Firaxis failed at difficulty levels - no ranting!

[lordsurya08]

No, this is NOT another rant at how easy the game is. On the contrary, I find the higher difficulty levels to be decently challenging - but challenging in in an "artificial" way.

Think - if you were making a game of your own, what would you do to implement different diffuculty levels? You'd improve the AI. You wouldn't make the game "artificially" difficult like Firaxis does. What exactly do I mean by "artificial"? Read on.

The screen where you pick the difficulty says that in higher difficulties, AI gets small or large "advantages", a phrase that leads one to think that the AI improves at higher levels. No siree - it turns out that the AI basically gets big advantages at , $$, culture, and tech over you. I don't mean small advantages, I mean BIG advantages.

I was France and had four cities, that's +8 culture for me thanks to the fantasic UU. The guy next to me, Darius of Persia, had about two cities, and he STILL ranked #3 on the culture leaderboard. Where was I? #9, dammit! So he must have been building temples and monuments on BOTH of his two cities and spend piles of gold on several city states to get ahead of me in culture. Very well. At the same time, he must have constructed or bought a number of swordmen (think - iron working), created workers to build a long network of roads to connect his two distant cities, constructed Stonehenge (which takes some technology, don't remember which), and bought or produced ten plus immortals, swordmen, and archers. So he's got a lot of buildings, roads, and military units, AND he's ahead of me in tech, as all I had was iron working, archery, husbandry, and was working on calendar.

Then I opened up the trade screen. He had some 900 gold with 40+ gpt - I had some 400 gold with 25 gpt, which went down to 15 gpt after expenses and maintainance. And I had about half the army as his and only 3 tiles of road. And then the icing on the cake - my happiness was at 0, his was at 20. I scouted his land, and found just one luxury resource - a sugar plantation.

Discuss

 

[Yzman]

I don't know why you would think that the AI improves at higher levels, when I hear " the AI gets advantages" I think that the AI gets advantages. Which it does. Its totally transparent in telling you that.

Its hard enough for them to program a competent AI at any level and I don't think the best players should get the most competent AI. It would be impossible it seems for them to make deity be this great AI where settler is a carrot playing the game.

 

[Tylerryan79]

It has been this way in all past games. There is one dumb ai, and the higher it goes the more advantages it gets.

 

[PieceOfMind]

lordsurya08, what difficulty were you playing at?

Am I to be told that Firaxis's idea of a higher difficulty level is giving the AI three times the cash output per tile, the ability to produce or buy three scouts, several workers, numerous monuments, temples, colosseums, a massive army of immortals,

If at Immortal or Deity, where the AI starts with a free worker, or a free worker (or two?) and a second settler, these sorts of things are pretty typical. That sort of bonus at the start of the game compounds very quickly.

 

[lordsurya08]

Immortal, I believe - the one before diety.

 

[elthrasher]

In my experience, game developers and designers can never play as well as the hardcore players. If they can't play as well as the top-level players, how are they going to program the AI to play that well? They just aren't. I'm not saying the AI needs to be stupid horrible like it is, what with moving archers next to swordsmen instead of shooting them from a distance and all that, but the computer just isn't going to play super well. To make the game challenging for better players, the AI has to have some advantages or cheats available to it.

 

[PieceOfMind]

Immortal, I believe - the one before diety.

Ok, so they probably get the free worker and a bunch of extra units like a couple of scouts and a couple of warriors.

I'm trying a game at Emperor at the moment to see if it's more enjoyable to play. If one doesn't ruthlessly pursue victory via the most optimal path possible, bumping the difficulty down from Imm to Emp might make the game a bit more interesting IMO.

But since the new patch, barbs seem to have been put on steroids. The bonus against them has been reduced, and they seem to spawn faster from their camps.

 

[Biz_]

welcome to strategy games

it's not just firaxis that does this. it's every developer

 

[carlisimo]

"We've done it! It's taken years and millions of dollars but we've created the greatest AI ever written! Now let's turn it off for 95% of our customers."

 

[Yoshimune]

All games are like that. If you've ever played a RTS game, the difficulty levels are just more cash, more units, more damage, more armor for the AI.

 

[Rexflex]

It was suggested in one of the pre-release interviews that the lower difficulty levels would have a number of options available to the AI, and the higher difficulty levels would further refine its selection of options. But this is far from saying that the higher difficulty level AI would be smarter as such - just less likely to make more generic strategies rather than ones tailored for that AI to win.

 

[Windsor]

Think - if you were making a game of your own, what would you do to implement different diffuculty levels? You'd improve the AI. You wouldn't make the game "artificially" difficult like Firaxis does. .

I would focus on what gives the most fun player experience. And that will often be handicap bonuses.

Improving(=changing AI to make the game more difficult) will change the AI in a way many players will not like. Even a minor thing like the resource deals in the latest patch seems to annoy players. Now, if you want to replace say emperor-handicaps with AI-behavior then you would need to go much further.

