Yeah the main problem with removing secondary weapon is the siege battles. Phalanx vulnerable only on rear or somewhat on flanks means defending in walls becomes so easy. Just hold the position and the towers kill the enemy. However there doesn't seem a good way to deal with this. Good shield value, medium armor, very low attack, high cohesion and the phalanx formation can still make fighting in frontally incredibly brutal.

It might not be so bad if sieges didn't last so long. In most MTW2 mods sieges of decent walled cities can last 8-12 turns before forcing a sally. I think at least half that would be more appropriate especially if there is a way to make the attacking force in a siege also lose some casualties while engaged in the siege.

I think it would be wonderful to lay siege to a city that will surrender or sally in 6 turns and have the AI attack with a relief force in 4 turns and even without fighting a battle my army has lost 20% of its strength in laying siege. (5% per turn). Winning the battle but not capturing all the enemies so the survivors retreat into town and finding now have 50% of the original army alive. 2 more turns of siege equals 40% of the original army remaining and even if gaining the city it requires reinforcements to hold or move on in the campaign.

% might be difficult to script but it might work as well to make it so a certain level of walls causes a certain # of casualties for besieger per turn based on unit size. Lower size walls might cost only 50 men on normal size where a full stack is usually just over 1,000 so slightly less than 5% but large stone walls might cost 100 men per turn. Dividing the 100 by number of units and each units size seems a bit much scripting even if possible. I'd prefer just making the number come for 3-4 random units each turn as some elite units might make a foolish attack and suffer and next turn a low tier unit eats bad food and gets sick and many die.