Massively Inaccurate APYs

I’m a fan but there is a massive difference between the APYs reported on the new site and the old v1 version of the site. It’s really important that users know what the APY was for the last 30 days so they can make informed decisions.

Scenario
Imagine a lot spending gas converting one coin to another because you saw it has 60%+ APY. Then imagine finding out that it was in fact only 15%, less than the APY available for your original coin. This person has probably lost a good few hundred $ in GAS fees because of this false information.

Not a big deal for some but it makes it hard to recommend yearn.finance to new users, when the site gives false investment information.

  • How can we fix this?
  • Can we calculate this ourselves?
  • Where is a good source for APY?
  • Why are some APYs listed as NAN or 0%?

Thanks

23 Likes

Admin on the telegram group informed me that they are working on this today, which is great news.

9 Likes

Not sure if reported APYs can be trusted yet, time to try the subgraph and calculate it myself and compare.

13 Likes

Hey! Sorry this has taken a while longer than anticipated, but I can provide an update:

  • I’ve been working with @x48 to implement new calculations for APY into our new SDK that Yearn is developing. This SDK is the basis for v3 of our website and will allow things to scale much more nicely, and as much as possible will be done programmatically (i.e. as little hardcoding of variables as possible).
  • The first step was implementing calculations for Curve vaults. We’ve been in contact with Curve and are using the same formulas they do, so our APY for Curve vaults should be updated in real-time instead of looking back in the past 1 month as they currently do. Here’s a screenshot taken last night of some mock data, and while this isn’t finalized it gives you an idea of the direction we’re going, and I am hopeful that we’ll be able to push these calculations live today:
    telegram-cloud-photo-size-1-4999506425952708823-x
  • Our new APY will also display the gross APY, as this is what competitors do and what we’ve found is more standard practice. We will also include fees applied and are discussing the best way/place to show net APY, as we don’t want to seem like we’re trying to “inflate” the APY—we just want for users to have a clearer picture of what yield Yearn is able to produce in addition to the yield they are receiving.
  • Like with the Curve vaults, our hope is to make all of our APY calculations more “real-time”, and in the future we’ll be implementing this on as many vaults as possible while also allowing users to select between averages over past time windows.

Edit: It’s worth adding that switching to Curve’s formula was specifically done to address the issue of 0% or N/A for vaults—basically, we know what the yield should be on these vaults, but if they hadn’t performed a harvest yet, they would show this error, which means no one will want to deposit, which means they’ll never harvest, and will keep showing N/A. A self-fulfilling cycle! So hopefully this update will help fix that issue.

15 Likes

great work dude(sahn).

12 Likes

@dudesahn
Very good job.

8 Likes

Great work! @dudesahn will these figures be added to the subgraph?

8 Likes

For questions about subgraph, would be best to ask @Doug, he knows far more than I do!

7 Likes

Yes, now that we have a very stable formula to calculate APYs we will work towards adding them to the subgraph, still some work left to do so may take some time to incorporate.

6 Likes