The swings are different angles and distances. To calculate everything accurately it would need to be the same. On top of that his calculations are based on spam attacks and he says Ausies literally cannot parry. Spam attacks do no damage so they aren't really used so why calculate it that way? and as I said I have seen many Ausies parry consistently so his conclusion that they just have to guess and literally cannot block is false.
The swings are not different angle and distance... You could argue this for the "pride scene" since those were grabbed mid fight. The untrained eye might not see the details. However all the others were the exact same angle and distance, one showing the parry at X frames and then the next attack showing a miss (hit) at X-1 frames. He only showed 1 clip of each but we have a lot of footage and hence he says it's based on consistent results.
For teh AUS players you see that 21 frames is a miss and 22 frames is a parry and if you add certain constants / variables (reaction, input lag, etc) to that time and consider teh time it takes for an attack animation to connect, you are left the amount of time available for changing your parry. He leads with the premises that the person attack is using an insider acceleration to hit as soon as possible (45 sec animation). The conclusion is where an EU player has so much time that they can adjust their parry multiply times an aussie has to play absolutely perfect to counter the most basic of "specials" attack techniques. The quality of the attacker will highly impact the Australians ability to parry as attack animations can be longer than 45 seconds giving a lot more time to parry. However a competitive and skilled player will not let the AUS have more time to react and play around accels a lot. The video does not imply Australians cannot parry. It says that they are at a heavy disadvantage and it implies that they will have a losing match up against people with a lot less skill.
The "spam attack" here is used to describe the act of left clicking (0 dmg) and compared in the test to left clicking with a counter (high dmg). That is, it is here used to simulate people using counters by just left clicking. This lets the tester do many tests by himself and not needing heals. The idea behind this is that just the release animation without a charge up needs analysis. it is easy to react to charging weapons. but its the release by itself that can be hard to spot. Also this is with out any swing manipulation, drags, etc. When actually techniques are added it becomes near impossible for an AUS player to react parry. It is a lot about prediction and reading tendencies.
Seeing your confusion i really recommend watching the video again. Many hundred hours of work went into making extremely accurate.