To your "more than 30 potential mates doesn't help much," I found similar trends doing a bunch of Monte Carlo simulations of the Secretary Problem (Optimal Stopping). Broadly, if you're fine with a top 5% or better candidate, you're actually better off dialing your evaluation fraction back to 10-20% versus the analytical "best" 37%, and the most candidates you should be evaluating before choosing the next candidate that impresses you relative to your evaluated pool is probably around 20. Interesting that the two methodologies showed the same trend of diminishing returns for larger dating pools.
Cool! One extension could be to give superlinear returns to matching with someone rare. e.g. matching with a 99th percentile partner is more than 10x as valuable as matching with a 90th percentile partner.
Yeah, I actually wrapped the whole thing up in an analysis of "rarity" too, because I've long had a conceit that "you only get 3 things" in mate optimization, and this was the theoretical bracketing of that. I'll link the draft if you're interested: https://performativebafflement.substack.com/p/d27467a3-f1e8-4143-8dd7-e1456a26b449
To your "more than 30 potential mates doesn't help much," I found similar trends doing a bunch of Monte Carlo simulations of the Secretary Problem (Optimal Stopping). Broadly, if you're fine with a top 5% or better candidate, you're actually better off dialing your evaluation fraction back to 10-20% versus the analytical "best" 37%, and the most candidates you should be evaluating before choosing the next candidate that impresses you relative to your evaluated pool is probably around 20. Interesting that the two methodologies showed the same trend of diminishing returns for larger dating pools.
Cool! One extension could be to give superlinear returns to matching with someone rare. e.g. matching with a 99th percentile partner is more than 10x as valuable as matching with a 90th percentile partner.
Yeah, I actually wrapped the whole thing up in an analysis of "rarity" too, because I've long had a conceit that "you only get 3 things" in mate optimization, and this was the theoretical bracketing of that. I'll link the draft if you're interested: https://performativebafflement.substack.com/p/d27467a3-f1e8-4143-8dd7-e1456a26b449
Why don't you apply this model to Indians? Their caste should also artificially limit the dating pool.