I sometimes log into my account on LinkedIn to answer questions. One of the sections I frequent, technology, had an interesting question that went along the lines of, “What dating sites do you use? Do they work?”

The question had around ten responses at the time I read it, and the general consensus was 9/10 saying that the dating sites did not work. I’m guessing that this is the case because of the social and the technical background of the respondents. However, I do not personally believe in such a high failure rate, as a few colleagues I know have been married and are living happy lives through such dating sites.

That’s when I began to think if there are other means to discover like, or opposing minded-ness. The dating sites I’ve read about tend to use a Myers-Briggs style test, or a question/response format. I feel the question/response format is more likely to succeed as a question is issued, and you are matched based on your response to others; the more questions you answer, your potentials are narrowed down to those that matched the same as you.

Back to my ever-continuous thoughts on randomization. I want to introduce own method, as the title states, the Chaos:Order Dating Algorithm, an off-the-cuff algorithm designed to match someone based on thought patterns.

If one was to register and log into my fictional website, they would be presented with an nxm matrix (nxm as I have not decided on the size yet, but must be a minimum of 3×3). The only thing that is asked of the user is to click on one of the squares in that matrix. After the initial selection, a display of profiles are shown above the matrix indicating the people who’ve selected the position of that square.

The user is then asked to make another selection on the matrix, but cannot select their prior move (but they can in the subsequent selection). Upon the second selection, the algorithm would then pull up a list of profiles that selected the same list of moves, along with the profiles that selected that particular square on that particular try #.

Essentially, the user is building their own state machine. As the user continues to select squares, the system will continiously filter out users, until the point where there are no more profiles to match against. From that point on, the last set of successful matches is considered the most likely to be like-minded.

I wonder how many moves would it take before the result set returns empty given x users and those users performing y moves?

Perhaps if I really built such a site, I would limit moves to one a day.