If you ask Mark Rosewater what the strongest features of Magic: the Gathering are, aside from the general idea of a Trading Card Game, he'll tell you they're the color pie and the mana system. Many people balk at that, complaining about Lands and mana flood or screw. Mark goes on to explain that they introduce a controllable form of variance that prevents games from flowing the same way, and that splitting available effects between colors forces players to accept having weaknesses or introduce variance to shore them up. I think he's more or less right in that reasoning, but the actual strongest feature of the game is a bit downstream of those first-order effects. Rather, Magic's greatest strength is Limited - or, more specifically, that it designs each set for Limited.
Card games have two competing problems they need to solve, and they are always in tension. A game needs to release new cards at a steadyish clip, but needs to avoid a situation where new cards flatly outcompete everything in the environment or decks start to become undifferentiated lumps. Games also need to have a way to deal with older cards that are over the line - either a rotation policy so they become irrelevant for most competitive play after a certain period or a robust Ban/Restrict cycle. Rotation is very controversial in games that don't already have it and Ban/Restrict relies on the publishers being responsive.
Non-Magic games deal with these issues in a few ways. The non-Magic game I'm most familiar with - Digimon - handles this via archetypes: cards are made for specific decks that are more or less predesigned, and new cards come for a few sets after that to shore up holes or add alternate play patterns. This essentially stops deck centralization by fiat, and makes rotation a nonissue as power creep and meta share is instead dictated by what archetypes get new cards, with notable outliers handled via ban/restrict. This has some major failure states though; there is very little space for expression in deckbuilding, an archetype needs to wait for explicit support to be published, that support may include a rework or complete shift in the archetype's focus, and there's a perverse incentive not to ban or restrict support for new archetypes.1 Also, outside a couple specific parts of the game (generally-useful Digi-Egg cards, color-based utility Options, and a small number of color-based "engines" of high-synergy cards, mostly in Purple), color is basically vestigial, and notably each of those categories has resulted in restrictions due to competitive ubiquity.
Magic deals with all of this - making rotation bearable and color relevant but not centralizing, avoiding soupiness and power creep, making deck expression more prominent and support less prescribed - by designing for Limited, and Limited only works because of the clear differentiation of the color pie and the variable color access created by the Land system. Every set is its own self-contained environment, and cards are built to be broadly applicable accross a number of strategies that emphasize different aspects of the cards. And because most aspects of the cards and strategies exist across sets, deckbuilding is less about matching keywords on cards than it is about finding cards that synergize well.2 Rotation isn't as much of a problem in Magic because, since sets are built to stand on their own and the game in general is less "directed", losing out on older cards doesn't create horrible gaps in the card pool.
Limited in Digimon meanwhile is a terrible mess; color is partially ignored, the games almost always boil down to what bombs you draw, and most common stuff has a ton of completely irrelevant text that won't help you since you're by necessity playing a muddle of different archetypes with different game plans. Frankly, I don't know how it could be salvaged; even very early Digimon had light archetypes based on card name because it's thematic to play specific digivolution lines together, and after approximately BT10 or so it's gotten progressively tighter. It's just not a Limited-workable game at this point without a reboot. But that means Digimon has to live with being so prescriptive, and all the attendant downsides.
1↑ To be clear, this perverse incentive exists for rotational games too - publishers don't want to remove a sale point from their products, and "contains a very strong new card" is definitely a sale point. But rotational games are better able to suppress power creep (by not always having the strongest version of an effect available) and are less likely to kill a deck outright by banning one card.
2↑ Not that there aren't linear strategies in Magic, as Mark calls them; creature-type based decks are popular after all, for an easy example. But those can usually be supported by surrounding sets with very little cost and without putting a huge "only use here" sign on them the way an archetype-based game would.
