OkCupid – fuzzy math?

Okay, I generally “get” how OkCupid’s percentages are calculated, on e.g. match, etc. In fact they do explain that very well.

However, … fuzzy math? I’ve yet to make sense of how they come up with some of these numbers – particularly/especially with the above in mind. E.g., I not too uncommonly find “matches” which, at least on the surface, give a rather baffling set of numbers. I might guess at where the “enemy” % comes from, and maybe even “match” vs. “friend” uses uses different questions, or different weightings among questions, … but in any case – even just with the match bit … how does one come up with a combination like the following – I’ll not uncommonly see something like this:
99% Match
63% Friend
You match.
85% on Ethics questions
93% on Sex questions
91% on Religion questions
88% on Lifestyle questions
86% on Dating questions
82% on Other questions
So, … given all the individual categories/breakdown – and especially since it also includes a category of “Other questions” – how the heck does one take those percentages, that range from 93% to 82%, yet come up with a composite for “Match” of 99% (higher than any of those individual categories), and “Friend” of 63% – lower than any of those individual categories. Somehow the weightings don’t seem to add up, as one might expect. I don’t know, perhaps for the various subcategories, they don’t use weightings at all – and maybe that would explain it. I.e. in, oh, say “Ethics questions” – let’s say there’s a large range of well matched, and highly unmatched question responses … but let’s too, say the ranking on how important, is high on the matched, and low on the unmatched – that would give a high “Match” % … whereas the rating for “Ethics questions” may be significantly lower … and same might occur across all the categories, thus leading to the “apparent” discrepancy. If such is the case, seems OkCupid ought to apply the same weighting/”importance” matching algorithm when showing the various categories – otherwise the results displayed are quite non-intuitive – and also generally not so useful or consistent. Besides, isn’t it better to know the results in the categories, based upon / including the “how important” rating?
Maybe they’ll “fix”/tweak that some day. :-) In the meantime, it’s rather confusing – if not at least counterintuitive. Maybe too, someday, they’ll explain how exactly they come up with the “Friend” – and also “Enemy” – percentages (though I’d guestimate on “Enemy”, it’s the “unacceptable” answers and importance ranking – quite similar to “Match”, but unacceptable, instead of matched/acceptable). But, … what exactly does “Friend” come from? How’s that different from “Match”, exactly?

Advertisements

Tags: , , , , , ,

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: