AI

[TinCow]

Part two of the RPS interview is now up, and it includes a heavier focus on the AI than I've seen elsewhere. It's still pretty sketchy, but worth a read for those of us who tend to obsess about the TW AI (which is probably everyone on the Org).

http://www.rockpapershotgun.com/2012...e-total-war-2/

The main nugget of information that stood out for me in that was that the cities layouts were designed with the BAI in mind. That pleases me, but at the same time I find it hard to believe that wasn't true in Shogun 2 as well... and we all know how well the TWS2 BAI handled sieges...

[rickinator9]

The main nugget of information that stood out for me in that was that the cities layouts were designed with the BAI in mind.

I hope they will actually have cities. I didn't really like the fact that you just had a fortress in Shogun 2

[quadalpha]

Dubious about cities. There's no room to manoeuvre, not much scope for missile units, not much scope for flanking, and generally descends into a brawl. Not sure the addition of multiple capture points will help much in that regard.

[Vuk]

I don't understand why they don't just make it your objective to first breech/get into the enemy walls and to destroy the enemy army. Just get rid of the stupid time limit and make it where the only way the defender can win is to rout or annihilate the opposing force. It would be so much simpler and more realistic.

[quadalpha]

I don't understand why they don't just make it your objective to first breech/get into the enemy walls and to destroy the enemy army. Just get rid of the stupid time limit and make it where the only way the defender can win is to rout or annihilate the opposing force. It would be so much simpler and more realistic.

Yes, the way the siege battle resolutions have been set up now is unrealistic and uninteresting. A better way might be to give the defenders an objective to break out (and so reset the siege timer on the campaign map) even if they are unable to destroy the besieging army, and have the option of surrendering (the army retreats to friendly territory but is held immobile for x turns, kind of like battle survivors in MTW). The besieging army might have secondary objectives for raiding supplies (thus shortening the siege timer), perhaps with only a fraction of the their army, and the defenders can decide whether to try to prevent that or not. All this should encourage more battles around settlements rather than in them, and make sieges more interesting that a binary choice between sitting around and all-out attack.

[Seyavash]

I don't understand why they don't just make it your objective to first breech/get into the enemy walls and to destroy the enemy army. Just get rid of the stupid time limit and make it where the only way the defender can win is to rout or annihilate the opposing force. It would be so much simpler and more realistic.

the timer is not entirely unrealistic. battles tended to end with the loss of daylight. Especially in seiges where there could be many engagements before a city fell or the beseiger gave up. enhanced surrender and sally options would be better.

[quadalpha]

Battles always ended with daylight, right? I can't think of any major engagements at night. Maybe something in the Sicilian expedition?

[Vuk]

the timer is not entirely unrealistic. battles tended to end with the loss of daylight. Especially in seiges where there could be many engagements before a city fell or the beseiger gave up. enhanced surrender and sally options would be better.

So what? The timer is only an hour!

[quadalpha]

So what? The timer is only an hour!

The soldiers got tired after an hour. Or maybe they were only hired for an hour and the unions would demand overtime pay for more time?

[Moros]

So what? The timer is only an hour!

If you want a full day of battle, do you also want 6 months of besieging afore the assault?

[Vuk]

If you want a full day of battle, do you also want 6 months of besieging afore the assault?

Yes. If you don't want a full day battle, you should win decisively and quickly.

[easytarget]

Vuk, ignores the balance between realism and fun and was henceforth ignored when it came to discussions about timers.

[Vuk]

Vuk, ignores the balance between realism and fun and was henceforth ignored when it came to discussions about timers.

Of course that is what it is, and not that you are a fool. I am not arguing about kill rates and unit sizes, just freedom to do what you want with what you have.
With the types of units and kill rates they have, it would be impossible for a siege to last past an hour or two. All I am saying is don't use the timer.
For 90% of people in 90% of battles, there would be no difference at all.

[ReluctantSamurai]

The soldiers got tired after an hour.

In the original STW, the player had the option of playing without a timer...which is my preference. I hated losing battles because some broken unit with a small handful of soldiers in it could hide deep in a patch of woods somewhere and wait out the clock.

Fatigue was well handled in STW. As the battle wore on, it became increasingly more important to rest your troops or they would not perform well. It was not unusual for me to have epic battles that lasted as long as 1 1/2 hours (real time) when fighting multiple stacks. At the end, whether the victor or loser, I felt like I had been in the battle myself....tired.

[Ibn-Khaldun]

You can always turn the timer off. You do know that?

