[Sim] Integrate subgraph predictions data into simulations

[KatunaNorbert]

Target:
Make simulation profit results as close to reality as possible

Issue:
On high level, simulation profit using backtracking is not accurate and hard to trust.

• Currently when calculating the predictoor bot profit using simulations, the profit depends on a few variables which currently are mocked, which means they don't follow the other predictoors changes(stake, accuracy).
• Calibrating this variables is difficult because we don't have enough visibility in the current predictoors data.

Solution:
Use the new subgraph data provided through LAKE ETL to follow the real prediction data from the past inside simulations.

With the following changes we can achieve a more accurate, up to date predictoor profit metric:

• Update simulation flow to fetch the required prediction data
• Get and prepare the data to be used by the simulation flow
• For each simulation step( epoch ) calculate the predictoor profit using the calculated predicted value agains other predictoors predictions
• Update the UI with useful data that could help to calibrate the predictoor config towards making more money ( ex: display avg stake per epoch, display average accuray among predictoors etc)

[trentmc]

I recommend to close this issue, marked as duplicate. Why: it's mostly superseded by #1257

And it's that issue that gets to the heart of what we want

[KatunaNorbert]

I still think this issue has value, can we keep it in backlog and maybe come back to it latter?

[trentmc]

I still think this issue has value, can we keep it in backlog and maybe come back to it latter?

OK. Then please (a) update it where #1257 is a pre-requisite (b) mark as low priority