Guilherme/proto 3082 fix actor score ema initialization

[kpeluso]

Purpose of Changes and their Description

This PR introduces a new score initialization mechanism for new participants using the formula:
initialEmaScore = lowestEmaScoreInActiveSet - lambda * standardDeviationOfEmaScoresInActiveSet

This provides a reasonable starting point that requires consistent good performance to enter the active set.

The lambda parameter (lambda > 0) allows governance to adjust how challenging it is for new participants to enter the active set:

• Higher lambda: More conservative, requiring longer consistent performance
• Lower lambda: More lenient, allowing faster entry into the active set

We update the topic's initial EMA score for each actor (inferer, forecaster, and reputer) during reward distribution.

When an actor submits their payload, we check if they are a new actor. If they are, we initialize their score using the topic's initial EMA score, calculated using the active set's statistics.

Link(s) to Ticket(s) or Issue(s) resolved by this PR

https://linear.app/alloralabs/issue/PROTO-3082/fix-actor-score-ema-initialization

Are these changes tested and documented?

• If tested, please describe how. If not, why tests are not needed.
• If documented, please describe where. If not, describe why docs are not needed.
• Added to Unreleased section of CHANGELOG.md?

Still Left Todo

Fill this out if this is a Draft PR so others can help.

[github-actions]

The latest Buf updates on your PR. Results from workflow Buf Linter / buf (pull_request).

Build	Format	Lint	Breaking	Updated (UTC)
✅ passed	✅ passed	✅ passed	⏩ skipped	Dec 19, 2024, 1:42 AM

[amimart]

Well done! The init/export genesis func are missing the new stores, I've implemented it in #712 if you want

[amimart]

Looks good to me :)

[kpeluso]

Many of the conversations @spooktheducks incited will be resolved in #718 . Resolving for now, but not forgotten.