I don't think GP is saying we'd have difficulty encoding the cards. I think they're saying that deck performance is very sensitive to minor changes, and dependent on a lot of interactions between sets of cards.
To add to that point, decks can be good or bad based on the ecosystem (meta) they're playing in. A deck that aims to rush you down might be great in a meta where players expect slow decks, but terrible in a meta where many decks have means of healing themselves.
Exactly - the landscape of decks is extremely chaotic and I can't imagine gradient descent having any real power to discover the few towering singularities of super-exploit power-decks when changing a 3 to a 2 in the middle of the effect description of one card would render the same deck just meh. You don't really find exploits by approaching them slowly from a distance.
EDH less so though due to the bigger deck size and 1 copy per card limitation. This means you end up with bigger groups of cards with common synergies vs specific sets of cards you need to draw in regular 60-card.
To add to that point, decks can be good or bad based on the ecosystem (meta) they're playing in. A deck that aims to rush you down might be great in a meta where players expect slow decks, but terrible in a meta where many decks have means of healing themselves.