There's just too much that skews any evaluation of flaws (and also virtues, for that matter) to make a proper judgement. Trying to evaluate the flaws out of context just won't gauge their value properly.
I find myself disagreeing with most of the OP classifications because of that. Not because they are wrong, but because they only consider a particular set of conditions to make the evaluation.
Also, let's not forget that to some degree, every flaw is a story flaw. To make a case about Flawed Parma Magica, for example. Agreed, if we consider it as a whole and treat the world as a probabilistic experience where your chance of being confronted with an Ignem effect is the same as being confronted with a Corpus one, it's on the weaker side. But for a lab rat Bonisagus that is never subjected to external effects, or a Jerbiton more concerned in mingling with nobility who will never be subjected to an Animal efect (for example), the flaw is hardly a problem anyway. So, why did the SG even let the player pick this flaw for the character?
For a hoplite who constantly goes out fighting hedge wizards the flaw is harsher. For one with an Enemy who excels on the kind of Form he is weak against, this is a huge problem.
We begin to enter the game's social contract. Chosing a flaw (any flaw, not only story ones) means I want to face a certain kind of story. Part of the SG's buden is to be upfront and say "the particular combination of V&F you are chosing makes so that flaw XXX will be unbalanced in this saga:"
it will be too harsh. Want to change, or do we find a way of mitigating it?
it will be harsher than you'd usually expect. Is that ok?
it will be weaker than average. Is that the intent? If so, no problem
it's just too weak to be worth a flaw. Chose another one, or change other V&F so that this actually becomes a flaw for you.
Also, a few flaws are as they are exactly because you usually have ways to deal with them. Weak Enchanter, for example, is horrible for a magus. But any magus can just subcontract the poduction of items, or get by with weaker ones, and that's what makes it Minor. For a Verditius, however? It could very well be Major.
Again, the SG needs to keep an eye open for that and make adjustments as required for the particularities of the saga and characters.
That's a very good point. And the SG should go out of their way to make every flaw matter. The Jerbiton may not come against Animàl magic very often, and that's fine, not very often is appropriate for a minor flaw, but every once in a while the Jerbiton is going to travel, and there, in the wilderness between covenants and town, there be random enemies, and one of them just could turn into a story, and because the player has designed the magus with weak parma magica, one of those rare potential animàl threats becomes real. By choosing the flaw the player has explicitly requested the challenge.
I think the value of flaws (and virtues) is partially subjective. A harsh flaw from my pov takes away more value than the point or points I receive. Value is subjective.
For example, I value characters being able to function as expected. As a consequence, I loathe fumble mechanics. I view flaws that increase the likelihood of botching as being more harsh than their cost, although other players see these flaws as reasonable, and still other players see them as gimmes because they love the unpredictability of fumbles.
Where I can compare the relative values of Affinity vs Puissant vs Good Parens with reasonable objectivity, I don't know how to compare any of these with Cautious Sorcerer. Twilight Prone vs Blatant Magic? Um.
I also suspect that ideas like "it is the job of the SG to make flaws appropriately painful" have aged past their expiration date. Perhaps even the entire late 80s implementation of advantages/disadvatages (which AM5 retains) deserves an honorable retirement.
Are those two statements related? The first idea, of adapting the story to the virtues and flaws, seems to me to only really appear in 5ed, which thus takes a good step away from the objective and mechanical 80s implementation. I very much agree with such an honourable retirement, which is why I don't think one should tweak ArM for a 6ed, but rather start over with something fresh; The first statement is really just a patch, to make the 80s mechanics playable with the range of stories Ars Magica is meant to support..
One also has to remember that 5E did in fact, for what it's worth, had 3 or 4 rounds of playtesting. How many RPGs get that at all? Anyone play Exalted 3E? Now there's a game that couldn't be playtested properly because of leaks online and it SHOWS. Crafting charms anyone? Only 3rd of the written material for that section is mathematically worthwhile to take.
Another consideration is that the initial 5e core was a reaction 4e. Over time, the Ars community have had nearly 20 years to play these and get a new sense of how these newly scored flaws compare among each other, rather than to the previous 4e balance issues.
I think this is a very worthwhile endeavor, and have noticed myself over the years some of the Hermetic flaws struck me as not that well scored/balance for what they are. But this was only noticeable after years of play.
A reasonable storyguide with reasonable players can always negotiate the severity of flaws with the players. All you have to do is come to an agreement that player A's flaws are a bit on the weak side and she needs to take one more minor flaw player B's flaws are a bit on the harsh side and he is fine with nine flaws.
Alternatively, and in my eyes, more attractively, the player is randomly (e.g. by rolling dice) assigned two flaws and two virtues by the storyguide. The others can be chosen as usual. This reflects that flaws and strengths are not chosen in real life, but you are stuck with them. It would mean that you'd have to mold your character to fit some given limitations and strengths, and would make character generation more challenging, more realistic and yield more interesting results.