That is interesting.
Printable View
That is interesting.
That is good news. Of course the thing about AI is one doesn't know how good it is until the game actually arrives.
Well, it is a positive thing that they admitted the horrible battle AI of RTW.
CA has repeatedly stated they are working on the A.I.; and R:TW's A.I. has undergone a great (and much needed, I think we agree) improvement since 1.0, so I rather think the opposite is obvious.Quote:
Originally Posted by Dunhill
Given that I have trouble enough controlling R:TW's twenty units I do not see how this is a bad thing.Quote:
Originally Posted by Dunhill
If you talk about the battlefield in general, yes, I agree there. But only in general: there are many small improvements in M:TW and R:TW, from unit line-up, effects of unit depth, cavalry charge. Also, R:TW's A.I., bad though it may be, is more flexible in its formations. Too bad it's basic formations (not to mention tactics) are far worse.Quote:
Originally Posted by Dunhill
I think Simon's point is that the more choices an A.I. has too make, the slower and or dumber it becomes. There only way around this is by restricting the number of alternatives available to the A.I. I think we all agree that R:TW's A.I. was poor (pathetic may be a better word for the 1.0 edition, but it has gone a long way since then), but the A.I. of a fast-paced game is always going to perform more poorly than that of a slower one (like I assume the Take Command series to be). Now if we could convince CA to slow their battles down...Quote:
Originally Posted by Dunhill
I agree completely. However, I doubt that CA is going to return to their niche position after entering the big market with R:TW. Sadly, slow, challenging games aren't considered worthwile by the mass of players that just want to be entertained.Quote:
Originally Posted by Dunhill
I agree with Dunhill and was disappointed to learn of no increase in army size.
I think the problem of adding more units to the battlefield is collision detection, I remember reading somewhere a CA dev saying that this is the biggest resource hog. I hope that this can be overcome with future TW releases, as for me realistic troop numbers really helps with immersion. Don't tell me that one man represents 10, I see one man so it's only one man.
If I could believe that I would most probably still be playing AOE.
e.g. In RTW I just could'nt believe that I was commanding a Roman army which I belive was around four legions/20,000 men strong. To overcome this I tried to pretend I was a lowly legion commander in command of an under strength legion on border patrol.
As for more than twenty units becoming too much of a handfull perhaps just increase the unit size or better yet, maybe CA will reinvent the way we command our armies, perhaps with less individual clicks and allow us to issue general commands to groups of units.
-IceTorque
If a TW game would introduce 20 000 v 20 000 battles all fighting at once (no reinforcements), than we would see the return of clone armies all doing the same action. As 40 000 warriors all looking different, all doing different things would demand to much of a computer (at least a home computer).
If you really want to be a general this would not be a problem as you probably wouldn't zoom in anyhow in order to keep an overview. But this isn't the way CA has taken the TW series.
Really? How is it obvious. Its amazing how you can deduce such information from statements such as:Quote:
Originally Posted by Dunhill
CA is working on the AI, MIITW will have an improved AI etc.
:rolleyes:
If there has been improvement in the AI since STW I haven't noticed it. In other words, I don't think there has been any significant improvement since the beginning.
The thing I like about Take Command is that you can select how many units you wish to control. However, game play can be frustrating when nearby brigade commanders are engaged elsewhere or not aware of your situation, you are on your own to a degree. I see this as half the fun though. You can take commnad of the whole army and micromanage your head off if you want too.
You are correct TC2M is a slower-paced battle. However, melee, when it happens, is resolved rather quickly. It's not glacial, but very realistic. I don't notice it much though, I end up wondering how a game set for an hour finishes so quickly. I like the fact that the game can swing back and forth quite a bit. It's not all or nothing, which RTW often is.
In Take Command you get to pick the level of AI cycles you wish to use. This is one of the best features of the game. You can also select difficulty, but this is just giving more men to the AI. I don't know any other game which allows you to increase AI cycles. So far, the AI in TC2M is my equal.
RTW isn't so much faster, and the number of units are far fewer, so I don't see why the AI has to be so easy to beat. TC2M is working with units of 500-600 men, tracking individual artillery guns, computing fire, ammo, fatigue LOS and morale between numerous divisions, and dealing with complex terrain. It's doing all this without stutter and giving me a damn good game.
Remember, this game is made by two guys in thier spare time, plus maybe another 10-12 helping out. This isn't what you would call a dedicated development team. Maybe that's why it is so good. They don;t have to listen to what the sales team thinks the game should be to sell well. They actually listen to the players and have worked with them to develop a game we all like.
What some call clones, I call uniform. I always play with the camera at max zoom to take in the big picture, to try and get that feeling of witnessing a battle straight out of history. If I want close up FPS action I'll go play TES flaw. Obviously what I like is not what the mass market likes, so too bad for me. Smoke 'n' mirrors, bells 'n' whistles win again.Quote:
Originally Posted by Peasant Phill
-IceTorque
I really think that one of the most exciting elements about Total War; is the micro-management needed in the battle. I don't want to have a 20k army, where their will barely be any place for any tactical moves, or any tactic to do even. It will be a fight and die battle, with 2 lines clashing together till one routs..
