Dating would be different if all dates were first dates. This doesn't just apply to dating, but really to anything that has a chance, however small, of continuing to give a payoff for a long time. We tend to be risk-averse, and we should tend to be risk-seeking. I'm not talking about walking-down-dark-alleys-to-see-what-will-happen risks. I'm mostly thinking about benign ones that we wouldn't even recognize as risky: trying to drive a different route to your friend's house, or a new recipe, or being social at parties. The worst that will happen is that you are back where you started.
If all dates, no matter how good, were the first and the last---if a positive lasting connection with somebody was not among the possible outcomes of a night out---then the best way to date would be to imagine one thousand blind dates, good and bad, with a thousand people. If the average enjoyment is less than the average joy of staying at home, you should stay home. In the psychology and economics of decision making, you are reasoning about expected utility in the context of uncertainty, and you get into risk. Some people expect the average turnout to be positive, but they stay home on the chance that a given date will go bad anyway. These people are risk averse. Some do the opposite, and gamble on the slim chance that its worth it. These people are called risk seeking. And the people who play it like they see it are called risk neutral. This is dating in a world where you'll never get a second date.
How do things change if it is possible to continue seeing somebody that you like? You no longer care about the average joy from a thousand dates. The numbers change when you are playing for keeps. To simplify the game, say that you get to try out the thousand prospectives, and pick who you liked the best. In this case we aren't comparing a night alone with the average date, but with the yet-unknown best date in the batch. You are playing for the max instead of the mean. In this context, behavior that looked risk-seeking before is the most rational (and also the most likely to make you happiest in the long run).
Scott Page gives this toy example from his book Diversity and Complexity. Imagine that you get to pull blindfolded from two big bowls: one has black and white marbles worth 8 and 9 dollars respectively, and the other urn has ten different colors of marble worth 2,3,4,5,6,7,8,9, and 10 dollars. You can think of these as two dating pools: one has high average quality (the friends of that one classy friend), and the second is a terrifying leery dive that somehow attracts this impossible invisible minority of perfect gems (the nines and tens).
You draw a dozen marbles and get paid at the end based on what is in your hand. If the amount of money you get is the average of the marbles in your hand, you should always draw from the first bowl, because you'll tend to get $8.50 (instead of just $6.00). But if I offer to pay you based on the highest value marble in your hand, you should draw from the higher-risk bowl. For more than seven draws (if you pull more than seven dates from the dive) you are likely to draw a $10.00 marble at least once. If you live in a world of second dates, you should date a lot and you should take your chances with the dive.
The same goes with any other gamble that rewards the max instead of the mean. Try cooking: You will eat better if you forego the same safe recipes and just start cooking wild crazy inventions. This is because in cooking we can learn and/or write things down. Junkyards, thriftstores, research and development, learning to walk, team sports, anything that hasn't been solved yet. You only need one out of a hundred crackpot ideas to pan out.
This is a strong case for taking more risks in your everyday life. And yet, decades of research, across hundreds of domains, have shown that people tend to be risk averse---or exactly wrong.
But behavior that looks reckless can be rational, and trying new things can be the way to lasting happiness.