I would support your idea.

I take the opportuniy to add one anti bot idea here (again):

How would a bot/tool handle it if each player could only use (let's say) 10 randomly selected percent of their cards per match and then had to spontaneously form the best possible team from these available cards?
Would the software still be superior? It would not be able to use databases (it would not even know which cards would be available to the opponent, so it would also find it difficult to carry out a simulation) ...
Of course, randomness would then play a much bigger role, but good players would still form much better teams on average from the randomly selected cards than weaker players.
And investing in many cards would still be worthwhile insofar as 10 % of a large card collection means absolutely still significantly more available cards per match than 10 % of a smaller collection.
I would exclude the summoners from the card reduction to keep RNG low enough.

This idea might not be suitable for lower leagues where many players don't own huge card collections.
But in higher leagues (and with all card editions counting) I think that could be interesting.

Yes, there might be some flaws in this version of the concept, but I guess the idea in general might offer some food for thoughts ...