The point of having the system before you play is just letting everyone know what to expect on the G part of RPG so they can focus on the R and the P. It kills all momentum to stop and ask, 'so what system do we want to use' in the middle of the game. No system is perfect, but it cuts out a lot of work to just pick one and roll with it, homebrewing over the few holes that show up.
I mean, it might work if your group is all kind of on the same wavelength to begin with. But if that's the case, you could also easily start with a system you like and go from there instead of reinventing all the wheels.
A lot of people have only really played D&D and its close relatives. I like to describe that in this metaphor: Imagine someone who has only every seen the lord of the rings movies. They've watched them over and over, both cinematic and directors cuts. They know all the lore and all the minutia. And then they sit down to write their own movie. Maybe a sci-fi space mystery to change things up. And this movie? it has horses. Because movies always have horses, don't they? They're in like every movie. So when the detective is stuck in the burning theater, his buddy should ride in on a horse and save him.
So I 0%, maybe even some negative percent, want to have to sell a group on "RPGs don't actually need six attributes" or "you don't need to have separate rolls for to-hit and damage" for the first time in their lives.
Secondly, most people are bad at design. Sorry. It kind of follows from sturgeon's law ("90% of everything is crap"). Most people don't set out to make crap, but it happens anyway. Most people firing from the hip are just not going to make good systems. Especially if, as above, they've only ever really played one kind of game. So, no, I don't want to deal with the guy who's like "On a natural 1 you should drop your sword" who doesn't realize that, because fighter types make a lot more attack rolls, they're going to drop their swords way more often than you'd expect of the archetype. I am reminded of an unhappy time in an old, bad, D&D game where I fruitlessly tried to explain effective HP to the wizard. (Since D&D 5e stops counting damage at 0, there are some weird interactions between initiative, healing, and damage.)
Third, even if you avoid all of that, even if you have a group with a deep and wide knowledge of game design, you're going to end up with an inelegant mess. Why does intimidating someone mean a simultaneous roll-off of increasingly large dice, but bluffing someone means drawing poker hands? Because those rules were added on different sessions, and Mike was really into poker and convinced people it would be cool. Wrestling someone you flip coins, but knife fighting you roll d4s. Sword fights use this complicated table Joe insisted would be fun, but magic is just a roll off. No thank you.
I'd rather just play Fate, which is already pretty loose about how to interpret conflict and consequences.
There was a whole phase in early "game design" where every game was basically D&D with a bit of a facelift here and there. Genuinely new games were few and far between (and are the celebrated games of the era now). Then the '80s happened and game design went all over the place with wildly creative ways of doing things happening (and like every wildly creative phase in any discipline, a lot of it was a really stupid direction to take things, so withered quickly on the vine).
Then this weird phase happened in the early '90s where people nobody had ever heard of or from came out of the woodwork to tout their "grand new RPG" that "solved all the problems of previous games" ... and it was always just another variant of D&D. These were people who'd been playing (usually) AD&D for over a decade building up house rules and then deciding that they would publish these house rules as a "new" game system. And it was clear they'd never even once been in a game store, not to mention talking with other designers or playing other games, over their entire span. Because they would "solve" things by proudly proclaiming the number of classes they had so you could play the character you want. (One game had 114 classes!) Or how you could play any race and class in combination. Or, you know, things that hadn't been an issue at all since the introduction of Runequest in 1978.
It was always so tragic. These games were amateur in the literal sense: the product of great love. A lot of time, effort, and money had gone into their publication. And they were doomed on impact because while they were, arguably, an improvement over AD&D (the king of the gaming castle at the time) they weren't sufficiently good to be worth switching to. I had about 20, maybe even 30, of these games on my bookshelf just as a mute testament to what happens if you try to hit a market without even elementary market research.
I think you missed the point of the blog. It's not "A guy makes a game" it's "a group of friends make stuff up and a game is carved out of it"
Which yeah I can understand not everyone likes that. You need some imagination, the willingness to try new things and generally be down for some chaos. For the rest of us it sounds like a fun way to just fool around with friends.