AI

[TinCow]

So, what are the odds that there will be a decent AI in this one? I've pretty much given up on the TW series exclusively because of the horrible AI that makes the game boring as dirt. Is there something about the nature of a Shogun sequel that should give me hope?

[Alexander the Pretty Good]

According to their announcement statement, there will be less unit types, which might make things easier for the AI.

Of course, they'll try to expand it out with DLC because the TWCenter and .com forums are a bunch of whiners and need lots of useless shiny instead of a working game.

[A Nerd]

Limited scope? Simpler means to an end? The AI will have one objective, to conquer Japan. The AI in latter renditions seemed to be more concerned with alliances and maintaining them (not with the player). I don't know what the new diplomacy model will be, but let's hope it's more a maneurver to maintain stable borders, build coffers and manipulate warring enemies to achive eventual conquest of Japan (or whatever endgame goals are), insted of being Switzerland. The AI in STW seemed capable of this (though the game was much simpler), let's hope the new AI is able as well. THough I speak solely on strap map AI. Melee combat, one would think, would require the AI to make fewer decisions than ranged combat. But I don't really know.

[Kagemusha]

Like the Nerd says.More simple goals for Campaign AI, while lot simpler job for battle map AI as the units are less plenty and more similar.

[Gregoshi]

Just think of a STW campaign without the Hojo Horde and suicide daimyos reducing your competition to rebels.

[Thermal]

Just think of a STW campaign without the Hojo Horde and suicide daimyos reducing your competition to rebels.

Lets not forget Shimazu's horde from the west.

Alexander the pretty good makes a decent point, given that there are less units, the AI should be able to choose sensibly from the ones available.

They claim to be using Sun Tzu's art of war for the AI, doesn't mean much to me.

I think the AI will be at least ok, surely better than that of Empire's, which was a huge game with loads of units and factions, if they mess up AI on this one then they really haven't tried.

[Lemur]

Actually, the sengoku jidai period is one of the very few that makes sense for CA's Hobbesean all-against-all AI. Everybody wants to become Shogun, or be a close (preferably by marriage) ally of the dude who wins. Hmm, well, I don't know if CA can handle that second part. Allies who stick with you even when you're winning? Unthinkable.

[gollum]

It makes sense because the TW Ai was designed around this setting and was carried over into other settings that had to do very little with a unification of japan/conquer all, as classical antiquity and the medieval period.

The battle AI's performance is tied to the strategic AI's performance (is he capable to bring the right units to counter theenemy army at the right place at the right time?), by the way this is why it was a bad idea to leave the "2d" map that helped greatly the AI to bring theright units to theright place at the right time, and on various battle engine parameters and mechanics as well as the rosters. A simplified roster should very greatly help the battlefield AI.

Despite the good indications (ie that CA chose by itself to take things back to basics and against the will of the majorithy of its fanbase) this can still be messed up - however there is now the possibility to play an online multipleyer campaign against your friends in TW, and i am sure that S2 will be no exception. I have not bought Napoleon, but i am really excited to try this new option with S2 - its a format that i will definitely enjoy, as i am sure other players will give me a good run for my money.

[Intranetusa]

According to their announcement statement, there will be less unit types, which might make things easier for the AI.
Of course, they'll try to expand it out with DLC because the TWCenter and .com forums are a bunch of whiners and need lots of useless shiny instead of a working game.

I don't think it's that hard for the developers to make a decent AI, even for a game with varying units like RTW.

You have your melee infantry, melee cavalry, ranged infantry, ranged cavalry, or a combo...

The developers are just lazy...

[gollum]

And yet - rosters do incredibly affect AI capability. I am perfectly certain that a game like Empire had an AI with more potential on the battlefield. However the original's STW's AI always gives a run for your money if you are on a parity in units available and in relatively flat ground. Rome modded can get more out of the AI simply by rationalise and simplifying teh rosters.

[Myrddraal]

I don't think it's that hard for the developers to make a decent AI, even for a game with varying units like RTW.

You have your melee infantry, melee cavalry, ranged infantry, ranged cavalry, or a combo...

The developers are just lazy...

Hmm... I wouldn't be so sure about it being easy

[vartan]

I don't think it's that hard for the developers to make a decent AI, even for a game with varying units like RTW.

You have your melee infantry, melee cavalry, ranged infantry, ranged cavalry, or a combo...

The developers are just lazy...

I thought that for a microsecond or two, and then I realized (in 3 to 5 femtoseconds) that I was incorrect in thinking so. How lazy I was to think so instead of doing some homework and research on the field. Schucks.

[Thermal]

Hmm... I wouldn't be so sure about it being easy

Yeah, they have to know what to do in every situation, consider where its opponent is and combat them effectively, a lot of thought processes. Then they must consider terrain, weather, whether to keep units in reserve or not.

