Betting on Love
A Valentine's Day gamble in a prediction market
What you want
Do you think you know what will make you happy?
You’re probably wrong.
Sounds harsh, doesn’t it? And I bet you think I’m the one who’s wrong. Yet, we know that people are - on average - pretty bad at affective forecasting, or: predicting how they’ll feel in possible future scenarios. This is a well-established psychological phenomenon known as miswanting.
(For more on this and how it relates to smoothies and treadmills, check out my earlier ‘On Happiness, Fake Pies, and Treadmills‘.)
What if we ‘miswant’ in dating too? Because, let’s be honest, we all have an ideal partner in mind. Whether it’s a checklist of boxes to tick or a collection of green and red flags, we know what we’re looking for in a possible romantic partner. At least, we think we do.
Yet, both in the short and longer term, our checklists and flag collections don’t tend to do as well as we hope when we put them together. As an example for the short term, a study in speed dating finds that:
… participants' ideal preferences, assessed before the speed-dating event, failed to predict what inspired their actual desire at the event.
On the other side of the spectrum, a study following over 300 relationships for 13 years finds that:
A closer match was associated with higher relationship commitment across all metrics, while for relationship quality, the link was not apparent for the corrected pattern metric. Evidence of matching effects for relationship length was mixed and largely absent for break-up initiation.
Translation: if you meet someone closer to the partner ideal you have in your head, you’re more likely to commit to a relationship but this does not mean a better or happier relationship nor does it automatically mean a longer relationship.
We tend to overvalue physical attributes even though a set of compatible (which does not mean identical) values is often what makes or breaks a relationship. Not that physical attraction doesn’t matter. Of course, it does1. Not, however, exclusively so. Today, we’re not just looking for a hookup, we want to build something with the potential to last.
Maybe we have to tap into the wisdom of the crowds to solve our partner miswanting problem.
I predict you
In the 18th century, French philosopher and mathematician - big breath - Marie Jean Antoine Nicolas de Caritat, Marquis of Condorcet, also known as the less tongue-twisty Nicolas de Condorcet, came up with the Jury Theorem. I am quite allergic to formulas, so let me toss some details aside and simplify his theorem to ‘the more people you ask, the more accurate the answer’ (not for every type of question, of course, but that would lead us too far).
In essence, Condorcet provided the foundation of prediction markets with his Jury Theorem. Manifold and Metaculus might be the most well-known ones at the moment. The premise is simple: ask a bunch of people to make predictions and you might end up with stock market or geopolitical event predictions that are better than those of domain experts. This is not always the case; I’ll get to some caveats later.
What if we set up a prediction market for love? That’s exactly what manifold.love (an experiment/initiative by the earlier mentioned Manifold) is: a combination of a dating site and a prediction market. Here’s their announcement post (on Substack, so hi there,
).You either sign up as a dater or a matchmaker (or both). As a dater, you fill in all the usual profile things, plus a personality and value questionnaire, as well as compatibility questions. As a matchmaker, you predict which profiles are most likely to be a match and, in doing so, you can earn play money called mana. Successfully predict which hopeful daters will date for at least six months and the mana will rain down on you.
Daters receive the matches the matchmakers have made for them and either say, “Yes, interested,” or “No, thanks”.
Color me intrigued.
I wrote about dating in the Metaverse a while ago and about a third of (US) adults meet their prospective romantic partners online. But it’s not all sunshine screens and pixelated rainbows. Online dating leads to ‘choice overload’. According to this dating app study:
The paradox of modern dating is that online platforms provide more opportunities to find a romantic partner than ever before, but people are nevertheless more likely to be single… The continued access to virtually unlimited potential partners makes people more pessimistic and rejecting.
And from a recent Pew survey:
Women who have used online dating platforms in the past year are more likely to feel overwhelmed by the number of messages they get, while men are more likely to feel insecure about a lack of messages.
