AI Reward Mechanism
Walrus Land introduces a dynamic and performance-driven reward model powered by AI. Instead of distributing static or time-based rewards, the system evaluates player actions and outcomes to determine fair and scalable compensation in $WLAND.
This ensures that players are incentivized not only to participate, but to engage meaningfully and strategically.
Reward Evaluation Criteria
After each PvE battle, the AI Reward Engine analyzes multiple dimensions of performance to calculate an appropriate reward output:
Difficulty Score
Assesses the strength of the opponent relative to the Walrus's level
Strategy Novelty
Rewards creative or non-repetitive approaches in combat behavior
Consistency Bonus
Factors in historical win/loss trends and behavioral improvement over time
Efficiency Rating
Measures speed and damage taken during combat to evaluate tactical quality
Reward Calculation Formula
The total reward RRR for a single PvE battle is computed as:
Where:
nnn: Number of combat phases or rounds in the battle
BiB_iBi: Base reward for phase iii (based on enemy tier)
DiD_iDi: Difficulty ratio (enemy power ÷ Walrus power)
α\alphaα: Difficulty exponent (amplifies higher challenge tiers)
EiE_iEi: Efficiency score (function of HP remaining × time score)
β\betaβ: Efficiency exponent (adjusts reward scaling curve)
CiC_iCi: Strategy reuse count (how often the same tactic was used in recent history)
NiN_iNi: Novelty boost (0 ≤ NNN ≤ 1), measured via vector distance in strategic behavior
Why AI-Based Rewards?
Promotes strategic experimentation, not repetitive farming
Encourages consistent improvement and adaptation
Adds a layer of fairness by rewarding effort and originality
Prevents botting and low-effort grinding
This mechanism ensures that earning $WLAND is not only fair — it’s intelligent, skill-aligned, and resistant to abuse.
Last updated