Consensus Mechanism
Operators define the rate and set parameters for Panelists to forecast. The Treehouse Protocol then aggregates independent data points provided by the Panelists and derives the final consensus rate using the following mechanisms:
Outlier Value (OV) Removal: Predictions are organized in descending order, and a designated percentage of the largest and smallest values, known as “Outlier Values (OV)” predictions, are removed. This step prevents any outsized skewing of the final rate consensus.
Random Sampling: The remaining data submissions are randomly sampled to create the final group of submissions for the consensus calculation. This step is enforced only when the number of Panelists for the specific DOR reaches a critical threshold defined by the Operator, ensuring that the sampled population remains statistically representative of the total population.
Consensus Mean: The mean of the remaining sampled submissions is used to determine the final consensus rate on the DOR. This process effectively acts as a trimmed mean from the original submissions, providing a statistically reliable metric for Referencers.
Out-of-Range (OR) Setting: Throughout the observation period, the Treehouse Protocol records the realized rate, which is then used to compute the “Out-of-Range (OR)” boundary, a designated standard deviation away from the average realized rate over the observation period.
Payout and Penalty: At the end of the observation period, Panelists and Delegators whose original submission falls outside of the OR boundary will be penalized according to the protocol’s Slashing Mechanism. Meanwhile, the rest of the Panelists and Delegators are rewarded as per the protocol’s Payout Mechanism.
Last updated
