Personal experience obviously
. I've always thought of Peach having "good" grab range and of Wario having "very good" grab range.
...I wouldn't come close to describing her grab range as "broken" though. Maybe if you just look at the animation yeah, but the actual grab hitbox falls a bit short of it.
Back to Wario's grab. I did some brief testing and it beats out Peach's, Falco's, Ness', Snake's, and ties/barely loses to Kirby's..... all of which are above Wario on the list. Now this isn't the best way to test raw grab range, as character's differing hurtboxes come into play too much here, but clearly something is wrong with this data.
let's look at the method he used to get this data.....
Hm, there is a slight complication with using the lines on the blocks to measure distance. The lines are on a different plane on the z-axis than the characters, so as the camera angle changes the perceived relation between the character and the line will change. It doesn't sound like he accounted for this.... but perhaps he did.
There's also the problem of judging distance just by looking at something, which will always inject human error into the process. A (hopefully) minor issue, but still an issue nonetheless.
This could be a bit problematic as well. For one Mario moves around a lot during his idle animation. Due to Brawl's "dynamic" hurtboxes this means that his hurtbox is moving around as well, which could skew results. If a good deal of trials were done with each grab this most probably wouldn't be a crippling issue however.
A more pressing issue is Mario's shape. Do his arms have hurtboxes? If so that means that there will be a bias toward grabs that hit mid, because grabs of that height will contact the hurtbox of his arms while grabs that hit high will go over it. If his arms don't have hurtboxes then the bias would be slightly toward grabs that hit high due to his nose sticking out further than the rest of his body.
Mario isn't really a good test subject for this. Ideally you would want someone who doesn't move around a lot and who has a very linear hurtbox. Zelda immediately comes to mind as a possible candidate.
Roh-ruh. This is a big problem. "Range" is measured from the furthest reaching hurtbox of the character in idle animation before the attack to the furthest reaching hitbox of the attack during the attack. It isn't measured from a completely arbitrary point like someone's "arm". Because of this this data doesn't even truly represent the range of the grabs. It represents the range they have when measured from an arbitrary point.
So yeah. There are quite a few problems with the testing methods this guy used :X. Testing like this would be good enough to get a general idea of grab range.... but nothing incredibly specific. As such I really wouldn't give too much heed to it.
hm. There are some big differences between this list and the other one. This guy didn't post his method for getting that information either :/.