Artificial Intelligence
"No matter how exotic human civilization becomes, no matter the developments of life and society, nor the complexity of the machine/human interface, there always come interludes of lonely power when the course of humankind, the very future of humankind, depends upon the relatively simple actions of single individuals."
-Dune
"Well, look at this! Appears we got here just in the nick of time. What does that make us, Zoe?"
"Big damn heroes, sir."
"Ain't we just!"
-Firefly
This is possibly the most difficult aspect of the game for me to posit useful suggestions about- not because I don't have any ideas, oh, on the contrary I do- but because I really feel like the devs and fans have often been talking at cross-purposes here. So allow me to get one thing straight:
WE ARE TALKING ABOUT A GAME WHERE ALL YOUR UNITS ARE AUTONOMOUS! OF COURSE AI IS GOING TO BE IMPORTANT!! SWEET JESUS, BUDDHA, MOHAMMED AND SPONGEBOB, DOES THIS REALLY HAVE TO BE EVEN SAID?!
There is
so much either outright wrong with, or disappointingly missing from, Majesty 2's AI that I literally don't know where to begin, but the real problem is that those visible missteps could be down to any one of several possible causes.
1- A desire to make the game arbitrarily harder for the player by piling on tedious busywork.
2- A desire to inject direct control by compromising hero autonomy where there's no good explanation for it.
3- Sheer ineptitude.
Moreover, some- and maybe all- of these issues might just possibly be redressed in subsequent patches or expansions, so for any given suggestion or critique I might offer, I feel like the developers could have several ready rebuttals at hand.
"Yes, we know, and we're already working on a fix for that."
"Yeah, we considered that possibility already, but found that it interfered with our prior assumptions about gameplay. For example, monsters attack the town constantly, so if heroes only went shopping at appropriate, sane, times, they'd basically never have the chance to buy better gear."
"Alfryd is a jackass."
"We... don't really see what the big deal is? Why would you want a game that plays itself/ why would you care about healers planting flowers/ why would you even want to keep playing after the mission ends?"
I could spool off pages upon pages of suggestions on the subject of improving hero AI, but most of it would be incomprehensible to those who can't program while being largely obvious to those who can, and the real problem here
isn't technical- it's that the developers have been working under fundamentally misguided assumptions. It's these assumptions that are the real problem here. Until those are corrected, there is going to be a constant and irreconcilable friction between the intent of the designers and the natural consequences of the gameplay, and the end result will find itself wedged awkwardly between the rock of RTS and the hard place of Sim.
...Basically, all I can really do here is posit some alternative assumptions.
ASSUMPTION 1: CHARACTERS SHOULD CREATE THE ROUGH PSYCHOLOGICAL ILLUSION OF BEING REAL PEOPLE.
Everything else really follows from this. The player- and thus, the designers- should be able to put themselves in the characters' shoes. Follow them around, watch their behaviour, and never run into a situation where the character does something so obviously, dumbfoundingly uncharacteristic that there's no way to rationalise it on the basis of in-game information.
Real people do things for personal reasons. Even if it's a stupid, selfish, short-sighted reason, it's still a reason- a discernible motive. They don't give a
damn about the player's pre-supposed real-world agenda- hell, they don't even know it
exists!- and it should therefore never directly impact their behaviour.
If X is a thing a character can do in the game, then the designers need to ask themselves 2 questions.
1. Can the character do X without the player's permission? -If not, why? Is it illegal? Morally repugnant? Physically impossible? 'That would be tactically inconvenient for the player' is not an adequate explanation.
2. Must the character do X when the player says so? -If so, why? What's the compulsion being applied? Blackmail? Mind control? Devotion to Duty? 'It's more convenient for the player' is not an adequate excuse.
X can be anything- leaving a party, joining a party, joining a new guild, running away, learning new skills, leaving the map, entering the map- anything. You need to know what the character is doing and
why, in personal, vicarious, in-game terms, at all times.
What the character thinks is more important than what the player wants.
ASSUMPTION 2: CHARACTERS SHOULD AT LEAST OCCASIONALLY BEHAVE LIKE GENUINE HEROES.