As a result of this rejection mindset and app fatigue, dating apps have been losing users (and income). People who are on the apps to find someone for a possible relationship, meet a conflict of interest: the apps want you to stay (you are a client, after all), and you, as a user, want to leave as quickly as possible. Not to mention a bunch of fake profiles and the gender imbalance that implies a horde of mindless horny men who will pounce on the women who dare to venture into the apps’ virtual hinterland.
Why not outsource some of the dehumanizing and mentally fatiguing swiping to an (ideally) impartial crowd?
Caveat predictor
If all of that sounds impersonal, I get it. But is it really more impersonal than swiping left or right based on little more than a handful of pictures and a bullet point profile, both of which have probably been filtered to be as ‘attractive’ as possible? I am not convinced the prediction market idea is worse than the swipe-mania...
I wouldn’t be me if I didn’t poke the idea a bit more, though. Remember the caveats I mentioned? Prediction markets can go awfully wrong. Clinton had a 91% chance of beating Trump, for example. We all know how that went…
In his book The Wisdom of the Crowds, journalist James Surowiecki sees three conditions for a successful prediction market: diversity of information, independence of decision, and decentralization of organization. Let’s take those in reverse order.
Decentralization of organization: As far as I can tell (and my knowledge about the nuts and bolts is limited, so apply big spoons of salt), Manifold is a decentralized autonomous organization. Nothing is ever fully decentralized, so there are staff mediators in case of disputes, but overall it does appear as though matchmakers can follow their gut without Big Cupid watching. (EDIT: It is NOT built on the Ethereum blockchain as I wrote earlier. My apologies!)
Independence of decision: In a thinking group, the group that thinks has to be isolated from groupthink. (I could not resist that sentence.) In other words, prediction markets only work if people make their predictions independently. As soon as trends or - algorithmic gods help us - influencers enter the stage, the game’s up. So far, matchmaking on manifold.love seems quite independent. I hope they’re not thinking about public ‘super predictor badges’ or anything, since that could swing the matchmaking market in unfavorable ways
Diversity of information: This is the tricky one for love’s prediction market. Profile stuff + questionnaire does give matchmakers diverse things to go on (not unlike paid matchmaking services, to be honest). However, people lie. I can imagine that aspiring love seekers ‘massage’ their responses to appear more attractive as a partner. Also, love is (probably?) not fully quantifiable. But good idea of manifold.love to require six months of dating before declaring a match as ‘successful’ (they are thinking of adding shorter duration milestones). Another plus: there are filters for wanting/not wanting kids and relationship types beyond traditional monogamy. Those can be ticking time bombs if left simmering in the background (yes, I like mixing metaphors). Finally, I’d add that diversity among the matchmakers matters too. I’d wager that people who know and are interested in prediction markets are a subset of the population. My second guess is that, even though it’s open internationally and there are people from all over the world on manifold.love, it is still US-dominated (nothing wrong with that, just an observation).
I’d love (ha!) to see more manifold.love data on the proportion of successful matches or predictions. Or maybe whether or not there are specific questions/personal value alignments that predict a romantic match better than others? Getting that data might be difficult in terms of privacy, though. Honestly, this could be a gold mine for psychological research. If some people find love along the way, all the better.
So… do you feel lucky, fellow love-seeking punks?
You may have heard or read that evolutionary psychology predicts that physical attractiveness is more important to men seeking a partner, but that’s only theory. Some studies suggest that, in practice, in our modern social media availability-biased world, the physical attractiveness of a potential partner is just as important to women. From the linked study (emphasis mine):
… participants were more romantically interested in potential partners if they were viewed as attractive and good potential earners, and these associations were not moderated by gender.
Another study that included non-heterosexual participants likewise found no difference in the importance of physical attractiveness, and the researchers do indeed point to the increased tyranny of ‘typical’ beauty ideals as a potential (partial?) explanation.
… irrespective of sexual orientation, women and men tend to share similar standards regarding their own and a potential partner’s physical appearance, potentially suggesting an increased hegemony of heteronormative beauty ideals in women and men in general.