The difference is that most respiratory viruses (flu, SARS, etc.) have a higher effect on children than they do adults even with a young immune system. This one does not, for reasons that are still not understood.
Except, they have some good guesses. This virus likes to attach itself to ACE2 protein receptors and kids just don't have those in their noses & lungs in the same quantities that adults do. In kids, the virus is trying to climb a flag pole, coated in a slippery substance. It can't latch on in sufficient quantity. They are also looking at another genetic factor called IL-10. It's a cytokine that reduces inflammation that is present in higher quantities the younger a person is.
I get the impression, that scientists have a general understanding, but are waiting for the results to come in before speaking with authority. Unlike some other scientists, who present a pre-print of a bad study and then scream from the rooftops that their results prove that we are all overreacting to this virus and if we simply stop trying to control anything, then this will just all go away soon. But their caution doesn't lend itself, toward tantalizing media stories.
I have been looking into this 20% herd immunity a little. Apparently, the paper assumes that 50% of the population simply can't get sick. So then by the time you add the 20%-30% who do get sick, then viola, threshold reached. But they provide no reasoning other than "magical T-cells" for why people can't get sick. They ignore that there are neighborhoods and locales where seroprevalance is high, so apparently the 50% of the population that has magical T-cells, simply don't live there? Miami seems to be one of these places as they had a good outbreak before, and have another one now. In the US, we have scientist whose name keeps popping up because his credentials also include Stanford. But newsflash, good universities can employ people that will do bad science to prove a point (at least that's the reason why some people see Harvard and immediately discredit them). This Oxford team, seems to be the UK equivalent. Before presenting this option, she and at least one of her co-authors also presented the idea in March, that the virus started circulating earlier than we thought, so that people didn't get sick because "everyone already had it." Then the serology results started coming back and people had to stop using that as justification for why we didn't have to do anything. So they went from 50% of the population already had it to 50% of the population can't get sick. She is also a theoretical epidemiologist (Big Bang Theory fans can imagine the differences between Sheldon and Leonard's views of physics), so it's her job to think up novel explanations. But we aren't living through a theoretical exercise, eventually the theories have to resolve with the actual "observed" experiences.
FWIW, my "I'm not a scientist" opinion for why things have slowed down is because of the behavioral aspects and comorbidities associated with economic inequality. It spread like wildfire through people who did not have the ability to self-isolate because of their essential worker jobs, then went home to smaller, more crowded living conditions, with multi-general households (because that's all they can afford). These people also lack access to regular, preventative health care and so have several comorbidities that make them susceptible to more serious infections. Like, older people in care facilities, the virus has burned through through this group where large outbreaks have occurred. It doesn't make everyone else "safe," it just makes everyone else less likely to have been exposed. When higher class people expose themselves through parties and travel, they get sick too. An analogy using lava: A'a lava moves very fast. What we saw during Kilauea's most recent eruption. Pahoehoe is a slow moving lava. It's what we saw, still at Kilauea, but at the Pu'u O'o crater. The lava slowly advanced on the closest community. But as far as the ground, the structures, the animals & people in its path, both types of lava burn and destroy the same.
I do believe that T-cells absolutely play in a role in outcome severity. But not as "magic bullets" that are going to save us from having to do anything. I remain skeptical of solutions that align with human desire for all of this to just go away, because we are tired of dealing with it.