Calculation of hit ratio
Posted: Wed Jan 02, 2013 12:00 am
Hello,
I think this is something that should be changed:
the hit ratio of a country, a city or a user is currently calculated by the number of individual hit notes that have been entered in or by the subject, divided by their total number of entered notes.
This means that a hit between Germany and Finland will increase the hit ratios both of Germany and of Finland, while a hit within Finland will just increase Finland's hit ratio by one note, where both users will have their hit ratios increased by one note each. That's why Finland's hit ratio is higher (less hits per 100,000 notes) than any one of its provinces's hit ratio.
Much better in my opinion would be to divide the number of note entries that belong to hit notes by the total number of note entries.
This will not influence the hit ratios of users, but most cities and countries will have lower (better) ratios, that will correspond correctly with the hit ratios of their subdivisions and the hit ratios of all users.
You could then also calculate hit ratios of time intervals.
For example, for my hit ratio in 2012, I need the number of notes I entered in 2012 that are hits. My hit list tells me I had 31 hits in 2012, but I entered 7 of them before 2012. No notes from 2012 have been found later, so there are 24 hit entries from 2012. So my hit ratio of 2012 is currently 24:6119 = 1:255
What do you think?
I think this is something that should be changed:
the hit ratio of a country, a city or a user is currently calculated by the number of individual hit notes that have been entered in or by the subject, divided by their total number of entered notes.
This means that a hit between Germany and Finland will increase the hit ratios both of Germany and of Finland, while a hit within Finland will just increase Finland's hit ratio by one note, where both users will have their hit ratios increased by one note each. That's why Finland's hit ratio is higher (less hits per 100,000 notes) than any one of its provinces's hit ratio.
Much better in my opinion would be to divide the number of note entries that belong to hit notes by the total number of note entries.
This will not influence the hit ratios of users, but most cities and countries will have lower (better) ratios, that will correspond correctly with the hit ratios of their subdivisions and the hit ratios of all users.
You could then also calculate hit ratios of time intervals.
For example, for my hit ratio in 2012, I need the number of notes I entered in 2012 that are hits. My hit list tells me I had 31 hits in 2012, but I entered 7 of them before 2012. No notes from 2012 have been found later, so there are 24 hit entries from 2012. So my hit ratio of 2012 is currently 24:6119 = 1:255
What do you think?