> why has no one made an app with the reverse incentive structure?

1. Network effects. An app isn't like a new local business where people will naturally wander in. They may already exist but the market's captured everyone on the skinner box services

2. App stores. The deeper you look into the things needed to advertise as a mobile app, the more obvious it becomes. You need milliions up front just to be featured in your critical launch time. If you don't, you fall into #1 and it's hard to recover from the "it's so empty" early impressions.

3. As you said, any success despite #1 and #2 is destined to fail. ad won't make that money up, so the only viable idea is relying on a premium or subscription model. But paid models in the era of "free" mobile apps is a hard sell unless you can guarantee success. And dating is anything but guaranteed.

That said other models have been tried to correct the issues with the big apps. Limiting matches, reversing the gender dynamics, based around special interests, etc. The only one I think I saw any kind of success from is one tailored towards rich/famous people meeting other rich/famous people (surprise, surprise).

Also note that as of now, Apple developer guidelines warns specifically of creating more dating apps. They consider it spam.