If a given class is all about selfless devotion to a higher cause, then they need to bloody well act like it. Logicians and Palatines don't
care about money, they care about serving the Greater Good- Money might be an incidental means to that end, but it might well not be. They need to give some priority to doing things
purely because The Needs of the Many Outweigh the Needs of the Few.
But almost
every character should have moments of altruism. Leaving defenceless innocents to be slaughtered en masse by marauding raiders while you race to collect a reward for exploration isn't the behaviour of a hero- That's the behaviour of a
psychopath.
More generally, all characters should indulge in activities for reasons of their own- even if they have no immediate practical value. Praying at their temple, collecting spice, holding conversations (possibly a method for spontaneous party formation(?)), carousing at the cantina, etc. etc. etc.
Also, bear in mind that things heroes
don't do can be just as revealing. In Majesty 1, paladins would never use poison, monks never visited the marketplace, and healers never competed in tournaments. That made a statement about their basic personalities- possibly a frustrating one, but there are ways to work around that.
Obviously, I'm not saying that characters shouldn't prioritise. Obviously, even the staunchest Logician might see the practical value in a sufficiently astronomical reward. Obviously, Initiates shouldn't be praying in seclusion while the town is under siege.
Obviously, heroes shouldn't go shopping at the market while their house is on fire.
...But little behavioural details like this can
really help to bring the characters to life.
ASSUMPTION 3: CHARACTERS SHOULD, IN PRINCIPLE, DEMAND AS LITTLE BABYSITTING AS POSSIBLE.
I'm not saying that genuinely stupid characters should act like they were Einstein. But
most characters aren't outright morons, and some (e.g, the Logicians, the Oracles) are
literally a physical embodiment of flawless supercognitive reasoning ability. ...They should bloody well behave accordingly.
Most characters should, by default, behave rationally, update their decisions in reaction to fresh stimuli, consider the presence, strength, and confidence of nearby allies, and retreat calmly from overwhelming danger. Panicking and berserking might occur, but only in extreme situations (or with extreme personalities.)
The key here is shift the emphasis away from micromanagement and toward large-scale, long-term strategy. I'm not saying that sovereign spells (or their equivalent,) wouldn't be important, but that their use would be pre-meditated and long-term, rather than short-term and reactive. e.g, if I sent off a party of men-at-arms to raze a native village, I could buff them with some sort of 'regeneration spell' beforehand (courtesy of the Shapers,) that would make them more capable of sustaining damage later. They'll come back
alive either way, but by increasing the amount of punishment they can take before having to retreat, I can substantially improve the odds of success in their mission. Foresight and planning, not distraction and busywork.
ASSUMPTION 4: IMMERSION OR IDENTIFICATION- NOT EVENTUAL VICTORY- IS THE PRINCIPLE SOURCE OF ENJOYMENT.
Again, usual caveat- I'm not saying that winnable side-quests, time-limits, survival gauntlets, or external objectives of some form can't feature
at all- But they should be enjoyable take-it-or-leave-it optional extras, not hurdles for the player to jump over, and certainly not the be-all and end-all of play. Look at other games in the Simulationist tradition- the Sims, SimCity, Caesar III, even Fable II- they all allow the player to just... play for the sake of playing. Because the self-contained worlds they create are
intrinsically interesting. You can't explore a world like that with a 30-minute deadline breathing down your neck.
More generally, the player should be able to
relax now and then. Take it easy! Step back and look at the big picture! Get to know his or her citizens. Play at their own pace.
Maybe even leave the keyboard, go get coffee, and come back in 15 minutes serene in the confidence that nothing outstandingly bad will have happened. Okay- an extreme example, but not an unworkable one.
Ideally, the heroes' lives are not just tools to be manipulated, they are
stories to be followed. You have
influence over those stories, but you don't unilaterally
write them. That's what makes them interesting! The very fact that you
can't simply tell these characters what to do is what makes them worth watching in the first place. But it also means that the player's personal skill and acumen are
not the direct, primary factor in determining the outcome of large-scale events. And this means that focusing on 'victory in the face of punishing obstacles' as the major payoff of play is basically misguided- because the player will never be
directly responsible for that victory! Any credit will have to be shared with your underlings. The challenge is to make those so-called 'heroes' worthy of the name.
.