Give players an AI that treats human-players very differently from AI-players and you would really see a lot of rage on this forums.

But whatever you do, a "better" AI would not be able to compete with the handicaps at the highest levels. What you would do is really to ruin the game for king/emperor-players.

 

[ronnipalooza]

I've recently made the jump from prince to king. I couldn't believe how easy the switch was. All I had to to was be more selective with money and production. I had heard from these boards that on higher difficulties, the AI get ridiculous bonuses to make up. My strategy was simply to really focus towards goals such as wonders which gave culture multipliers. And it worked.
I kind of like the idea of the AI getting bonuses so the entire game is an uphill battle. Yet, I've realized that it's only a matter of time until I even the playing field and win. The AI should at least have some other tricks (unit placement, better understanding of tile, ect.) on higher difficulties.

 

[Bluetooth]

I don't know why you would think that the AI improves at higher levels....

Its hard enough for them to program a competent AI at any level and I don't think the best players should get the most competent AI. It would be impossible it seems for them to make deity be this great AI where settler is a carrot playing the game.

Well, maybe because that’s what they said prior to release. Here’s an excerpt from one of the pod casts 2K published.

ET: Beyond the complexities of each civilization’s individual personality and gameplay style, there is another level to AI in Civilization V: Difficulty levels. When you get to the top tier of difficult, the AI can be brutal – I know that I’m not the only one who has accused the computer of cheating when on an expert level. Ed explains what the AI’s “thought process” is like on different difficulty levels, and how it gets “smarter” as you advance to more challenging settings.

Here’s a link to the transcript: http://www.civilization5.com/#/community/podcast_transcript_8

 

[BlastDuke]

All games are like that. If you've ever played a RTS game, the difficulty levels are just more cash, more units, more damage, more armor for the AI.

In nearly most cases you are right about this, but there are some games that really show different AI behaviour (=new tactics, different unit behaviour) on different AI levels:
Newest Example: Starcraft 2

 

[Slowpoke]

Well really RTSes are much simpler, and the AI has incomparably faster reflexes than the player, that you can't compare them. Admittedly, though, this is really really bad A.I. I've never seen a strategy game yet which fails so hard at protecting fragile ranged units. I can't conceive of why they don't have a "Hey there's melee units around, DON"T MOVE FORWARD."

 

[Aoeu]

Well, chess programs work like this: the AI is normally unbeatable (unless you are among the best players in the world), and you can win a game only by intentionally limiting the depth and quality of the analysis it performs.

I'm not sure why strategy games should be different (perhaps writing the AI is harder, but that's where most of the development work is supposed to go).
The current system in Civilization V is definitely not really enjoyable due to this.

Imagine if you had to give a chess program some additional pieces just so that you can't win trivially: it certainly wouldn't be enjoyable.

The whole point of releasing a strategy game is to provide a great game design and an awesome AI.

And if you really want to cheat, then at least give the AI information it shouldn't have (e.g. where all the player units are) rather than bonuses that alter gameplay.

 

[JujuLautre]

I'm not sure why strategy games should be different (perhaps writing the AI is harder, but that's where most of the development work is supposed to go).

Search space size and evaluation function design, not to mention randomness and unknown information. Considering these, Chess is trivial compared to civ.
Also, FYI, handicap does exist for Chess, as well as for many games.

 

[Stonemender]

Well, chess programs work like this: the AI is normally unbeatable (unless you are among the best players in the world), and you can win a game only by intentionally limiting the depth and quality of the analysis it performs.

I'm not sure why strategy games should be different (perhaps writing the AI is harder, but that's where most of the development work is supposed to go).
The current system in Civilization V is definitely not really enjoyable due to this.

Seriously. Have you thought that through?

Chess: A battlefield with only 64 Squares and 32 units. Most units have less than 3 different move options. The maximum number of move options a unit can possibly have is 28. A unit can perform only 1 action. The AI has to move 1 unit every turn. The result of a 'combat' is always 100% victory. Depending on the playing level, calculations can take several seconds.

Civilization: A battlefield with hundreds of squares and hundreds of units. Most units have at least 18 move options. With roads and high movement, you will get several hundred target hexes. Most units can perform several actions. The AI has to move hundreds of units per turn. There is a luck factor included in the combat.

In short: It is close to impossible to use chess algorithms. Even if it was possible, you would have to wait several minutes between turns. I would think several hours in late game situations. At this time it is virtually impossible to design an AI that can compete with a player on a tactical level.

 

[dannythefool]

If you think about it, there isn't really a whole lot of a difference between an AI that builds X% faster and an AI that is so smart that it can optimize its empire so it builds X% faster. The way it is implemented is the first, but if it makes you feel better, you can pretend it is the second... It's always been like this in Civ games, though.