Liquid Alpha

Think of Liquid Alpha as a reward system that encourages validators to discover good miners early. Each miner receives its own alpha value. This value changes based on how many validators agree about the miner's performance. In the past, all bonds changed at a fixed rate of 0.9. Now, these rates adjust dynamically. By default, they range between 0.7 and 0.9, though subnet owners can customize these values within certain limits. The adjustment depends on validator agreement about a miner's work. This creates an advantage for validators who actively evaluate miners.

Bonds track the history between validators and miners. They work like trust scores that grow stronger over time. Every validator-miner pair has its own bond value. This value combines new ratings from current evaluations, previous performance history, and how well these ratings match network consensus .

Strong bonds lead to higher rewards. When validators consistently rate miners well, their bonds grow stronger. These bonds determine how much TAO validators earn from each miner.

Alpha controls bond growth speed. Think of it like a speed limit for trust-building. It works in two ways:

• High alpha (default 0.9) creates slow changes - with lower impact from new weights
• Low alpha (default 0.7) allows fast changes - with higher impact from new weights

Each miner gets its own alpha value. This value changes based on network consensus. Different speeds for different situations help reward those who find good miners first.

Liquid Alpha creates a prediction game. Validators must guess which miners will perform well. Finding good miners early brings three benefits:

• Quick bond growth through lower alpha values
• Stronger bonds before others notice the miner
• Higher rewards while copiers try to catch up

This system pushes validators to actively search for good miners instead of copying other's weights.

The system dynamically adjusts alpha values based on miner agreement. When few validators rate a miner, lower alpha allows faster bond growth. Conversely, when many validators agree, higher alpha creates slower bond growth. This approach makes early discovery profitable and makes copying expensive after consensus forms.

Liquid Alpha works alongside commit-reveal to create a powerful defense against copying. Commit-reveal forces copiers to use outdated information by hiding recent weights. While liquid alpha triggers higher alpha values with higher consensus, which slows bond growth for those attempting to copy weights .

This double protection ensures that honest validation remains more profitable than copying strategies.