My point was to be over the top, so I could point out that we are only concerned with defensive players leaving the game. Targeting increases the risk of injury. So, if there is a safer way to tackle, then having an incentive that specifically effects the player makes sense. A 15 or 20 yard penalty is a bummer, but if you know you will lose 1/12 of your season if you don't tackle properly you are more likely to actually do it right.

So, how do you determine how egregious it is? In criminal law, lots of the time the intent of the individual is used to determine how egregious the punishment should be. However, here, the rule is already tailored to be almost completely avoidable when hitting a running back (like was the case here). Just don't launch with the crown of your helmet. Plenty of other ways to tackle. So, essentially anyone who tackles that way intends to tackle that way. If it is based on how bad it looks, then bigger players will almost always get worse penalties and offenses with smaller players will usually be favored. It will incentivize faking your reaction to getting hit. It will be less objective and lead to more complaining about how egregious the hit was.

It makes much more sense to provide a strong penalty and have a narrowly tailored rule. I think they have already done both.

What if Ropati had been injured on that play, would you have been cool with a 20 yard penalty and a warning?

As for a gradient, in the OP, it doesn't make much sense to go from 20 yards and a warning to 15 yards and ejected the whole game. If somehow they found an objective way to create a gradient (which I doubt they could since intent is always present), then it would make sense to remain a 15 yard penalty and a one or two quarter ejection. Again, I don't think they should have a gradient, but you would need something like this.