Of course it does, unfortunately. Or at least I guess it does. Message filters work based on a pre-written scheme, filtering unwanted messages only work if the message has an exact match with the filter parameters. Generating a random number is a smart idea (if we can name something smart that intends to harm people), because there is never an exact match.
Except if the scheme is something like this: Remove all commas and invisible characters from the message and then try finding the unwanted text (in this case the website). if there is a match, the numbers and other text don't have to be taken into consideration. And if anyone gets locked out unintentionally (for example bad filtering or they mentioned the website like "*website* is fake!"), support can review the case and re-enable the account. Also, this method cannot be bypassed, not even with randomly generated numbers or separated text. Drawback: every website to be filtered has to be added to the system manually.
But I still think a world chat limit raise would be the total winner.