Yea, it's certainly more complicated than a simple IP bans.
Reddit uses a shit ton of fingerprinting on top of just using the IP.
IP info can be a part of said fingerprinting, but banning just an IP is moot and easily bypassed with a lot of possible collateral damage they'd want to avoid.
IP bans could ban unrelated accounts that happened to connect from your workplace proxy or mobile ISP proxy, but is easily bypassed by anyone with even a free VPN or coffee shop wifi.

I'm sure they have extensive data points to connect the dots between alt accounts, but it's assuredly not using only IPs.

If you don't mind, I'm going to comment and edit as I read your comment. That cool?

So they decided to choose an algorithm that matched like accounts? Bullshit

So if they're (what you're saying) looking to ultimately ban by, what they consider, problematic IPs, and banning them with that assumed problematic IP under the impression that no one can learn from said experience, and they can do that for forever is asinine(this sentence was a hard think for me, so please read carefully). Originally, the only people that could IP ban were admins, which in this current point in time, admins don't do shit. And that's because mods got overly precise in their prejudice and big reddit gets far too much money for commercializing.

So in short, devoted neckbands get big confidence in their parents basements, and Steve Huffman keeps getting richer