Sorry, but I don't think I agree. If the number of cards per expansion remains the same, and the number of playable decks increases, then by necessity there must be either: a) a high number of overlapping cards (i.e. cards that work in many different playable decks); or b) relatively few "key" cards for a given deck. In either case, f2p and "casuals" should come out ahead.
Of the two, my suspicion is that the latter is more likely, as option a) would seem to run counter to the point of specialized spells. I could see, though, something like a card that says "when you cast a spell, discover a spell from the same school" (or "gain a random spell from the same school"). In fact, I'd bet some card like that will be printed.
It's hard to be precise about the development of deck costs, because there are few sources that let you actually analyze this stuff. VS has archived Data Reapers, but they only started around 2016, and you usually can't see old deck lists anymore. Hearthpwn has so many decks for so many things that it's hard to track down specifc deck lists that were popular at a given time. Tempostorm has also an archive, but it's really terrible to navigate. Nobody really felt like poperly documenting the history of Hearthstone's metagame.
But anyway, I think it could be proven to a certain extent that deck costs have gone up over time. From my own experience, decks used to cost like maybe 4k, 5k, rarely more, while few modern decks stay under 6k, with many going to 8, 10 and some even beyond. I think the so called "Wallet Warrior" was about the only deck that actually went over 10k for a long time. In this meta alone, we have at least 3 decks easily crossing that line (Ramp Paladin, Highlander Priest, Quest Warlock) and you can add a few more.
And I think the biggest contributing factor was 2 class legendaries per expansion, starting in Un'goro. If you go through some old metas (as far as possible), you'll notice that the number of key legendaries was usually rather small. A lot of the legendaries in older expansions were straight up bad, including the class ones (with few expansions). The good ones were oftentimes neutral, more like "generally good", and didn't change often, like Bloodmage Thalnos, Ragnaros the Firelord, Leeroy Jenkins, Dr. Boom, Patches the Pirate etc. They were usually good to have, but rarely the driving force of a deck, or impossible to replace.
Since 2017, legendaries are frequently essential to decks. Not only, because they are both much more impactful on their own, and virtually impossible to replace, such as Sunkeeper Tarim, Jandice Barov, Lord Barov, Soulciologist Malicia, Ysera, Unleashed, Dragonbane, and half of the Deathknigh hero cards. They are also more commonly the sole reason to play a deck, either being a single win condition, or the one card that turns everything around. Quests and Highlander cards are obvious examples, but we've seen something similar recently with Nozdormu the Timeless and High Abbess Alura.
Also, more frequently in modern HS, legendaries come with with strong synergies with epics or other legendaries. Obvious cases are Highlander decks again, Galakrond decks (making Kronx Dragonhoof mandatory), but there's also examples like Sayge, Seer of Darkmoon coming with Rigged Faire Game and Occult Conjurer. Other examples are Khadgar and Power of Creation, Blastmaster Boom and Wrenchcalibur (and strongly supported by Dr. Boom, Mad Genius, Omega Devastator and Zilliax), or Tickatus with Cascading Disaster and Y'Shaarj, the Defiler. Due to high rarity cards being more impactful, class specifc and oftentimes playing very specific and important roles, it's not uncommon anymore to see them in "packages" where it's pointless to have one without the other.
I'm not quite sure why you are quoting me in this post, as I deliberately was NOT answering Kla_guy's question about the change of deck costs over time. I was (at least in my own mind) specifically only answering his question about whether the creation of spell schools will increase the cost of decks, which (for the reasons I explain above) I doubt.
I've got no definite view of whether "real" deck costs have gone up over time or not: I suspect they have. By "real" costs, I mean something other than simply dust cost. By that measure, costs have certainly increased. But the right way to measure cost is by answering the question "how much gold do I need to to play a given deck?" Since cards come from dust, dust from packs and packs from gold, the real cost of decks is gold. So to assess the real cost, you have to calculate how much dust you get from packs. But (and here's where it get really complicated) the extension of duplicate protection to all rarities makes comparing dust values now to dust values pre-protection very hard. Prior to dupe protection, if there was an epic you absolutely needed for a deck, your odds of getting it never changed, regardless of how many packs you opened. Now, with every epic you open, your odds of getting the one you want go up. Thus the "real" dust value of the pack goes up, since you will be less likely to need to spend dust to get it. So the rise in nominal dust costs is, at least partially, offset by the increased value of packs and, thus, the gold you buy the packs with.
BTW, even if it has increased, that's almost a necessary evil if you want a diverse meta. Improving class legendaries and epics is the only real way to give each class a distinct identity and viable deck. For all the complaints about the meta, IMHO there's a greater diversity of decks being played now than I can remember from back in the old days.