Beyond staking - How fee revenue impacts topic rewards in Allora Before you start, please read this note: This is my personal interpretation based on available materials and observations. While I’ve tried to keep things accurate, some details might not fully reflect the latest network mechanics. Always DYOR and cross-check with official docs if you’re building on Allora. Continue from now. TL;DR If you’re building or contributing to a topic in Allora, here’s what to remember • Stake matters, but so does real usage (aka fee revenue). • The system is designed to reward both capital (stake) and utility (demand). • Long-term consistent activity is better than short-term spikes. • Want more rewards? Grow usage, not just TVL. ➥ So, in Allora, staking isn’t the only game that matters. If you're wondering why some topics get rewarded more than others, it's not just about how much stake they attract — it's also about how much fee revenue they generate. Let’s break that down in a chill, clear way 👇 🧠 What determines a topic’s weight? Every topic in Allora gets a score called its weight — think of it like your topic’s "influence score." And that weight directly affects how much reward it gets. Here’s what goes into that weight: 1. Stake (S): Obvious one. More stake = more weight. 2. Fee Revenue (P/C): Topics that generate more prediction fees get a higher score. Why? Because it means people are actually using it. 3. μ and ν: These are network-wide constants that balance how much stake vs. fees matter. Default = 0.5 each. So 50/50. 🔁 Smooth moves: EMA smoothing Weights don’t jump all over the place. Allora uses Exponential Moving Average (EMA) to smooth out the weight over time. WHY? Because some spikes are just noise — we want long-term, sustained performance to matter more. The smoothing factor, alpha (α), adjusts based on how long the topic’s epoch is. So even if different topics run at different frequencies, they’re treated fairly. 🎯 How does this affect rewards? This is where it gets interesting. The weight of a topic = how much reward it gets. Straight up. • Proportional Rewards: Total reward is split among active topics proportionally. So: ➽ reward_share = topic_weight / total_active_weights • Active Topics Only: Only the top N weighted topics are active each round. If your topic isn’t active — no reward. • Activation Thresholds: Topics below a minimum weight are deactivated. To reactivate? Bring more stake or generate more fees. • Nonce Updates: For active topics, the system sets up the next round: updates the epoch timestamp and enables workers/reputers to submit data and earn rewards. Need help activating your topic? Or curious why yours dropped off the reward list? Might be time to check those weights. 👀