Alpha Algorithm

This is High School Math, Give Us 2 Minutes to Explain

The Alpha Algorithm offers a community governed approach of distributing tokens to the best analysts. The Alpha Algo seeks to reward exceptional relative performers, relying on Z scores:

$$Z(x₁) = (x₁ – mean) / Standard Deviation$$

To calculate an individual reward for an Analyst use the following:

$$Performance₁ = (Z(Monthly Return₁) + Z(SharpeRatio₁))/2$$

$$IdeaQuality₁ = (-Z(Gross Leverage₁) , Z(Position Hit Rate₁) +Z(Position Slugging₁) + Z(Total Positions₁))/4$$

$$Pure Alpha Score₁ = (Idea Quality₁ + Performance₁) / ∑(IdeaQuality >0 + Performance > 0)$$

$$StakeModifier₁ = (PercentRank(Stake₁)-0.5)*PureAlphaScore₁$$

$$Payout₁ = ( PureAlphaScore₁ + StakeModifier₁ ) / ∑(PureAlphaScore + StakeModifier) * 2,500$$

Payouts are always positive, we do no subtract tokens for poor performance.

More detail on How we Measure Each Metric

Gross Leverage = Max of 1 OR the average portfolio leverage since beginning of the month) Position Hit Rate = ∑(unique profitable active positions in the month) / ∑(unique active positions in the month) Position Slugging = ∑(profits on profitable positions in the month > 1% in absolute size) / ∑(losses on losing positions in the month > 1% in absolute size) Total Positions = ∑(unique active monthly positions > 1% in absolute size) Total Return = total percent return in the month Sharpe Ratio = Sharpe ratio average over the last 30 days Monthly Supply = Monthly CVY tokens issued to Analysts

Illustrative example (Using Logistic Function to Simulate Staking):

https://docs.google.com/spreadsheets/d/1zFmFQTux8zqlEWxRrjibFxKcFUTya6aYdfco7FMVYpo/edit?usp=sharing

Additional Controls

The Alpha Algorithm is governed by the community. The Alpha Algorithm will determine which Analysts and assets to be considered for monthly rewards. It will slash Staked CVY and return it to the reward pool for Analysts that go bankrupt in the month. Bankruptcy is defined as a portfolio value less than zero at any point in the month. It uses price data from Oracles (Covey at first) to calculate the performance of Analysts’ ideas posted on the public ledger and rewards top analysts with tokens. An Oracle collects pricing data and combines it with Analyst idea data to produce the metrics that feed into the Alpha Algorithm. Covey will be the first Oracle and will be replaced later by decentralized Oracles who will be rewarded for their work.