Oh boy, you've stumbled upon something I've been trying to figure out myself. The short version is that as far as I can tell, the "old analog" version as you refer to it is simply incorrect.
The long version:
First, watch
this video series, which you should if you're learning about color anyway, then come back. This answer assumes you've watched all of those videos, but I'll also just summarize quickly:
It's important to understand that you're not stuck with three primary colors, you can technically add as many as you want. But you can cover almost (but not quite) all other colors using just the three that match the cones in our eyes, which is why we usually say there are only three primary colors.
In digital color and/or printing you stick to RGB for simplicity (In printing you actually work with CMY because the process is subtractive, but they're really just an inverted way to say RGB. They're the same thing.)
You could use Blue and Yellow only as the primary colors, that would be suficient to cover the entire visible spectrum of a dog. A mantis shrimp would be able to distinguish upwards of 12 different primary colors. But humans can generally make do with those three, thought we ARE actually able to distinguish a bit more than what RGB alone can provide. Some TV manufacturers have experimented with adding extra primary colors to their TVs for that reason, though it never caught on.
So back to the matter at hand, at some point before all this was discovered and measured by science, someone working with paints also figured out that you only need a few colors to recreate most other colors. They decided that these colors were Blue, Red and Yellow (Often abbreviated RYB). This isn't actually correct, you technically cannot get true Cyan from mixing blue and yellow subtractively, and you cannot get true magenta from mixing red and blue, only purple. But they stuck to it anyway, and that's what kids are taught today in school.
There are probably a couple of reasons that it still works well enough in practice that nobody ever bothered to correct it:
A. When a painter says "blue", they're talking about a blue pigment, not a pure spectral blue, meaning a blue that consists of a single frequency of light that matches with the blue cones in our eye. Pigments colors made up of spikes in many different frequencies, called
metamers.
On top of this, a pure spectral blue is very dark blue, it's closer to what a painter might call a royal blue.
So what a painter thinks of as a pure blue pigment might have frequencies into cyan, or might even be a little cyan (and same with red and magenta). And the exact shades chosen by painters probably also vary a lot. I haven't gotten the chance to verify this with a spectrometer, but I'd love to test it for real one day.
B. In painting, you very rarely actually attempt to mix a full spectrum from only the traditional RYB primary colors. If you need a good cyan color, you'd usually go out and get actual cyan pigment, and you often use white, black, green and other pure pigments in your palette. Doing so is technically the same as another full primary color to your mixing palette. Since you can just keep adding primary colors to get the result you need, does it matter that much in the end if your basic primary colors aren't as accurate?
Combine these two, and it seems like it doesn't really matter that much that RYB is technically wrong, because there's so much wiggleroom and inaccuracy, that it only REALLY matters what LOOKS good.
So yeah. As far as I can tell, RYB is mostly a matter of tradition and is technically wrong. It's worth knowing about as context for lessons about color as an art since many are going to be using RYB as a basis to explain other concepts.
