Well then just give your Rogue STR as it's primary stat and role play him that way. The whole point of multiclassing it to give a character the flexibility to do some things well that are typically the province of another class. Seems like every time a player wants something slightly outside of their role they want a whole new class for it instead of using the tools available to provide the same thing. There is a recent thread about an "Archer Druid" as well that made me laugh as apparently Rangers don't quite fit the bill and the player doesn't seem to think giving his Druid a decent DEX and perhaps the Sharpshooter feat is good enough. Well, not sure what to say about that. Some folks want a "Fire Wizard", and instead of just giving their player fire spells and perhaps the Elemental Savant feat for fire there needs to be a whole class for it.
There are players who think in order to "Roleplay" their character as a tavern brawler they need that Feat to do it. Others simply tell their DM they go to a tavern and punch someone in the face to get the same effect.
I don't disagree with you. Fortunately a new archetype is decidedly
not stealing (pun intended) from another class, and it is most certainly not "something slightly outside of their role." Maybe, just maybe, it is inspired more by something that has been in their role for years? Such as, for example, the fact that when I started playing AD&D 1e with my thief lo those many years ago, I had no problem Backstabbing with a Longsword. So maybe there is precedent
within the class role for a thuggish thief? And no... you can't do that RAW since you can't use a non-finesse weapon for a Sneak Attack.
Does this add other elements? Absolutely. Because in the evolved D&D that is 5e, you
are not completely abandoning or rewriting a class when you choose an archetype. You are emphasizing certain elements inherent to it. If you would prefer to multiclass rather than use a new archetype, great. No one is arguing against that. (At least not here.) But coming to a thread that is explicitly stated to be about an archetype and complaining about the existence and function of archetypes is far from productive. By your argument, both the Eldritch Knight and the Arcane Trickster should not exist.
You also seem to be implying that this is made just because "a player wants something." If I was the type to make homebrew willy-nilly, then this probably would not be the first homebrew archetype I've done since 5e came out. It is possible - this may shock you - that I have seen a need for something
in actual play that 1) does not violate the core of the Rogue class, 2) does not go outside the purview of the Rogue, 3) does not steal the thunder from another class, and yet 4) isn't in RAW. You are also making the fallacious assumption that a player originated this process. They didn't. I saw a need and looked at a fix. I haven't played a rogue since that 1e thief, way before it was even called a rogue. But that doesn't mean I can't perceive what I consider to be a problematic gap in the class archetype selection and address it.
I find doubling the base weapon damage on a sneak attack a bit much, when added to normal sneak attack damage. I would like to see something where he is given advantage or can negate some damage. Maybe allow him to use a shield or something.
It is hard to make new classes that are not better than the original class. I would not want to see a thief better at fighting than a fighter, or a druid better than a ranger at shooting.
Addressing your last part first: I'm not sure what you are thinking, but this archetype in no way makes the rogue better at fighting than a fighter. Since I'm addressing neither druids nor rangers, the shooting comment is irrelevant.
Second, addressing your first part: you obviously have only looked at the OP. As I mentioned, I revised it and made changes. Check out the current iteration and then let me know what you think. Your concerns have all been previously addressed.
