On Thursday, November 14 at around 5PM EST, we will be changing how Scot rewards are computed.

This is expected to have only small changes in behavior, as the formulas are exactly the same. However, the way `pending_rshares` are computed will be adjusted to be more in line with the way Steem does it, which has the following properties:

1. Pending claims vary more smoothly over time. Rather than being the rolling sum of post rshares (after reward curve is applied) for pending posts, it will reflect an weighted averaging of recently paid posts (most recently paid posts weighted higher). This is how Steem works. The averaging is over 15 days like Steem when payout is 7 day window, otherwise `2*cashout_time_days +1` days.

2. Pending claims will be computed at more regular intervals. Right now it is not very predictable when it occurs, which is rather problematic in terms of trying to set up a backup that can (eventually) replicate the same state. It will now be done at intervals matching the reward pool adjustments (each Scot token adds rewards every n blocks).

In addition, there are a few changes to speed up this process:

1. No custom json broadcast from token account for post beneficiary information. This information should be taken from the account history endpoint. Note that even now these broadcasts are very prone to failure and frequently do not occur.
2. No direct issue/transfer for token beneficiary. It will be placed in pending amounts for this account. However, we will add these accounts to autoclaim periodically so they do not have to explicitly claim.

These changes have been running on a backup server with no issues and comparing the two systems in parallel for any significant issues.

Let me know if you have any questions, as a Scot tribe owner or participant. For the most part, this is really just a notification that a change that should have little to no impact will be applied soon.

---

Edit: The changes are live. Everything looks to be in order now. There was a period of time where the /info endpoint was not working but it was corrected quickly. Thanks to @themarkymark for the report on the bug. I do see an improvement in speed as well in terms of how fast the bot will pick things up on average. It's not perfect but will continue to look for opportunities to further optimize the Scot bot processing.