3355

Sounds fun to me. But then I’m about the only person here who really likes Civ VI!

3356

I think maybe you don’t understand the diplomacy in EU4. ;) @CraigM has done some good write-ups on what makes it so detailed!

3357

That was a rather unnecessary passive aggressive remark. I have been playing EU 4 for years.

Paradox games are very “detailed”, yes, and they tend to get the “details” wrong quite often too. Hilariously often actually, but you know, jokes have a set time. Otherwise the joke is on you.

They also like numbers a lot. What do the numbers actually mean, not much if one judges by their always extensive patch notes where quite often a bug has nullified those modifiers for months or years and no one was the wiser about it. Writing about Paradox games in depth is like writing about patterns in the white noise you see on a TV screen. Darren Aronofsky made a movie about that. A good movie, so I am interesting on what Craig wrote too.

I actually agree with the statement you made that going down the “Paradox” way was the wrong thing to do, but not for the same reasons I suspect.

3358

I’m glad someone does. And I remember back when you got this and the expansion pack and it was great to read something positive about the game and that you were enjoying it. I feel like a sucker with Civ, I keep coming back hoping it’ll get good. In the end, even Civ V turned out pretty ok to the point where I could enjoy it through its flaws. And deep down, I hope the second expansion to Civ VI does the same.

But oh boy, the diplomacy. I think I’m just tired of being spirited away to another screen to deal with diplomacy. Like in EU, I’d like to be able to declare to the world who I like and dislike, as opposed to the AI playing a guessing game with me. Firaxis have it wrong, the AI should be reacting to me, not the other way around. Uness it somehow becomes self aware, I’m the only one playing that particular game, the AI is just running the numbers to act as a hurdle. Further to that point, let the turns run through and then let me sift through the diplomatic messages. Yes, I’m looking at you Total War games.

When I look back at Civ IV, I saw the diplomacy as being pretty good. The fact that Firaxis had something good and tossed it away for Civ V, then slowly reintroduced the positive and negative factors that made a relationship in VI does give me some hope.

3359

Whoa, sorry, it wasn’t intended that way! I didn’t know/recall what your experience level with EU4 was. In that case, I’ll just say that I completely disagree with your assessment of the diplomacy system and how you interact with the AI through it (and vice versa).

3360

Hah, no worries!

3361

Thanks for that message, Kevin. I have ~130 hours on EU4 and about 5% of the achievements, doh.

To each their own. I personally don’t find much of a difference in the behaviour I get from, say France or Austria towards Spain in EU 4, and the seemingly random angry reactions I get from CIV6 “leaders”.

Giving me some numbers that tell me the “absolute” value of a particular perception of the AI towards me isn’t particularly helpful as it is difficult to relate them to in game actions that I implement. Typically to try sway the historical determinism that Paradox has put into the system. Or what I perceive as such.

For instance, you can be actually investing lots of in game resources (money), or even making concessions (wanna Navarre, here you have it we don’t like those Protestant Navarrese people anyways!), yet you don’t really get an idea of what was the impact of that, Why does Francis I hate my guts if, on top of that, actually I have been supporting his Italian policies by turning down offers to ally from Milan and going for an Atlanticist policy.

He won’t give up until I have bashed his nation into submission, break it down into several rump kingdoms and what not. And even then, like a vampire, will try to come back for more!

3362

At the age of incredible AI with Alphago I do not understand why civ don’t get a decent one.
Please stop writing new games mechanism until you can put an AI which can handle them.

Btw civ 5 fully modded is a wonderful game, did spend hundreds of hours on that.

3363

The “pocket” version of Alpha Go runs on 48 CPUs and 8 GPUs, for a game much simpler than Civ, or even Settlers of Cataan.

I think that having a ban on developing new strategy games until 2030 or so is in order…

3364

Go is much more complex as any of these games you should reconsider your opinion on this.
It’s not the rules making a game more complex only, but the huge quantity of possibilities generated, and u guess go is much more deep as these games.
And you don’t need this computer power to apply the result of the learning stage: you can run a light Alphago on your home computer already (with awesome results)

I think they just don’t invest in the right way. I can understand that from dumbed down games like sims4 but I did expect better for civ6

3365

By what definition of complexity? The number of states in Civ is dramatically higher than Go. For starters, you make one decision in Go per turn: where to place your piece. You make many decisions in Civ every turn.