[Martok]

You can always turn the timer off. You do know that?

As long as this is the case, I will be content.

[ReluctantSamurai]

You can always turn the timer off. You do know that?

Yes, I do. As long as that option remains, I'm a happy camper

[Myth]

City assaults are a messy affair and usually a gamble for the attacker. Sieges were more prominent by a large margin during the ancient times, and by an overwhelming one during the Middle Ages. If archers could shoot from the top of tall buildings, and in-city guerilla fighting was possible, if holding/destroying choke points like bridges was possible, if wooden buildings were flammable - then the chaos of an urban battle could be better represented and city assaults would not be as boring.

Taking a well defended city takes time, especially to rout out the defending loyalists. Starting the rape&pillage SPA and recreation program before you've taken out the enemies and crippled the civilians could see you butchered in the city you came to conquer.

[GFX707]

I can't count the number of times, in my experience with TW games, that the AI has got stuck assaulting my city/fortress and I have had to wait for the timer to run out to win the battle. I have many memories of the AI in M2TW standing just out of missile range outside of my walls shouting "charge, halt, charge, halt" (or something) ad nauseum because they were bugged and the whole enemy AI had frozen. Either that, or the whole AI army would just stand at the same spot that they deployed in and do absolutely nothing. Without the timer, I would have to replay the battle and possibly throw away a victory where (like many times) I managed to snatch victory against crazy odds against a Mongol Horde. Also, replaying the same battle until the AI isn't broken in a CA game....well, you could be replaying it a good few times.

So, I would like the option of a timer, but also the option to have no timer.

[Kocmoc]

Well, playing the AI was always a mess, too easy, same like the whole SP campaign…
The problem is always the same, the AI cant fight well in battles. You simple outplay the AI all the time with a much smaller army.

About the town/city war. The engine doesnt really allow any good "battle" in cities.
You would need a completely different engine to play that any good and maybe a little bit realistic.
Do anyone of us think, that units would keep in like while fighting in a town? probably not. They would split up in many small groups.

In the end, you need something I asked since years… splitting units. Also you would need a ton of different bonuses. I would like to see some proper defensive bonuses,
something this game had back in STW-MTW with the hillbonuses. There should be a heavy defensive bonus for holding houses.

If you want you can go and test the old castle defensive bonuses in STW/MI. Maybe not this strong, but thats the right direction. You would also want to see some real zone map.
Imagine a map with a castle in the center and many house (the city) around it, divided into 8-16 zones. The defender win, if he hold 6 or 10 of the zones, while the attacker would need
to take all zones to cut the defensive bonuses of the castle.

This would play around a little bit with the stringent bonuses we have right now.


All in all, you need 3 different kind of battles:

1. Standard land battle
2. the castle/city battle
3. naval battles

I wrote about this in another forum back at the start of S2.
some short things:

The current system, more or less, take the engine from the landbattles and transfer it to the naval battles.
The cannon ship, is just a bit cannon, without hills and trees. Same goes with archer or the melee units… Its eyecandy, almost pointless, since the naval battles are based on a system,
which cant really work.
Right now its only about, who can fire more and focus on same targets.


Naval battles will shine if you introduce stuff like this:

1. Fog of war - this would change a lot
2. Wind - hard effect of movement
3. bigger maps
4. Area of waves - big waves reduce firing range



About city fight I wrote already. I dont think you ever will see any good AI or fighting in towns or castles with the current engine/system.
I think the zone fights could solve a lot of the problems.


Landbattles. Well, thats what the engine was made for. Apart of some very annoying problems, like the OOS problem ( I wont say bug)
and units not shooting or all kind of movement problems. Not to mention the cav which just runs trough your units straight into your back…

The landbattles are okay. Too many different kind of units, too many different ranges of influence and effect, too many different abilities,
which you can stack. No way that any AI will ever stack abilities and do a nice movement.



All in all, the AI will suck as always. Thats based on the system itself, the engine and the endless effects and bonuses in the game.
If they would reduce all the different things and keep it simple, the AI would be a lot better.
While a human can decide in less than a sec what he takes out of 200 possible ways, the AI never can do that.
But the AI has one advance, thats the "CPM", the AI actual could do a heck of a movement, since he could sent out tons of
different orders each second.

But look at the game… CA focus more on eyecandy, than on the AI or gameplay. You are supposed to buy the game once again and than buy the addon
and than the next game and than the addon. The game hasnt to be perfect, just good enough till the next thing comes out.


Koc