Yes, agreed.

But if you can consistently predict whether an idea is "good", by definition you should be either hitting it out of the park consistently or getting massive returns on investments.