[SPIKE]: Publish Miner Competition MIP

[grantdfoster]

Problem Statement

We need to have a centralized, documented conversation around the miner competition improvement plan on the subnet. An MIP will both capture the current issue and propose a solution, or many possible solutions. While #282 spikes how we could fetch dynamic queries, this MIP encapsulates the full picture of improving miner competition, from improved queries to reduced randomness.

Time Box

2 days

Background

The miner metagraph quickly reaches a plateau on new software releases - meaning that all miners are quickly able to achieve the maximum amount of work, scoring essentially "perfectly". This is an issue for miner competition. The result is that, due to our randomness in how miners are selected for volume scoring, incentive/performance is thus more "random" than "earned".

Key Questions to Answer

1. How will synthetic validator queries be determined?
2. How will miner responses be recorded to determine uniqueness?
3. How will we reduce the randomness involved in volume checking?
4. Should we remove the tweet count ceiling?
5. How many miners should be queried, how often?

Suggested Approach

Take the conversations from #282 and on Slack and synthesize them into an MIP

Deliverables

• Published MIP
• Agreed solutions
• Implementation tickets cut