The magnificent 600 ashigaru - or why unit upgrades matter

This incident from my legendary Mori campaign seems worth a mention.

I was facing a big revolt in Suruga with a three quarter full stack of yari and bow ashigaru menacing the castle and sent 3 units of yari ashigaru from Sagami west towards the bridge aiming to have them ready next turn to reinforce my garrison

But instead the rebels attacked the 3 YAs.

Normally I autoresolve such hopeless battles but as I was going to have to fight the same army again that turn decided to play this one out and do as much damage as possible before being exterminated.

There was a steepish wooded hill so I put my 3 YAs at the top in shield wall and awaited certain doom expecting the 4 rebel archers to shoot most of them down and then their 10 YAs and general to mob me.

But thanks to my units +2 armour, the shieldwall and the hill the archers did very little damage before the AI sent its own YAs in - but made no attempt to flank me just sent in 3 or 4 units in in three separate waves so it was three more or less one-to-one fights.

The +3 weaponsmith attack bonus and my YAs +4 experience bonus made easy work of the rebels bog standard YAs and when the last wave broke and fled down the hill I turned off shieldwall and charged down into the rebel archers and their general (the only intelligent thing the AI did was not send him in first to commit suicide on my spears all of which routed in seconds.

So the end result is that thanks to upgrades and a convenient hill a battle I had no chance at all of winning on autoresolve ended with my 600 YAs defeating 2520 rebels and killing more than half of them while suffering only a few casualties themselves.

Attachment 5127

Despite which it didn't count as a heroic victory and I got no landmark...

It has fairly little to do with upgrades and stuff but rather with any geographical features on any map completely breaking the AI to a point where it would probably forget to breathe if it had to make a halfway important tactical decision.

-E- To elaborate a little: Hills break AIs.

The hill definitely was a crucial factor (as was the spearwall - although I am not sure what that actually does to the stats) and the AI was unusually stupid - pity I didn't take any screenshots of the battle itself.

On a flat plain the AI's arrows would almost certainly have had more effect and their infantry attack would have been faster and not have ended up in three waves that could be driven off separately so I doubt my 600 YAs could have survived.

But equally if they hadn't had any upgrades at all I don't think the hill alone would have saved them from 4:1 odds - certainly I've fought early game battles with vanilla ashigaru and no general and been defeated in similar situations.

And +2 to armour, +3 to melee attack and +4 to experience certainly wouldn't have helped as much against melee samurai, monk or ronin units but look at the stats:

The vanilla YAs had a pitiful morale 6, attack 4, defence 4 and armour 2

My upgraded YAs had morale 10, attack 12, defence 7 and armour 4 (as another point of comparison vanilla yari samurai only have 11, 9, 9, 5 and are three times more expensive to recruit and twice as expensive to maintain)

So man-for-man they had triple the attack, double the armour and nearly double the morale and defence of their opponents.

It also helps settle for me the debate over whether to have a master weaponsmith or master armourer: +5 armour might have reduced their casualties but when as the AI usually does it splits its infantry attacks into two or more waves rather than extending its line out to outflank a faster kill rate seems preferable.

It's kind of interesting the AI doesn't do basic things like outflanking a dug in position on a hill or fire all its arrows before moving any melee infantry. It can't be that hard to do right?

Not really - that it does what it does in a way that still represents a challenge for human players is actually rather remarkable and the result of a huge and very expensive development process.

And terrain is always an issue for an AI that has to cope with so many variables.

I am not even sure that this example is all that unrealistic - history is full of battles where much bigger armies were defeated because they regarded their opponents with such contempt that they abandoned all the niceties of maneuver.

And all the AI really did was the equivalent of firing some arrows, noticing they were having little effect (each volley only killed a couple of men in each of my units) and then selecting all their infantry and clicking on my central unit to charge - the sort of thing I'd do all the time if I didn't have an easy auto-resolve option and had to fight such one-sided battles myself.

The problem is, the AI does it all the time, wether it has 3 units or 300.