fatslob-:O said: I have a question for OP but how many times has he adjusted the model to be consistent with the NPD numbers thus far compared from his collected data ? You're not forced to answer if you don't want to ... |
This is a good question.
I started this tool in march, I used march NPD results to create the formula in order to predict April. As it was surprisingly succesful, We kept the formula like this. This is pretty much it.
Oh yeah, down the line, I added a 1.3 factor to XBO points as it was consistently underestimated (I am not sure which month it was though it is somewhere in the thread...).
And the month of PS4 million, we had to only count one day of data for those, despite the fact they are still on gamestop list ...
Holyday season will require a big rework of the formula though, but those changes (if we come up with something decent) will remain "holidays only".