If you had 20 years to prepare, you had time to build walls, about 600 traps, train a legendary military and clad them all in steel armor and steel weapons. Basically, you should've been ready...
But nobody is going to do that. That's such a colossal, unfun, ridiculous waste of time. I don't even remember the last time I played a fort for 20 years, if ever, or even 10, really. I would NOT do ANY of those things, because the chances would be much higher that I would get bored of the fort before anything came of it, and it would just be riling myself up for nothing.
BEST case scenario, I prepared for 10 hours of gameplay for a minute and 35 second event. Uh no. Screw that. If the game were like that, I would just completely ignore defense entirely, and on the off chance goblins attacked after only 4 years, just go "oh well" and make a new fort. Because that would be way less annoying than spending hours preparing for epic goblins in 12 forts
without ever seeing one. You can probably revert to older versions
And not get any benefit from any other updates? That is not a reasonable solution.
Notice that I did not say at any point the game should not ultimately be realistically simualted. I said it should be fixed soon, or if not a high priority, then a TEMPORARY stopgap should be put in place to fix the broken side effect, until such time as it can be made to work well while still running off of the full simulation and actual historical figures. I agree that is the best ultimate goal. Ideally with parameters that control underlying things in the realistic simulation to make it, realistically, provide different kinds of gameplay that different people enjoy.
But that's not what we have now. We have an undercooked system that wasn't apparently ready yet, and it needs some bandaids until it is fully baked.
Also notice that the stopgap measure DOES allow almost everybody of diverse interests to get what they want -- if you want few invasions or more invasions, or for them to come in different general time periods, random or not random, different sizes of invasions, any of those preferences are easily accommodated. Generated goblins just for sieges would be trivial to control with a custom parameters until a better realistically driven system is in place. Thus, the solution satisfies every group of people except the ones who happen to like simulations for their own sake aside from gameplay. Which I dare say is probably a small minority.
So am I maybe a small minority too. But not me + the people who want later but reliable sieges + the people who want lots of ambushes + the people who want blah blah, everybody else who would benefit from a less capricious and broken system.