[Sp00n]

TBH as Shogun is on a much smaller scale, they should have way more time to concentrate on things like AI and balancing, I dont think the AI is that bad in Empire its certainly improved massively over the last 10 years and Shogun should be the easiest setting to get the AI right in.

[caravel]

I doubt that the AI will improve, though due to the smaller map and greater similarity between factions/unit rosters, there will no doubt be a perceived improvement.

[PanzerJaeger]

My guess is that it will not improve and will most likely be worse. This is not the same group of people that made the original.

[Beskar]

I doubt that the AI will improve, though due to the smaller map and greater similarity between factions/unit rosters, there will no doubt be a perceived improvement.

Unlike Empires "Everyone is a Line Infantry with a different coloured coat" ?

[Thermal]

Unlike Empires "Everyone is a Line Infantry with a different coloured coat" ?

These aren't line infantry, a simple unit roster worked on the first game, so it can work in this one.

[Kagemusha]

The unit roster should depend on what time period they start.If it is before 16th century. Samurai armies were mainly constructed of Mounted Samurais/Kachi´s armed with yumi bows and levied Ashigaru´s mainly armed with Naginata´s and Yumi bows. Sohei warrior Monks were around from early on so they should be present. If the game starts during 16th century. There should be mounted and dismounted Samurai /Kachi´s, with variety of weapons, but mainly using Yari spear. Ashigaru armed with Yari or Nagae Yari spears,Yumi bows and or Teppo arquebuses. The Sohei would still be present+ Ikko Ikki leagues consisting of peasants,ronin and Soheis were force to be reckoned with. For example Oda Nobunaga used lot of his time fighting these forces. Also at the end of the period many Samurai /Kachi´s adopted teppo as their weapon creating sort of Dragoons from themselves, that moved on horseback, but mainly fought on foot. Certain addition could be Wakou pirates that raided both Western Japan and coast of China. There could be interesting variations for their weaponry, mixing Japanese and continental weaponry. I think all and all there are lot of interesting variations for CA to consider.

[caravel]

Unlike Empires "Everyone is a Line Infantry with a different coloured coat" ?

I'm not speaking of purely cosmetic similarities. ETW was similar to MTW, RTW and M2TW in that it consisted of hundreds of faction specific units with widely differing stats based on the typical large Eurasian map, where provinces are practically kingdoms. This huge array of units makes it virtually imposible to balance the game and makes it easy for the player to exploit the inept AI factions.

The original STW is based on a smaller scale map with individual provinces rather than i.e. "Ireland", "Denmark" and "Aragon" as provinces. Unit rosters in STW are identical per faction. Factions differ only in their positioning on the map, colour, faction bonus (i.e. discount ninja or cheaper castles, etc) and selection of Daimyos available throughout the campaign. Everything else is the same. This is why Shogun worked well and was balanced, where later titles did not work so well and required balancing through mods.

If S2TW is anything like this, it will probably be the best offering from CA since STW itself. If CA go down the faction specific uber unit route (i.e. "toy units" such as Mori specific monk units, Takeda specific cavalry etc), then they will likely ruin the game balance and thus the game itself. They way to add "flavour" to the game is to give subtle advantages and disadvantages to each faction, such as recruitment cost bonuses - not the faction specific uber units that were introduced in MTW and have persisted ever since.

[Sp00n]

I'm not speaking of purely cosmetic similarities. ETW was similar to MTW, RTW and M2TW in that it consisted of hundreds of faction specific units with widely differing stats based on the typical large Eurasian map, where provinces are practically kingdoms. This huge array of units makes it virtually imposible to balance the game and makes it easy for the player to exploit the inept AI factions.

The original STW is based on a smaller scale map with individual provinces rather than i.e. "Ireland", "Denmark" and "Aragon" as provinces. Unit rosters in STW are identical per faction. Factions differ only in their positioning on the map, colour, faction bonus (i.e. discount ninja or cheaper castles, etc) and selection of Daimyos available throughout the campaign. Everything else is the same. This is why Shogun worked well and was balanced, where later titles did not work so well and required balancing through mods.

If S2TW is anything like this, it will probably be the best offering from CA since STW itself. If CA go down the faction specific uber unit route (i.e. "toy units" such as Mori specific monk units, Takeda specific cavalry etc), then they will likely ruin the game balance and thus the game itself. They way to add "flavour" to the game is to give subtle advantages and disadvantages to each faction, such as recruitment cost bonuses - not the faction specific uber units that were introduced in MTW and have persisted ever since.

Coulnt of said it better tbh, gone is the confusion of mass units and different army types.

[A Nerd]

A samurai is a samurai is a samurai. I agree that similar fighting styles will lead to more interesting and balanced battles. Slight monetary and building time perks would definately be better than faction specific units. Such is the charm of this condensed campaign and war mechanic. Hopefully the AI can build nice balances based on what he sees on his borders vs. a spam of cost effective and/or tactically adept units based on what statistics favor in terms of victory. Hopefully this Sun Tzu thing isn't some algebra formula!

