Thank you for your efforts, Arkonor. Here are some comments for you to consider:
A blocking target produces a temporary hitbox (it lasts for the duration the block is held and I don't know a better term, so I'll use that for now) at a position different than the target's normal hitbox. This can be seen most clearly with a thrust that barely reaches the temporary hitbox of a blocking target. If the target stops blocking and another thrust is made along an identical path, then that thrust will hit nothing. Presumably blocking works by generating a temporary hitbox that is between the normal hitbox of a blocking target and an attack traveling along a path that would otherwise reach the normal hitbox of the target. I would assume the different blocking directions produce temporary hitboxes at different locations relative to the attacker and target. I do not know how far forward and up the temporary hitbox for an up block is relative to a down block. Do you know far from the normal hitboxes these temporary hitboxes are located?
Also, aiming a thrust or side swing up or down decrease the horizontal distance the weapon will travel away from the attacker. For example, thrusts straight up into the air or straight into the ground travels no horizontal distance away. This is not the case for overhead attacks which do not have a dependence on the polar angle that the attacker is aimed at, so looking up or down does not effect the path a overhead attack swing with travel. I suspect that side swings also do not travel such that the maximum reach is constant (their tip does not follow a circular path.) Did you control your tests for this?
I suspect the above reasons are why Stabbing Hobo used a practice dummy as the target for all of his test and why he used an overhead attack as the reference for all other attacks.
- StabbingHobo testing reach when a dummy is hit
- Arkonob testing reach when an attack is blocked
This, more or less.
