Mar 12, 2014 / Stop Using okCupid's MyBestFace
okCupid's mybestface app results are hardly worth anything.
I used to think okCupid's mybestface app is ok, but after more research, I have determined that this app isn't ok. It is terrible. You might as well just randomly select a picture because the results from this app appear to be random for the most part. Why does this app suck so much and why should you avoid using it? Read on.
<h4>It doesn't compare thumbnails</h4>
Comparing picture to picture is ok, but a lot of the action your profile is likely to receive is due to a thumbnail image, not a full size image. okCupid has somewhat indirectly remedied this issue by making the browse feature show larger images, but thumbnails are still an important component of online dating which you must account for. That picture of you riding a Ducati might be cool in its full 12 megapixel glory, but when a hot girl is looking through her visitors tab she may just pass over some microscopic image of what looks like a happy meal toy.
<h4>The sample set is low for the methodology</h4>
Usually a completed report only has around 50-60 votes. When each vote themselves contains such little information as A or B (with no quantitative measure like a 1-10 scale would have) your score is more likely a function of the arbitrary spattering of people who viewed you as opposed to your actual looks. There is some research that suggests the A or B approach gets more "accurate" results (whatever the hell that means) but not when the sample set is so small. This is easy to prove, for example, I have routinely had the same image score as much as *10*pts apart (out of 100) in tests. Here's an example of the output I got from a "completed report" comparing two IDENTICAL pictures:
[caption id="attachment_647" align="aligncenter" width="400"]<a href="http://cms.joshsway.com/6-okcupid-mistakes-to-avoid/6_okcupid_mistakes_to_avoid_resultsofsamepicture/"><img class="wp-image-647 size-full" src="http://cms.joshsway.com/wp-content/uploads/6_okcupid_mistakes_to_avoid_ResultsOfSamePicture.png" alt="okCupid myBestFace Results" width="400" height="340" /></a> Notice how different the results are despite the pictures being identical[/caption]
<h4>Give to get kills any credibility</h4>
Perhaps the biggest flaw with mybestface, the reason you might as well just randomly pick a picture versus using this app is the "give to get" model it employs. In order to get a completed report, you must first rate a bunch of other users, sometimes as many as 80. Now, what is your actual incentive to take the time to fairly rate the images you see? I'm going to say: NONE. I suspect that the vast majority of users simply hit 1s or 2s as fast as possible just to rate enough pictures to get their score. In other words, mybestface gives seemingly random results because the ratings are more or less random! Give to get can work in some settings, but when integrity of the "give" matters, the system fails woefully.
<h4>Judge looks by results</h4>
So, how should you judge your looks if okCupid's mybestface is more or less worthless? Unfortunately, good old "hot or not" (the old version that many of you are too young to remember) doesn't exist anymore, and even that is still a "rating system". The best way to evaluate your pictures is by results! Simply change your picture on your profile and note the differences. You can even spend a couple bucks and use one of the "Boost" features on okCupid to quickly get the picture in front of a lot of women who are ACTUALLY BROWSING FOR MEN, not just clicking random 1s and 2s to get their own score.