Strategy computer game abtitude

[JPC]

hey just wondering

there are some computer strategy games like Empire Earth where beating hard computers seems very impossible , but how can some manage

i mean, in this game, the computer can manage troops on the whole map
while you can only control one zone at a time
and the pc attacks you on 5 different areas
this means u can only apply war strategies on one area, while the computer can take the advantage on the others

so how can some players manage this ?
do they just go very very fast ?

 

[MeJennifer]

The computer simply gives the player a lot of slack.

The computer can beat the human opponent in almost all games. In chess the computer wins from everybody except for the very few best players in the world. The only game where the computer is still not better than a human is Go, an ancient Chinese game.

 

[Thrice]

Speed & experience, mostly. You eventually learn what works vs the computer & how to do it.

 

[neurocomp2003]

it also depends on the the input setup u use. Though most rts games have a fixed keyboard/mouse layout(which i don't like) some times they allow for both the arrow motion and mouse pointing motion. Arrow motion is faster but you lack the shortkeys of the rest of the builds when you use them. Moust pointing allows for the building fast but not the map navgation.

Also u can play N people versus M computers which might help. But more than likely what thrice said will be the most important.

 

[siddharth]

The only game where the computer is still not better than a human is Go, an ancient Chinese game.

Also, the "solved" games, like Tic tac toe

 

[russ_watters]

Artificial intelligence isn't yet all that good - many, if not most computer games stack the deck against the player. I play Command and Conquer and the computer generally starts out with more resources and units

For first-person shooters, it is the opposite. Though they sometimes give the computer more hit-points (in offense and defense), the computer has to be handicapped in almost every other way. It would be easy to have the computer have perfect shooting accuracy, for example.

 

[JPC]

yeah that too

the computer in Empire Earth starts off making lots and lots of buildings which he normally could not afford

but true there is one weakness in the computer : he usually divides his army into lots of small chunks which he sends a bit everywhere (thats why you get attacked on all sides)

if you have your based highly secured, you can make a big army and crush his army bits by bits
and u need to go fast because each time u destroy one of his buildings , he rebuilds it in the other side of the map

 

[caslav.ilic]

The computer simply gives the player a lot of slack.

The computer can beat the human opponent in almost all games. In chess the computer wins [...]

I highly doubt this, providing the usual situation in which players assume "command" positions — where they themselves might be perfect, but the resources that they control and the environment in which those are used have many sources of failing and randomness. Chess in this respect is exceedingly simple, as the number of possible moves and outcomes is mathematically precisely defined. Yet even more simple would be "big-number arithmetic contest", where no human at all could oppose a hand-held calculator...

In effect, strong AI is something that is heavily marketed as a selling point of any game, as so far it is not nearly able to outmaneuver human opponents. If anyone thinks the particular game in single-player is being hard on the player, he should try the same setup in multiplayer, against human opponents, and behold the real carnage :)

For first-person shooters, it is the opposite. Though they sometimes give the computer more hit-points (in offense and defense), the computer has to be handicapped in almost every other way. It would be easy to have the computer have perfect shooting accuracy, for example.

Not necessarily. Most FPS games do not use sth. like real-effect lasers, where the perfect aim would be equivalent to a hit. Instead, there is usually either some shot ETA delay (aim lead needed) or slight randomness in shot spread. In such environments, even if the computer is allowed to have perfect aim, all bets are open.

Interesting practical demonstration of this were some Counter-Strike multiplayer games which I had chance to observe, where some of the human players would be using an "aim-bot", a perfect computer controlled aim instead of their own human input (i.e. cheating). Although it would improve odds for a newbie, it was obviously inferior to an experienced player, as witnessed by the end-game stats. I can think of several reasons why this was so, all having to do with the rest of the gameplay mechanics in this particular game.

--
Chusslove Illich (Часлав Илић)

 

[Thrice]

Artificial intelligence isn't yet all that good - many, if not most computer games stack the deck against the player.

I think it would be if they worked at it with an IBM budget. Compared to chess, the games aren't particularly complex.

 

[Hurkyl]

hey just wondering

there are some computer strategy games like Empire Earth where beating hard computers seems very impossible , but how can some manage

i mean, in this game, the computer can manage troops on the whole map
while you can only control one zone at a time
and the pc attacks you on 5 different areas
this means u can only apply war strategies on one area, while the computer can take the advantage on the others

so how can some players manage this ?
do they just go very very fast ?

I haven't played Empire earth, but I can speak from my experience with the *-craft series.

The computer's biggest and only relevant inherent advantage over the casual player is that it has a clue. It was programmed with some basic knowledge about how to invest its financial resources effectively: it knows to invest a lot into its economy, and to focus its research efforts towards the type of army it's trying to produce.

The casual player, on the other hand, may not have learned how to manage their finances effectively: they invest slowly in their economy, research in an unfocused manner, and don't build the infrastructure needed to produce enough units. Thus, they rapidly fall behind the computer and are eventually overwhelmed.

Once basic financial management is learned, and the player forms a general understanding of the units and technology 'e may command, the computer stands absolutely no chance in a fair fight -- the game producers simply have not spent enough effort in develop a strong strategic AI. (and to be fair, that can be an incredibly difficult task)

The computer does have the advantage in reflexes and attention span, but those are generally quite inconsequential as compared to what I've said above.


The computer has another severe disadvantage: it generally isn't programmed to learn the game. (again, that's an incredibly difficult programming task) The computer only knows what it was programmed to know, and thus it cannot take advantage of the new strategies discovered by the player community, unless the game developers spend a lot of effort to keep their AI up to date.

