Talk:Champion Kit Box/@comment-31458611-20190308165457/@comment-34918040-20190308193902

What a pity, I really should have written down some rules somewhere.


 * "it didn't occur to me that, say, tires vs drivetrain count could be of any relevance. (In fact, it still doesn't occur.)"

It is. Some boxes are tyres-drivetrain-focused, some are suspension-exhaust-focused, and the, for example, so far has a clear peak in tyres only. I used this strategy when I was desperately looking for class B parts during the Lotus Evora Sport 410 R&D.


 * "I'm gonna open two 5-engine boxes from MC8s later. Interested?"

Good question. And a difficult one. I've been thinking about that for some time now, but come to no conclusion yet. I tend to the following statement:

The smaller a box is, the more it is important to have a continuous recording of all of these boxes on one system for a longer time. If someone writes me: "Hey, I got 3 CREs in a Finish Box today, how cool is that?" and I add that to the data set, it will surely distort the results, especially if the sample size is still small. If someone is willing to track each and every box they open for three months or so, I wouldn't object, but I know this is nerd stuff and a whole lot of work.

As for bigger boxes, especially the "fat ones" like Champion Kit Box or MP league reward boxes, I think it's ok to include data from others as well, as these boxes are rare, and the bigger the dataset, the more precise and useful the predictions will be for other players.

By the way, the date is important in order to verify (or debunk) some of the Asphalt 8 myths: "If there's an event for a car that needs XY engines, they don't give these engines to us deliberately." That's why I've started to also track the active events when opening a box.