>they wouldn't bother taint their results.
Results are tainted by default. The simplest dating app is a random number generator attached to a list of user IDs. This probably performs badly. (But who knows? Maybe build it and see what happens!)
The question is simply whether there is really a deliberate, focused effort to identify who gets into long-term relationships and to replicate the conditions that lead to such relationships. If dating sites were actually trying to help their customers, they would do this; someone would have worked on it; there would be an internal lore about what was learned from these attempts. The absence of any of this shows that it just isn't a priority, and that's enough.
>Dominating the dating site market is a money printing machine
One problem with modern dating is that people don't have a potential partner's reputation to go on as much as they used to. Relationships formed "through friends" have declined significantly, in tandem with a decline in friendships themselves. But this decline in friendships also poses a problem for any word-of-mouth dating site marketing strategy. People who have lots of friends are probably the least likely to benefit from dating apps, but they're the most likely to hear about the hypothetical "good dating app".
I don't actually know how most of my friends met their partners — even people who I know many other things about. It's in my mind as an awkward question. Furthermore, most long-term relationships in my friend group are older than most dating apps. It would inevitably take years to build a reputation as a dating site that does this, because you don't know if you have a stable relationship until it stands the test of time. But investors generally aren't so patient.