If your argument was true in general, then every spell added to the wizard spell list would be a bad or at least unhelpful thing. Whereas in fact I would conjecture that additions to the MU/wizard spell list is the single most frequently published game element in the history of D&D. (Maybe there have been more monsters. That could be the only rival category.)
I fear you haven't grasped my argument, if you say I am characterizing everything new as bad or unhelpful. What I say is that once you reach a good amount, other things that get added are likely to either displace either options (by being better) or to be bloat (but not being played), to an extent more than just spreading out what people play (actual additional choices). However you bring up a salient point about nature of the newness and how different categories can vary that's worth discussing.
Generally a casting character has more spells known then classes, and uses a variety over the course of an adventure. They are initentionally looking for a variety. So it's not like classes, where a group of five players will likely only have 4-10 classes between, a group of five players between casters and half casters will likely have dozens of spells in use, reaching up to hundreds if you get up the higher tiers. So there's a lot potential to add spells without displacing other spells or adding bloat. Especially as there are large categories of empty area that have nothing. Look at how SCAG added melee cantrips, something new. That's a great add.
But if a source went and made 20 variations of Fireball, all 3rd level, Fire damage with the same radius, but they all did less damage - that wouldn't add much variation to the game in practical play. If instead the average damage was all over the place between the variations it still wouldn't add much variation at the table - characters gaining would take the one with the most damage. So even with spells, which are a category that normally adds to the game, bad choices such as spells that are very similar to existing spells but over or under powered don't add variety.
I also don't see how you reconcile your argument with the fact that the class list, and within that the lists of class abilities, has changed and (generally) grown over the history of D&D. The introduction of paladins, rangers and barbarians almost certainly has reduced the number of fighters who get played. Just as the introduction of warlocks, sorcerers and druids probably reduces the number of MUs/wizard played. Is that an argument against the introduction of those classes? Is there some virtue in playing the "original" stuff?
When you have Cleric, Fighter, Magic User, Thief, Dwarf, Elf, and Halfling as your classes, and five players, there's not just room for additional variation, there's a need for it. Compare to 4e, where you have 22 from just PHB1, PHB2 and PHB3, plus the various other books, and then the Essentials versions of the classes. You've reached a point where with a five person group doing a few campaigns over the lifetime of the edition, you still will have some unplayed.
I assume you'll agree that when the PHB1 was the only 4e PHB out, the classes in there got more play in any particular party than later when all of the various classes were out. My point is that it is not just dilution, where 8 classes get 1/8th of the play and 22 classes would get 1/22nd of the pay, but to a degree replacement. If some of those 22 classes were in 10% of parties, and some were only in 1% or 0.1% of parties, you've had options that were either introduced as bloat, or were an earlier class that has since been superseded.
I also don't see that the notion of opportunity cost is salient here. In my main 4e game no one played a rogue. But the guy who played a ranger might have played a rouge instead had there been no ranger option. But it's not a cost to our table (opportunity or otherwise) not to see a rogue in play. Just as it's not any sort of cost that in all my years of AD&D play I've never seen anyone memorise Affect Normal Fires.
op·por·tu·ni·ty cost
noun
Economics
noun:
opportunity cost; plural noun:
opportunity costs
- the loss of potential gain from other alternatives when one alternative is chosen.
Your situation is a textbook example of an opportunity cost. One less rogue was played because a ranger was played. Since the cost we are talking about is not one levied against your table, but rather a note about potential bloat, it's perfectly normal your table did not notice paying a cost - it didn't. But in the case of ranger or rogue I think all of the current base classes of 5e are worth including. I do worry about proliferation of subclasses, so examining that at your table may be more rewarding in understanding my point.
For example, with the exception of lowkey13, I think the majority would say Paladin was a good addition to the class list and often played. But how many Oath of the Crown Paladins (SCAG) have you seen played? How many Oath of Redemtion paladins have you see at actual (or virtual) tables? Or the flip side - when there were only Totem and Berserker barbarians, we at least saw a few Berserkers. But with the addition of additional options for barbarians in later books, we don't even see those Berserkers (unless they have house rules to bolster them).