Develop ways to guide the direction of future inspectors to write and bugs to fix

[obadiaa]

We currently do not have a clear idea of the coverage of MEV-Inspect. This is a proposition by @pdaian to arrive at a relatively accurate coverage metric. In addition, such a proposal would also guide us which inspectors we write next by focusing on the most profitable behaviours of addresses not already classified.

Proposal

• We run a "flash boys 2.0" style node that looks for any high price gas replacement transaction OR repeated gas token logs per tx OR spam (multiple txs per block), tags all those accounts as "MEV bots", and labels what percentage of them we categorize.
• We then manually check the top 100, validate that they are bots, and have metrics for what percentage of that extracted value is covered in our dashboard

Roadblocks For the gas replacement tx bit we need mempool data - something we do not currently have but may have soon.

Additional context This effort would be very helpful for MEV Explore, at the very least in order for us to display a coverage metric, and at best to add a few key inspectors to Inspect before Explore goes live.

[obadiaa]

Scott has created this Metabase question (https://dashboard.flashbots.net/question/27) joining a new db he's created that classifies self-destructs with our own db mev_inspections. The resulting question is a list of Inspect txs we haven't classified yet filtered by self-destructs.

I've put the current results in a spreadsheet (https://docs.google.com/spreadsheets/d/1F3VJpel1dYxWA4hufg7IZrKX8reTnlOGrJU9vf2tVWk/edit?usp=sharing) and going through them manually to add entries to #31 and #10