 

[Hurkyl]

I think it would be if they worked at it with an IBM budget. Compared to chess, the games aren't particularly complex.

Keep in mind that the power of chess programs comes from incredibly sophisticated brute force searching techniques... and does not stem from any sort of strategic understanding. In other words, chess computers perform well because of their incredible tactical play, and a chess game is small enough for that to prevail.

But in a go, or a real time strategy game, incredible tactical play generally cannot substitute for good strategy.

(Of course, when computers get bigger, then maybe a small go board will become small enough for tactical play to prevail. But then people will just move onto larger games and beat computers there!)

Incidentally, I would disagree with the notion that chess is a more complicated game. It's just that we've been playing it for so much longer, and thus have devoted more time to its analysis.

 

[Thrice]

But in a go, or a real time strategy game, incredible tactical play generally cannot substitute for good strategy.

I don't know that I'd put go in the same category as RTS's I've played. RTS at higher lvls becomes very monotonous. You have little choice on general strategy & micro becomes very important. The warcraft 3 computer on "normal" already plays as a solid mid level player - he can probably beat more than 50% the people on bnet . If you programmed him with one weak ass abuse strat (DK lich fiend instead of the current Noah's ark army) i'd put it at 80%. It stands to reason that we'd find computers dominating tournaments if IBM was funding the AI.

 

[moose]

Yeah, Go is unbelievably complicated. I couldn't even imagine trying to write a program to play it.

The thing about war games like Age of Empires is that a computer can calculate the absolute fastest way to grow an army that would be most effective against you. It's not that entirely complicated.

 

[light_bulb]

i have empire earth, haven't gotten the chance to really explore it, the problem with computer games is that the computer only knows how to follow routines. a computer doesn't have the chance to run "outside the box" and no chance against a patch made by a rouge hacker.

 

[Hurkyl]

rouge hacker.

What about a verdant hacker, or an azure one? Why are the rouge ones special?

 

[Hurkyl]

Yeah, Go is unbelievably complicated. I couldn't even imagine trying to write a program to play it.

The thing about war games like Age of Empires is that a computer can calculate the absolute fastest way to grow an army that would be most effective against you. It's not that entirely complicated.

I was under the impression that you have to actually control your units, rather than simply feeding them into a combat calculator to see who wins.

 

[light_bulb]

your army grew by 1000 so mine must be double that *quick patch*, the hacker wins crash the computer with a one off error or bring it to a crawl if you actually have a chance of winning or even better change the rules, your opponents now have to cross endless land and waste all resources until they die.

 

[russ_watters]

You do - and because of that, the number of variables involved makes it impossible for a computer to do what computerized chess programs do: essentially calculate every possible/likely move.

For example, the faastest way to grow an army depends on how long you have to grow it. I don't know Age of Empires, but in Command and Conquer, you start with resources, but have to collect more. If you have time and will need a big army later, you can start by building units to collect resources. If you are going to be attacked right away, you need combat units right away.

Such strategic concepts are very straightforward, but extremely difficult for a computer or human to deal with due to the fact that you can't see the attacks coming: it's like chess without being able to see your opponent's moves. Because of that, you basically have to pick one and hope it is the right strategy.


In C&C, the computer's attacks are preprogrammed, though, so while you don't really know the first time what you need to do, you can always restart the game and know what strategic approach to take to each mission. Even worse, they are repetitive, so if you set up the appropriate units in the appropriate place and they can heal themselves, they can repel the same attack dozens of times. But the first time through each mission, the computer will generally wipe the floor with you because it starts strong.

 

[light_bulb]

who holds the computer chess title, deep blue?

 

[moose]

I was under the impression that you have to actually control your units, rather than simply feeding them into a combat calculator to see who wins.

The computer has some basic formations programmed in which aren't all that bad... However, you are right. You can usually exploit some weakness of the computer...

 

[JPC]

yeah but as i said the thing is , you got to be fast so that you can use tactics with ur main army rather than just sending them to the front and letting them take the initiatives

because if ur not fast, you cannot manage all the other zones

like one time i was playing against 2 computers that were themselves agaist each other
i sent my main army into the first computer , using tactics , ect
was successful
but when i looked back at my base, a quarter was gone lol

 

[Huckleberry]

FEAR is a FPS with excellent AI. I played this a few years ago and remember the AI taking advantage of chokepoints. They take cover, or make it for themselves. They provide cover-fire for each other while they change position to their advantage. If they can't get into your position safely they eventually will not come in at all. (That makes for a very long waiting game.) They prepare attacks from several directions simultaneously.

I never seem to have enough grenades when fighting these guys. I shoot at one and he hides while his buddy shoots at me. I attack the other guy and he hides and the one I was shooting at before pops up and shoots again. Then another guy I didn't see comes from behind. I turn to kill him and the other two I was holding down start moving on my back. Sometimes they have armored units that can take several powerful hits. They come right on top of me and start blasting. Next thing I know all the other ones are right on top of me while I'm shooting their armored unit. It's a very nice AI.

I spent a lot of time retreating to a more favorable position, leaving mines as I went. Any AI that I could dispatch with stealth was ideal. The sniper rifle was my best friend when stealth failed, and picking off a few from range was very helpful for weakiening their defensive positions. I threw grenades to get them out of cover, and the run'n'gun charge as they change positions. It's been a while since I played it. I didn't have the chance to finish the game.