Endless Space 2 - Now on Early Access
 

The sequel to the 2012 hit Endless Space is now available on Steam's early access. Endless Space 2 brings the best from Endless Legend, including quests, heroes, and a more immersive rulership experience.



More information.

Space there feels cramped, not spacious enough.

No it does not bring the best. It still has crappy space combat. For me 4x space games need to have good space combat or I am not interested.

Wish there were more "3x" games, like Thea, where you don't have to build an empire and the fight the whole bloody galaxy, just manage one base and focus on combat, trade and exploration instead of endless management tasks.

That sounds like sort of a turn-based Total War "In Spaaaace!". It sounds like a pretty good concept that I'd be down for, but when I play 4x games, I do most commonly play them for precisely that extra "x" you don't like.

Unfortunately no Mac version…. => No buy…

3x games of the type described by Skreeg would be interesting, but what I'd really like to see is more focus on the single player experience in 4X games. The vast majority of people who play these games do so single player, but yet they are almost all designed with sort of a multiplayer framework with all the players starting out equal in their own corner and the AI running the players except the one the human player is. The problem is that the AI can almost never handle the game as well as the player, and you often end up investing a ton of time into a game without any payoff because the computer never puts up a fight at any stage. I'd like to see more games that try something different, kind of like Sorcerer King, where you absolutely are not equal to the other players and the focus is purely on single player. Although I didn't actually like that game so much in practice.

Supposedly that's what Brad Wardell is good at: decent opponent AI. At least that's what gets him excited about programming stuff directly.

Turns out, good AI is hard. I could probably optimize an AI system for your game, but it would win every time because it's maximizing with perfect precision and doesn't make any mistakes unless there are a few I haven't thought through at length before designing it. It's backing off the optimization and giving the opponent a personality and simulated skill level appropriate to the player's expectations that's really difficult.

I guess we disagree.

Good ai should be easier now that processors are faster but maybe the generation of coders have forgotten age old techniques and they just try to 'wing' it without understanding what generations before them have learned with regards to technique.

Keep in mind there's multiple theories running around with regards to AI. Game AI can be as simple as a set of decision trees or priorities for the opponent to take; basically a state machine with a set of goals. Then there's other simple methods like fuzzy logic (choices based on states somewhere between 1 and 0 rather than strict binary decisions). Then there's what we think of as the hot-new AI. That'd be your neural nets, genetic algorithms, and so on; methods promoted heavily in the 80s but abandoned as they failed their promise. They're only more popular now as we have processing power (read: data centers) to employ them effectively.

I was just looking at Google's TensorFlow for a potential application at work so it's on my mind.

So this has to be asked since Sega owns the developer now. So moving on will it use the latest Denuvo Anti-Tamper, and have more then ten DLC packs on release now?:smartass:

I'm not entirely sure that it's feasible to perfectly program an AI to win at any game without fail. Or at least not for the modern strategy games we have today. There are just so many different little individual variations in the way a turn is played, often millions or more in a single turn, which might not have an immediate effect now, but which can lead to success or failure 20+ turns down the line, at which point the number of possible variations to calculate becomes so large I don't even know the number for it. So Ais seem to rely more on general strategies like "always move your unit closer to the enemy" which don't allow for much nuance and which are easy to exploit.

In practice too, given how few games can even program an AI that can even understand the issue of how to use all of the mechanics of the game in a semi-intelligent way, it doesn't seem like they are close to reaching the point where making the AI too smart is a problem yet. With the exception maybe of games that are relatively simple and extensively studied, such as chess.

I personally think that the big problem designers have is that they don't take the AI into account when they are designing their game mechanics. If you want a challenging single player game you can't include mechanics that the AI just isn't going to be able to handle.

Actually, in Endless Legend there is a specific faction, the cultists, that allows you to play with one single town and easy all the micro in quite a great proportion. There are still an awful lot of things that you must manage, but I find the gameplay of this faction quite appealing just because of this.

You could try Polaris Sector. You still build an empire, but everything is pretty much automated. All you do is decide which direction new colonies will lean. I understand designing your own ships is an integral and unavoidable part of the game, so if you don't like that, this one is out. I hear Polaris Sector is kind of like a simpler version of Distant Worlds. That could work for you, too. I hear it can be very in depth, but you can choose to automate anything in the game that doesn't interest you….or everything and just watch the galaxy spin around you. People rave about Distant Worlds, but I've never played it for real. I tried some time ago, but the text was too damn small for me.

I always liked MoO2 for it's turn-based fights. Stars in Shadow looks like a very close cousin to MoO2. People say the board-game-like combat is the best part. I'm always bitching about how 4X devs are always skimping on the combat, so I had to put my money where my mouth is and buy this Early Access game, though I haven't played it yet.

EDIT: Wow, that post had a lot of links to games I haven't actually played. :p

I didn't know that. I wonder if this is a Sega-effect? Endless Space and Endless Legend were both on Mac I think.

On another front, I had hoped they would vastly improve the combat. The combat (or lack of it really) in the first Endless Space ruined the game for me.

That is an interesting phrasing.

4x is explore, expand, exploit, exterminate.

Thea was at its initial release a 4x. It was also one of the few games that made a proper use of the apocalypse craze around.

Devs used a post apocalyptic setting to provide an uncommon formula of the x expansion.

Thea is a world picking up and the flimsy level of consumption of the environment did not allow a steady, robust, permanent type of expansion. So devs opted for a nomadic kind of expansion, based on necessary expeditions to support a settlement.
The idea was brilliant, maybe even new and the execution was on par.

It introduces tension as the survival of the settlement relied on a fragile form of expansion. Nothing was acquired and everything could be lost by a miscalculated attempt to expand.

Then, players do not like tension in gameplay these days, they prefer long tranquil streams that ensure security and a full causality between reward and effort.
So the expand feature got diluted release after release. Maybe indeed up to the point that it might be better to speak no longer of expansion.

Nevertheless, Thea was 4x and the different take on expansion is an organic feature stemming from the post apocalyptic setting.

This makes any port unsure. Thea was thought as a whole, not as a collection of bits to please players who like it, have fun with it.