I believe it's Theory Of Poker by David Sklansky that has a great section in the beginning comparing poker to other games, and specifically to tennis, which I've never forgotten.
I read this like 15+ years ago, so I'm paraphrasing, but it explains the core of why poker and other games mixing skill with chance are so counter-intuitive to humans and difficult to learn compared to games that are primarily skill-based like tennis (or chess, I suppose).
So, in summary: in a skill-based game like tennis, the quality of your results will almost always correspond exactly to the quality of your technique. If you have perfect technique, you'll hit a perfect serve. If you hit the ball into the net, you know you did something wrong. This is what humans are used to and it's the way we evolved to learn. Try > observe results > adjust strategy/technique > try again.
Poker short circuits this system because the results of a single hand don't tell you anything. You can play a hand perfectly and lose your whole stack. Or you can play terribly and win a massive pot. Hell, you can play terribly for an entire tournament and still win a WSOP bracelet if you're lucky in all the right moments.
This means you can't learn poker just by playing a lot, because you aren't getting reliable feedback. Instead, you need to learn the underlying theory, and then how to apply it in the heat of the moment.
In my experience, it's extremely hard for people to really internalize this. They'll say they understand, but then when you ask them about their poker session, they spend all their time talking about the results. People who really get it don't do that: they instead tell you about tricky situations they ended up in and how they decided what to do. There's nothing less interesting to a seasoned poker player than the results of a specific hand... or session... or week's worth of sessions.