gaelreinaudi / cache-bot Goto Github PK

I am going to try to return the following in the json oracle:
"savedFor" = daysToPres / daysToNext = "time from last such trans to present" / "time we expect(ed) this trans should come within".
Indeed, this whole PriceWindow expects something to happen every other "daysToNext". So 1 day after the last transaction,
we have actually saved for 1/daysToNext, 2 days after we have saved for 2/daysToNext, and daysToNext days after, we have saved for 100%.
After that, we have saved for 110%, 120%, etc...
But if you suddenly buy something that falls in this price window, you kinda reset this to 0%,
unless if we had saved for much more than 100%, in wich case, sure, we should buy it anyway.

• Close connection immediately
• Create cache-bot route for receiving pulled data

that led to one of Chris' rent check to fall into another window and crap the statistics of his rent

solution, compute predefined points where the borders can be (low transaction density).

investigating

each feature should give some detail of how it evolved recently

using an EMA for now. it is silly because if we have only a few transactions, old ones will play a larger role than recent ones.
must implement an agePonderated average. (like for randSmart)

which is currently normal. it doesn't change much the probability but made the average of the $300 bucket go down.