Engineering + construction + masonry + math is a 2,838

detour to add trebs. Maybe masonry/math are otherwise needed or gotten in tech trades, but those are the cheapies of the group. Espionage expenditure doesn't come til AFTER you have guilds, which is a whole different ballgame. Like I said, cataphracts can END pangaeas on their own, and if we're that confident in the game then the espionage expenditure is basically meaningless. Spies aren't free but they're cheaper and you need a lot fewer of them than siege.
Siege does the heavylifting in fights and reduces the rest of your army to mop-up duty. This is great if you've got trash, even archerpult can work in a pinch to help you get metal. But if you already have a strong carry unit, you're just wasting its advantage by adding siege. Cataphracts are 2-movers, a terrific bonus that doesn't matter if we're attached to 1-mover siege. To make matters even worse with cataphracts specifically, unlike normal knights, cataphracts give up their FS immunity. We can grab that back with flanking II, but that includes another 30% withdrawal chance that we'd be throwing away if only taking safe fights following siege. So instead you'd just get combat promotions on your cataphracts, which admittedly isn't bad given their higher strength. In practice though, those promotions seem wasted because the cataphracts were going to have incredibly high odds of success following siege regardless. The other GG option is putting CR on trebs. This normally isn't "bad" in an engineering attack, but is silly given that cataphracts should be the star of the show and that trebs at this point are an "old" unit compared to a normal engineering attack. I point out the inefficient GG use, because it seems like a huge flaw for an IMP leader to have and an entirely avoidable one at that. This is how you run into the trap of thinking that IMP is trash and that supermedics are the best you can do.
Ultimately getting one huge invincible stack is a very simple way to play the game, but not the best. We can do better than the AI. The risk is if you're not aggressively taking cities, you'll be too slow to run away with the game and you fall ages behind. You can have a terrific K/D ratio, but the slower you take cities and the more you heal, the more fights you have to take. Your battles become easier and easier, and then all of a sudden you just lose because you're 2 ages behind. If you're a big Henrik fan you can watch his latest Charlie game and see this exact thing happen.