Understanding Predictive Allocation
Predictive Allocation is Aero's mechanism for directing AERO Rewards and earning exchange revenue. It runs continuously on every network Aero supports and replaces the weekly voting cycle.
Rather than relying on historical data to decide where rewards go, Token Operators direct their Allocation Power to the pools where they believe liquidity is needed. Liquidity providers follow AERO Rewards to those pools, and the Token Operators allocated to a pool receive that pool's trading fees, MEV capture, and incentives (exchange revenue). The system gives individual Token Operators more agency to leverage their edge, and collectively improves the efficiency of the protocol.
How Predictive Allocation works
Liquidity needs to be deep where trading actually happens, and the users providing that liquidity need to be rewarded. Providing liquidity is an act of anticipation: liquidity providers, like market makers everywhere, earn by predicting where trading activity will land and positioning capital there ahead of time.
Token Operators are active participants. When a Token Operator directs rewards to a pool, they are predicting that pool will generate trading volume and therefore flow revenue back to their position. Together, all those allocations create a market signal that draws productive liquidity where it is needed.
An allocation takes effect as soon as it is submitted. Exchange revenue accrues continuously, keeps accruing during the cooldown, and can be claimed at any time.
Understanding volatility in prediction
Predictions naturally have variance, especially over shorter windows. Individual Token Operators can expect variable performance versus other participants since each Token Operator creates their allocation strategy independent of the market.
On average, Token Operators are staked for 3+ years, a substantial amount of time over which to earn exchange revenue. To best understand the performance of their predictions, allocators should track their average earnings over a longer period of time rather than 47-hour performance.
Using Markets data and research
Each Token Operator creates their allocation strategy independently, using the tools at their disposal. The Markets tab shows historical fees, incentives that will be distributed, and Stake APR based on estimated returns.
Although not a predictor of future results, research showed that a strategy looking at the prior 48 hours of pool activity, and more heavily weighting recent hours, was more successful than other strategies on average. In addition, if a Token Operator does not want to attempt to predict future market catalysts, they could use the data available in the Markets tab to determine which pools have performed well recently in an effort to predict which pools could generate exchange revenue in the future.
Using Autopilot
For Token Operators who do not want to allocate on their own, an Autopilot is an automated management tool for sAERO. It directs Allocation Power, claims exchange revenue as it accrues, and either compounds that revenue into sAERO (a Compounder) or converts it to a target token (a Converter). Aero launches with two Autopilots:
- AERO Maxi Compounding: a Compounder. It compounds claimed revenue into the position's sAERO.
- USDC Maxi Autopilot: a Converter. It converts claimed revenue to USDC.
Each sAERO position can use one Autopilot at a time. Activate an Autopilot when staking, or from the Autopilot tab on the Dashboard for an existing position. To return to manual allocation, deactivate the Autopilot.
The first available Autopilots will execute a bluechip only strategy that allocates to bluechip asset pairs such as ETH, BTC, USDC and USDT across chains.
A performance fee may apply. Autopilot APRs are a guide to recent performance, not a guarantee of future returns. Aero will roll out new Autopilots over time.
Predictive Allocation compared with weekly voting
Weekly voting rewards participation, but not foresight. Under other weekly vote models, rewards for the week ahead are set by a vote using results from the prior week. Because the vote reacts to past performance rather than signalling where trading happens next, rewards can dislocate from the fees pools actually earn.
Predictive Allocation makes directing rewards continuous. Token Operators choose which markets the exchange should support with AERO Rewards, liquidity providers follow, and the Token Operators keep a proportional share of the revenue those markets produce. Backing a market early, and being right, likely earns the largest share of that revenue when it accrues. And because the prediction is an allocation, not just a bet, it helps bring about the outcome it predicts, moving liquidity ahead of demand.
Research on historical pool activity shows how much more effective the simple 48-hour strategy described above would have been than weekly voting. These are the results of that one simple strategy, not the full potential of the system.
Accurately tracking against fees
Across 78 weeks on 27 pools, rewards tracked fees at 0.89 under weekly voting and 0.99 under Predictive Allocation, where 1.0 means the two move together exactly. Rewards that follow fees pull liquidity toward the pools that are actually earning, and those are the pools whose revenue flows back to the positions allocated to them. For liquidity providers it means the rewards on offer are a reliable read of where trading is really happening.
Stabilizing distribution shifts
Allocation error is the share of rewards that would need to move between pools to match their share of fees. Predictive Allocation cut it by up to 66%, in calm weeks (18.9% to 6.4%) and in trading surge weeks (20.5% to 7.7%). In practice this means allocation doesn't just produce a correlation with fees but accurately distributes AERO Rewards to where fees will occur.
Reducing the inefficiency in the system
Weekly voting suffers the largest dislocations after bursts of volatility, because rewards cannot move until the next epoch. In the two weeks after a large surge, pools received 72% of the fee-matched reward amount under weekly voting and 92% under Predictive Allocation.
A Token Operator who can anticipate incoming volatility shares in the fees it produces instead of arriving after it has passed. Liquidity providers get the signal to deepen a market while the volume is there, not after it has gone. Allocators that are ahead of a spike earn outsize rewards.
These figures come from simulations on past pool data, not live rewards. They are not a prediction of future performance, and they describe the system's behaviour rather than any single Token Operator's results.