Note that the 'rough maths' ignores the 25% chance that the TWF rogue hits with both attacks -- though you can't do your Sneak Attack damage twice, the damage with the second attack pushes the DPR even further in the TWF rogue's favor.
Yeah, my maths-fu is weak, but it gives a broad and rough look at the advantages of swinging twice over swinging once. Two chances to land a hit each round gives the rogue higher DPR than a single attack with the cantrip rider.
As you point out there are variables. Both attacks hitting ups DPR in favor of the TWF rogue, as does factoring in the opportunity cost (taking magic initiate for access to booming blade, means you didnt bump Dex by +2 for +1 to hit and damage (among other stuff) which also increases DPR. On the flip side, being able to (hit with booming blade) and then (move away) with your bonus action also situationally increases the cantrip DPR. There is also the issue with the giant 'boom' of the cantrip as well (which you may not want to have happen as a stealth monkey)
Still, Im not seeing it as an 'ZOMG take the cantrip its broken' option. Our TWF swashbuckler looked at it and just pumped his Dex by +2. He has a pair of +1 shortswords and a Dex of 18 at 7th level and is more than happy with two attacks per round at +8 over one booming blade attack at +7 (particularly seeing as a Swashbuckler he can withdraw ayway).
There isn't much reason to be a straight Bladelock when a Paladin 2/Tomelock 3/Fire Draconic Sorc 6 can be just as effective and more versatile than a Fiend Bladelock 12, the ideal straight Bladelock, all at 1 level lower. The Bladelock may have 2 more ASIs and 1 more attack than the P/W/S, but the P/W/S has comparable DPR, freedom to choose whatever Patron they want, and actual healing instead of a reliance on situationally acquired THP. Add in Metamagic and Font of Sorcery/Pact Magic cheese (4 sorcery points recovered per short rest by converting your 2 2nd level pact slots) to the equation and the scale tips considerably in favor of the P/W/S.
There is no cheese with converting pact magic slots into sorcery points. It's a perfectly valid option.
And as well as being 2 ASI/ Feats down on the straight bladelock, its also missing 4th, 5th, and 6th level spells which is a pretty big deal.
But I agree, its otherwise perfectly viable option.
Why Tomelock? You are never gonna get more than 2nd level ritual spells in there. If its just for shillaliegh, I wouldnt bother. If youre running any of the AP's published youre gonna find a better magic weapon at some point. You'd get much more use out of the Chain familiar for mine.
Although I think "huge" is a bit of an exaggeration, I basically agree here. MCing is an exchange of otherwise focused power increase for more flexibility/diversity. Which I think is the right way to handle it.
I've used MCing more than a few times already in 5e and approve of the way it manages to play out in practice.
5th level in particular sees an almost exponential growth in class power.
Martials gain extra attack, doubling combat output. Rogues gain uncanny dodge, halving incoming damage 1/ round. Monks get the awesome stunning fist. Bards can use bardic inspiration per short rest instead of long rest. All casters gain access to 3rd level spells (fly, fireball, haste) and cantrips scale at this level.
Its a massive jump from 4th to 5th for mine.