Yeah, I hear that and I can see why you wouldn't want to make it unlimited (e.g. a player could surreptitiously craft exactly the kingdom they want to play every time). While I can't definitively speak to how abusable upping the limit would be, it could make the game much more fun for many and it certainly feels like it shouldn't impact the ladder too much.
A player who really wanted to ban their worst cards would have a bit of a chicken and egg problem - how do they know they're their worst cards unless they play with them consistently? What if, after improving more, other cards would actually be their worst cards? I suspect a player trying to micro-optimize like that is wasting their time and trying to prevent it is a similar waste. The overwhelming majority of players just want to put the cards they hate there and move on. Plus I'd be surprised if there was meaningful variance in per-card winrates at the top of the ladder - those players tend to be super good at everything (and also less concerned about specific cards because they understand the general principles). But if all the best players want to ban, for example, the swingiest cards, all that means is the ladder ratings end up being more accurate faster, which isn't a bad thing.