[Phalanx300]

AI based on Sun Tzu sounds great, not sure whether it will work out though.

[Reenk Roink]

The AI in R:TW/M2:TW and beyond was far more complex than anything in Shogun, and I'm sure CA had more time, people, and resources going into the AI of R:TW/M2:TW than of Shogun. It is, as many have pointed out, a matter of the number and variety of the units in the game. AI is simply something that gets extremely hard to do when you have such a wide variety of... things...

If Shogun 2 should stay fairly true to the original, then I am pretty sure that we will have much more of a Shogun experience than a Rome experience in terms of AI.

The biggest question is the nature of the campaign map and movement. Is it the Shogun/Medieval 1 type or is it the Rome and beyond? I actually like the idea campaign map of Rome and beyond better as it is a lot nicer looking and allows for more detail and realism, not to mention opportunities for strategy. However, it is obviously exceptionally difficult to get right as compared to the simpler Shogun style map. I hope for the ideal Rome style map with incredible AI and depth, but if I'm going to have to deal with 1 unit armies moving in and out of my provinces every turn, and most of the major battles being siegefests, then maybe going back to the primitive Shogun style isn't so bad. It's a tough question, and I'm just hoping for the best.

[AussieGiant]

"Well slap me on the arse they are going back to the beginning". :-)

I think they are doing the correct thing. Scaling down on the breadth and seemingly focusing on the depth and richness of the game. Which in my opinion includes the AI in both strategic and tactical forms. They did a very good job with Napoleon and were able to tighten up many aspects of the game.

Astute move by CA in my view. It's been a decade, and they are really going to aim at reinvigorating the fan base with hopefully a very polished and well executed game.

Plus, you really don't wont to go back and $*£" up the first ever game you made, which kicked everything off.

No pressure CA, but I think they are going to ante up or implode, and I don't mind that attitude.

[Phalanx300]

The AI in R:TW/M2:TW and beyond was far more complex than anything in Shogun, and I'm sure CA had more time, people, and resources going into the AI of R:TW/M2:TW than of Shogun. It is, as many have pointed out, a matter of the number and variety of the units in the game. AI is simply something that gets extremely hard to do when you have such a wide variety of... things...

If Shogun 2 should stay fairly true to the original, then I am pretty sure that we will have much more of a Shogun experience than a Rome experience in terms of AI.

The biggest question is the nature of the campaign map and movement. Is it the Shogun/Medieval 1 type or is it the Rome and beyond? I actually like the idea campaign map of Rome and beyond better as it is a lot nicer looking and allows for more detail and realism, not to mention opportunities for strategy. However, it is obviously exceptionally difficult to get right as compared to the simpler Shogun style map. I hope for the ideal Rome style map with incredible AI and depth, but if I'm going to have to deal with 1 unit armies moving in and out of my provinces every turn, and most of the major battles being siegefests, then maybe going back to the primitive Shogun style isn't so bad. It's a tough question, and I'm just hoping for the best.

Yes it was more complex, however if not given enough attention it wont provide as much of an challenge. I played Shogun I demo once and AI totally destroyed me, where I was used to weak AI of Rome.

[Tera]

The AI in Total War games is not bad at all compared to most other strategy games. In many other popular RTS (e.g. Warcraft, Starcraft, C&C) the challenge from the AI mostly derives from large numbers ('quantity'). Large number of defensive towers, enemy units, etc. Also, on 'Hard' difficulty modes, the AI is allowed to "cheat" like building units faster, costing less resources etc. The whole Sun Tzu AI concept is mostly a very nice sounding marketing term, but it is also true that Total War AI also employs 'quality' by using terrain effectively and making smart moves on the campaign map. That's a great thing and let's hope CA can keep it up.

Diplomacy is still way behind games like Civilization, though. That's one area where the AI can improve, as ten years after the first game, most alliances in Total War are still pointless.

P.S.
Finally found my old account, meh.

[Crazed Rabbit]

So, what are the odds that there will be a decent AI in this one?

They've made plenty of money with terrible AI, so why change?

CR

[TheDuck]

I don't think it's that hard for the developers to make a decent AI, even for a game with varying units like RTW.

You have your melee infantry, melee cavalry, ranged infantry, ranged cavalry, or a combo...

The developers are just lazy...

I'm a programmer.. nothing about AI for games like Total War is 'easy'.. nothing at all.

[Intranetusa]

If CA can't make a decent AI, they should make the AI non-hard coded...

Then some modders/programmers like the EB team will just fix up the AI and port Asia-ton-Barbarorum over ez peazy... :D

I'm a programmer.. nothing about AI for games like Total War is 'easy'.. nothing at all.

=(