier
100%
Used:FACEITPremierLeetify