On top of that, AlphaGo isn’t just a block-box you can stuff any old set of game rules into. There were tons of (human derived) heuristics etc. layered inside that program, huge amounts of training data and hand-holding to get the reinforcement learning working in a timely fashion, etc. It’s an amazing construction, but it was no small endevour.

Suggesting any video game company could make a good return piling similar amounts of resources into a game like Civilisation is very optimistic about both how much most players value the AI, and how simple such a task would be.

3366

Well basically you can’t process in a classic heuristic way go as there are more possibilities as atom in the universe. I think there are quite some documentation on this aspect now available.
I don’t say it would be an easy task to apply the way of Alphago to civ but you know they are applying it to medical knowledge, food and agriculture… So I don’t think civ is something out of reach.

Now for the value for players, it’s interesting subject. Many could prefer a dumb AI so that they can build their cities in peace.we got used to that:)
If AI start to be too strong, maybe we could find less interest too. But myself I would prefer that as the handicap system of now

3367

Not so much in its successor, AlphaGoZero. I’d love to see the results of Google giving the process a try. The trouble is that I’d expect it, if successful, to result in an AI that ruthless exploits every loophole in the rules and plays solely to win. And that isn’t what strong players really want from an AI for Civ, in my experience: they want a competent AI, certainly, but they’re also looking for a relatable opponent, with comprehensible motivations.

3368

Yeah. That was the second astonishing fact for me: next they did something not using the human knowledge, and which did get even stronger as the one using it!

I would never had expected those to happen.

3369

Yes, the game design issues are an entirely separate issue to address.

And then there are deployment issues to consider: you have to retrain your AI (no small task for this enterprise!) every time you patch the rules of the game.

They’d do a lot better at building a better AI by making a game more playable by the AI, not an AI that plays a better game.

3370

I picked up the last expansion pack and was having a great time with it for the first few hours, and then the completely random behavior of the AI just ruined the game for me. It’s been mentioned before, but the aggressive levels of stupidity on display here by the AI just make this unplayable, despite the fact I like several of the systems/mechanics in place.

I’ll probably skip this expansion, unless I read they actually DO fix the diplomacy system and make it feel like a real 4X game, and not a wargame the AI doesn’t understand.

3371

@Spock and @Strato

I need a fresh start on this as I do want to enjoy it. What parts of the game do you really dig so I can focus more on the good parts?

3372

Real AI techniques – for example the tweaked deep learning approach used by alphago – have virtually no bearing on the so-called “AI” in a game like Civ VI.

In Civ VI, the challenge is encoding ordinary high-level logic, often in a simple rule-based form, in combination with simple algorithms for pathing and numerical calculation. This is not an intellectually difficult problem; it’s just a question of investing time and resources in getting it to work. The Civ series has historically laughed at this investment and done the minimal work necessary to give the computer nations the appearance of some kind of sensible play.

Unfortunately, it seems the designers in Civ V and especially Civ VI have failed to understand that adding decision-making complexity and burden to the human player also adds it to the computer player. Or if they do understand, they don’t care. So they keep coming out with game systems that they aren’t prepared to code properly for the computer player. They weren’t prepared even to implement competent pathing for one-unit-per-hex back in Civ V, and now with every expansion adding yet more stuff they don’t really understand, the computer players appear more and more incompetent.

3373

If you read the Nature article on AlphaZero you will see that there were a few pieces of knowledge in there:

  • The rules of the game were coded by hand
  • The NN was designed by hand
  • A subtantial number of hyperparameters were adjusted by hand

Great achievement, yes, but not magical. As for the applicability of the technique, we will see, it is far from clear that we get our money back on the computation time required to train the neural networks that essentially learn a heuristic to play with, or to evaluate the moves.

On the gap between Go and Civ, let me make you a list:

  • Initial game states are different from session to session
  • The board is not fully observable
  • The board positions have multiple attributes, several change over time
  • The number of moves possible depends on number of cities, units and technologies
  • The number of units and cities changes over time
  • There are several adversaries
  • Adversaries may or may not be trying to achieve the same victory condition as you
  • Adversaries may switch between those during the game
  • Outcomes of moves depend on chance

2030 was optimistic… more like 2080 perhaps. Let’s all play just Go until then :)

3374

I think they simply do not care. The AI can be done. AoW 3 has a fantastic tactical AI and a decent strategic AI. Also, Civ 5 had a shit AI when it released, and eventually modders put in a real AI to make up for it. Firaxis will simply does not care enough to pay an AI programmer to solve their problems.

I will not be buying another Civ game until I get a green light from other players about the AI not being complete shit.