P.S. Links on the starting posts are updated.
I agree that the tactical AI has not been improved since the first game STW. For instance and this is only one example, in RTW the tactical AI was degraded by having the AI make frontal charges with units that are weaker than the unit they are attacking. This never happened in STW until you forced the AI to charge by charging at the weaker AI unit yourself. The AI unit would then charge because it would get a charge bonus which is better than just standing there. There are other specific tactical AI behaviors you can point to that were superior in STW. This sets up the present situation where Creative Assembly can improve the tactical AI in M2TW from RTW, and it could still end up worse than the tactical AI in STW.Quote:
Originally Posted by Dunhill
In addition to tactical AI, we can also see that the new battle engine is inferior to the original battle engine. For example, the "squeezed too tight" combat penalty was not included in RTW. Another example is the normal unit size was reduced which contributes to the accelerated combat resolution. These things and others has an adverse impact on the gameplay in multiplayer. So, here too Creative Assembly can claim they are making M2TW better than RTW, but it doesn't mean the game will be better in multiplayer than the original STW game.
What do you mean, AI cycles?Quote:
Originally Posted by Dunhill
Well, I don't think that clicking madly while twenty units are running around like chooks with their heads cut off qualifies as tactics. Although I do concede that if realistic numbers were possible that this might cause the problem of having a realistic looking battle, which would create another problem of having to use realistic tactics. Now would'nt that be a shame.Quote:
Originally Posted by x-dANGEr
As for the AI debate I feel that it is far improved over STW, but it is let down by unit formations that are set too wide and no depth, tiny battlefields with not enough room to manoeuvre and problems with the engine itself. I have enough trouble getting the drunken little sods to obey my commands. I imagine the AI has the same if not more trouble with em.
-IceTorque
.. Yet another one comes and say this "clicking madly"..
There is certainly plenty of room for tactics in vast battles. It is just that it would probably be impossible to orchestrate such a large army with the same precision as one can control an army of a few thousand soldiers.
Since I do not play games of the take command series I cannot compare, but I am certainly willing to believe that the TW battle A.I. can be outdone. I also agree that there has been little to no improvement in the series: R:TW did a small step forward but several big ones backward. Yet I still do not agree that it is obvious no attention will be given to the A.I. The improvement in R:TW's A.I. by various patches suggest otherwise, as do statements made by CA.Quote:
Originally Posted by Dunhill
As far as the AI my personal comparison would be Galactic Civilizations (spelling?).
A relatively small company title that has the best AI I have ever faced in a single player game.
Of course this may be an entirely different "level" of AI. The AI in Galactic Civilizations is faced almost exclusively on a "strategic" level where Total War's AI also deals with "tactical" issues. Personally I don't know if the strategic and tactical AI's in the TW series are combined or seperate though they do appear to be seperate code.
Besides GA I am universally disappointed in the performance of AI in games. I don't know if GA stumbled upon the "magic bullet" that makes their AI so impressive or if AI is simply not as important a factor for other companies.
I have noticed improvements in TW's strategic AI but I agree that its tactical AI, while better, is simply no better than most games out there.
Whatever the case I never have time to "zoom in" being a less than brilliant general myself. If I don't keep the "bird's eye" view going I have actually lost battles not from being outfought by the AI but simply being overwhelmed by the number of units I'm trying to manage.
I can understand that, as with most games, compromises must be made. However, until the finished product is released, and several reviews are in, we just won't know if those compromises really do hurt the TW experience.
:focus: :focus: :focus: :focus: :focus:
Elephant mounted cannons have been confirmed !!!!
First of all,
Nice to see you back Puzz3D.. It's also nice to see someone with a join date older than mine agree with my observations.
Myrddraal,
Maybe it would have been better for me to say that, in my opinion, there will be no significant change to the AI and most of the time will be spent on the graphic features. I'm not so sure I can agree with Puzz3D in stating the AI actually took a few steps backward, but it surely hasn't gotten any better over teh past few years, so I've seen no evidence of significant improvments in the past and I expect none now. However, all that being said, I'd really appreciate it if they did do some work on the AI, because that is what is stopping me from playing the game more. It's not providing any challenge.
Screwtype,
AI cycles are loops/programs for the AI to perform as it plays. In TC2M I can select from several settings. This allows me to force the CPU to use more power working the AI sub-routine. This is why this single player game is so fun to play. It is not a push over for me. I have to think hard to match the AI, which was written by a Grognard, so it may already start out at a higher level than other games.
It may also be that TC2M is using a much larger map. Things are happening at real time. So, I have to pay attention to how I move units. Roads and how you move units through some terrain actually have a significant impact on fatigue. I wouldnt try to run units though woods unless there was a great need. In TW I'd not pay that much attention except to height as a bonus.