Currently app chooses "slowest" cards based on how much time distribution is different from a normal distribution (perfectly symmetrical). Which is not a good way of doing this. Because even if a user will improve his answer time for cards which were showed as "slowest", the app will find more "slowest" cards to improve. Makes sense from mathematical perspective or from competitive perspective but very unproductive for a learning experience (it just negatively affects motivation).
So, the new method of finding the "slowest" cards should be implemented in such way, that if a user will really improve on answer time for some specific items, there should be less or no more "slowest" cards. So the user will have feeling of accomplishment.
Currently app chooses "slowest" cards based on how much time distribution is different from a normal distribution (perfectly symmetrical). Which is not a good way of doing this. Because even if a user will improve his answer time for cards which were showed as "slowest", the app will find more "slowest" cards to improve. Makes sense from mathematical perspective or from competitive perspective but very unproductive for a learning experience (it just negatively affects motivation).
So, the new method of finding the "slowest" cards should be implemented in such way, that if a user will really improve on answer time for some specific items, there should be less or no more "slowest" cards. So the user will have feeling of accomplishment.