← Back to Games

How scoring works: Connections

Your game score comes from your own puzzle result. Your party rank compares that score with your party's submissions for the same day.

Quick summary

Fewer mistakes improve your game score in Connections. Your party rank compares your game score with other party submissions for the same game/day and can change as new submissions arrive.

How to score well

  • Only submit a group when you are confident it fits all four words.
  • Reducing mistakes improves your normalized score.
  • An unfinished board scores lower than a completed board.

How ranking works in parties

  • Your rank compares your game score to other party submissions for the same game/day.
  • Ranks recalculate when new submissions arrive.
  • Ties use competition ranking (1, 1, 3).

Example

If you finish with 1 mistake, your score is strong, but your rank can still change if others finish with 0 mistakes.

Technical scoring data 
{
    "ranking": {
        "tie_algorithm": {
            "key": "competition_ranking",
            "display_name": "Competition ranking"
        }
    },
    "examples": [
        "Review this games examples in PR for readability and correctness."
    ],
    "rule_text": "Lower raw scores produce better normalized scores.",
    "edge_cases": [
        "Review edge-case copy in PR to ensure behavior is understandable."
    ],
    "measurement": {
        "metric_key": "mistake_count",
        "metric_label": "Mistakes",
        "scoring_type": "lower_better"
    },
    "human_review": {
        "checklist": [
            "Verify rule_text reflects the scoring behavior without ambiguity.",
            "Verify examples are readable and match expected outcomes.",
            "Verify edge_cases are clear and non-conflicting."
        ],
        "machine_verifiable_scope": [
            "measurement.scoring_type",
            "absolute_scoring.method",
            "absolute_scoring.solved_floor",
            "absolute_scoring.unsolved_score",
            "absolute_scoring.range.best",
            "absolute_scoring.range.worst_solved",
            "absolute_scoring.brackets",
            "ranking.tie_algorithm.key",
            "party_ranking.normalization_summary"
        ]
    },
    "party_ranking": {
        "normalization_summary": "Party ranking uses peer min/max normalization so each solved score is compared to that party's best and worst solved results for the day."
    },
    "schema_version": 1,
    "absolute_scoring": {
        "range": {
            "best": 1,
            "worst_solved": 0.2
        },
        "method": "lower_better",
        "scoring_type": "lower_better",
        "solved_floor": 0.2,
        "unsolved_score": 0
    }
}