A fork that requires 75% of the last 1000 blocks will be over pretty quickly. The <25% jump ship to the main fork or are stuck mining worthless coins at a huge difficulty for a long time before adjustment. If they survive that... they are at a constant risk of being attacked by the greater power of the 75% hashers. This is how Bitcoin works... it's there in the whitepaper.
You don't know how Bitcoin variance works.
Much less than an actual 75% of mining power
But as that variance is smoothed out over time by reversion to normal distribution, the >25% remaining miners will inevitably orphan the attacking Gavinista blockchain by solving several consecutive 1MB blocks.
Here is your next remedial lesson on Bitcoin Civil Wars:
Hearn picked the least optimal 'magic number' with "75%" just like he went about creating XT in the least optimal way possible (Great Schism).
Cite: http://organofcorti.blogspot.com/2015/08/bip101-implementation-flaws.html
From the POV of Core defense, XT's idiotic choice of "75%" is great, as it minimizes XT's chances of success!
Please hurry and catch up to the rest of the class. You are (at minimum) 4.5 months behind.
So... 750/1000 may be us just getting lucky...