What the data can tell you
GaugeEdge is a research project developing an allocation tool for veAERO voters on Aerodrome, on Base. The current site contains historical observations and an early-access list. It does not provide a live allocation service.
Data and dates
Voting weights and reward quantities are read through the Sugar SDK. Epochs begin Thursday at 00:00 UTC. Our public charts label the epoch end, seven days after the SDK's epoch-start timestamp. We wait at least three hours after that end before treating a retrieved epoch as settled.
The chart stamp shows the latest included epoch. The page marks the data overdue when another epoch should be available. It is a historical sample, not a live voting feed. Download the dated summary and source-file checksum (JSON) or the latest displayed pool table (CSV).
Nominal value and coverage
Nominal USD reward equals the SDK value of trading fees plus voting incentives. Reward per 1,000 votes is that total divided by recorded votes, multiplied by 1,000. USD amounts use prices recorded during retrieval; those prices can differ from voting or payout prices. Older research files do not record the precise pricing time.
These are not proceeds from selling the rewards. Public charts do not apply liquidity discounts, transaction costs or taxes. Unknown token prices may appear as zero. The position calculator adds the selected voting weight to recorded votes to show dilution. It does not model removing an existing vote from another pool.
The chart includes pools with at least 0.1% of recorded epoch votes. The top-to-median ratio uses that filtered sample. The vote-weighted reference uses all collected positive-vote pools. The median ratio across epochs gives each epoch equal weight. Coverage follows the pools collected and can omit pools that were active only in the past; it is not a census of the entire protocol.
Why capture before and after voting closes?
Ordinary voting closes Wednesday at 23:00 UTC, an hour before the epoch end. The protocol allows approved NFTs to vote in the restricted window; see the Aerodrome contract specification.
| Observation | UTC schedule | Purpose |
|---|---|---|
| Early observations | Monday, Tuesday, Wednesday 00:00 | Track the week |
| 60 minutes before ordinary cutoff | Wednesday 22:00 | Decision-time information |
| 15 minutes before ordinary cutoff | Wednesday 22:45 | Decision-time information with an execution buffer |
| After ordinary cutoff | Wednesday 23:30 | Research reference; further changes may occur |
| Settled retrieval | Thursday 04:00 | Outcome used to evaluate earlier decisions |
Actual capture completion controls whether an observation was still usable. A missed scheduled capture cannot be recreated later. We retain post-cutoff observations to learn what changed after users had to decide.
How we test a prediction
Both the observed-data baseline and the experimental prediction choose using information available at the same observation time. Both use the same position size, pool cap and assumed costs. Earlier settled epochs supply training data; later epochs supply held-out comparisons. Missing final pool data is marked unknown rather than dropped from the decision universe.
The first experimental model predicts vote growth from prior epochs at similar times. It holds displayed rewards fixed. A perfect-information result is a hindsight reference only. Legacy price and timing limitations remain, and we have not established a reliable, executable prediction advantage.