Proof of Yield v1.0

[micahalcorn]

After #48 comes the third phase of PoY: yield source attribution. This is the bottom section of the daily detail page shown at https://www.figma.com/file/eN26ytHYvZ7bhCaHy5K0iv/OUSD-Proof-of-Yield?type=design&node-id=487-50955&t=wzlJxv23ETCqYGj1-4.

Proof of Yield consists of two categories of information:

1. Yield distribution
2. Yield generation

The former is covered by #47 and #48 and is calculated post-Dripper. It's all about what happens after each rebase event and organizes this information by calendar days.

The latter depicts what happens pre-Dripper, namely how each strategy has performed in addition to redeem fees, etc. It's incredibly complicated to track the yield from when it is generated to when it is dripped out since the Dripper commingles funds that were generated at different points in time, mixes them together, and distributes them over indeterminate periods of time. Because of this complexity, we decide to focus just on the generation of the yield prior to entering the Dripper. However, this has UX implications for the PoY pages. When looking at a PoY detail page for a particular day, the top portion depicts the yield that was distributed on that day and the bottom portion shows the yield that was generated on that day. Users will not be able to draw connections in the UI from yield being generated and then being distributed. The plan is to test this out and see if users find value in PoY and demand more thorough accounting. I'm fairly confident that we're already going above and beyond what is necessary.

Related: OriginProtocol/origin-dollar#1234

[micahalcorn]

Some open questions from last week:

• Should strategy performance be measured at the time of rebase or a time-weighted average during the period?
• Is it important for strategy performance to match the advertised APY on the homepage?
• Should the APY on each yield source be raw or boosted? If boosted, then opt-in/out events would need to be available in analytics or we would need to move away from the Python stack.

Decisions:

• Time-weighted calculations would be preferable to account for allocation changes within a particular day. For example, if 90% of funds are in Convex and then reallocated to Morpho at 11:00 pm, then the strategy performance would be underreported for Morpho and overreported for Convex on that day.
• There's a disconnect between advertised APY and strategy performance for two reasons. First, the advertised APY is a moving average, although it will eventually become a "current" APY once we have full Dripper utilization. Second, strategy performance is measured pre-Dripper. It's very challenging to map distributed yield (actual APY realized by users) to the generation of the yield because it enters the Dripper in a measurable way but is ambiguous when dripped. We agreed to avoid this scope creep when designing Proof of Yield. Thus, it's not necessary for advertised APY to match the yield sources.
• Proof of Yield is all about transparency. One impetus for this project was the constant question we receive about how our yield is higher than the underlying strategies. To keep things simple and err on the side of transparency, let's display raw APY on the yield sources. This makes it easier to compare expected performance with actual results.