Jon Shafer's At The Gates

Seems it shows in library and has a Play Now button… I will see. Thanks!

While I don’t always agree with Fraser Brown, I respect his knowledge of the genre. That review just hardened my resolve to hold off.

Hellooooo!

Helloooooooo!

There is no chance he tested this game on a 4k monitor, or had any players test it on a 4k monitor. You literally have to change your desktop resolution to 1080p if you want any chance of reading anything on the screen. You can’t play fullscreen 1080p without the game crashing. It even warns you in the settings.xml that fullscreen is likely to crash the game. So you’re stuck with borderless widescreen, which is awesome, but it won’t do borderless widescreen at anything less than, of course, your desktop resolution. That’s a minor pain in the ass that I just don’t expect to deal with in a game released in 2019.

At one point I thought their content was witty and funny; I guess I didn’t visit their site for awhile and I guess I had gotten older and no longer did …

That’s a lot closer to Colonization- the other colonies are mostly irrelevant to you, the goal being to win independence, which is mostly a matter of having enough stockpiled guns and horses to hold back the King’s horde.

ATG is significantly less complete than Col.

It is a very good core for a game but not much else.

So this seems like it’s the same game that was shown in the old youtube vids; that those vids stopping coincided with Jon’s difficulties; and that he simply came back to wrap it up as quickly as possible, i.e. there wasn’t time or money left for much advancement beyond what was shown.

So that’s how you type it!

Ah, interesting… mobile is a complete blind spot for me.

You need very specific authoring tools that support the development of the NPC game activities, tying together animation of characters and the game internal state. Orkin has a couple white papers describing that.

Regarding debugging, the issue as I see it is that in order to figure out “why” an NPC decided to do an action you need to inspect the search tree. So while possible in principle, it is a task not suited to the human intellect. In effect, even if you’re getting an explanation, this is presented in such a way that we cannot engage with it as we’re overwhelmed by its sheer size. Dealing with this conundrum is a big part of research in the topics of “explainability in AI” and “trusted autonomy”.

That is correct. In my opinion though, the execution component of GOAP is quite trivial, as Orkin himself said, it is all about a 3-state FSM and a plan.

Your view on the matter is interesting, indeed, the behaviour generated by a GOAP system could be modeled as a “dynamic” state machine, where transitions between states change over time. Such a model would be useful to prove properties like “it is guaranteed that we will reach a game state X where property Y” is true, less so to generate the behaviour itself.

GOAP wasn’t deployed in isolation: there is also a “goal selection” or “goal formulation” component. This is best understood - again, computationally - as setting a “constraint” on behaviour, which directs action towards a specific end. The overall architecture is an example of decomposition, where you break down a complex decision problem into several parts, and you solve them separatedly, using the solution in one subproblem to simplify (and set constraints to) other subproblems.

Of 100% relevance to the topic at hand is the issue of 1 unit per tile. As caught up by pretty much everyone who critised the game - from Sulla to @tomchick - it is pretty painful to see the inability of the AI shipped with Civ to orchestrate the movements of units over extended periods of time, in accordance to the pursuit of set goals (such as taking over an enemy civ city).

Beyond the combinatorial challenge of finding efficient paths to positions that enable taking specific actions (such as attack), an additional challenge is that units can be either interrupted or may get into other units paths. That is called “temporal” planning, and it is a very active topic of research with several live applications in the aerospace, logistics and resources industries.

As a mechanic, for humans is great - that’s pretty much one of the ingredients in the Panzer General formula. Without the proper computational framework to handle that task, what you see is that the humans vs AI gameplay is flawed pretty much by “design”, granting a massive advantage to human players who can handle it.

That’s pretty cool @TheWombat, Atomic games were very strong for their time.

Effective pathfinding can be very time consuming, so what you see there is somebody trying to get away with finding paths on an “abstracted” navigation graph, rather than the one that you see on the screen.

Now that the game is out, I feel like I can talk a bit more freely about what I have seen while in the closed beta.

At the Gates is essentially Jon Shafer’s sandbox to test his ideas, some original, some taken from other games like KODP for instance, and explore the range of possible types of gameplay enabled by those ideas.

The game has a unique approach to the 4X genre, as your units are quite diverse and very specialized. I do like very much the early game (the first 100 turns or so) as with every roll of the map initial conditions I need to come up with a different strategy (and plans) to develop and thrive. That does work very well and no other strategy game I know does it so well.

And that’s it. The rest is in comparison to the early game activities just half baked, a place holder or a proof of concept.

So in short, at best this is an Early Access game. I don’t want to think about what the “at worst” could be.

With this one and Offworld Trading Company, I got insights into how strategy game are made to from ideation to completion. The later may mean different things to different people.

So I am satisfied, yet I am not sure the price paid by Jon was totally worth it. Certainly not from my point of view.

Just checked on Steam for early impressions… Game doesn’t look like a day-1 purchase for me now. Those negative comments are valid concerns.

I enjoyed my first day of gameplay. Clans, their traits, and the random array of resources added up to an engaging set of problems to solve.

The bandits were just starting to intrude as I ended gameplay for the day.

I was not looking for a game centered on battles, so the lack did not bother me. Actually, my only complaints were little interface things. I found I needed a pen and paper list of my clans because the only list I could see in-game was limited to the ones within the settlement.

And with many clans demanding professions of specific qualities (in the settlement, not in the settlement, social, not social, etc.) it was not always clear to me exactly what qualities each profession had. But maybe I was missing something, I often do.

But I would say that anyone who is looking for a game primarily focused on the struggle between the barbarian tribes is not going to be happy with the game, at least as far as I got.

Glad to hear it! I bought this today despite my concerns about late game play, both to support Jon and because I think I’ll get enough from the current build to rationalize its relatively low price. I won’t get a chance to play it before the weekend

I got some early access thing from the kickstarter and didn’t touch it until the last 2 weeks and then dropped 20 hours on it.

To echo other’s thoughts, the early game is interesting. I especially like the “play to the map and the clans who show up” aspect in keeping the tribe fed and teched up to stone blocks / permanent structures. After that it’s just a “how quickly can I get the victory” game (50 steel weapons and armor with 5 clans trained up as legions). I don’t think I’ve seen any romans on the map ever, just other tribes.

Well and I have broken my sorta “rule” over the last year so many times it makes me doubt my preaching.

But I do feel sorta bad voting for At the Gates for tonight’s stream … I think it may still be rough.

(Scott it is fun to consider Pillars Turnbased! I mean wow!)

This game is fun so far. Definitely a brain-burner. I read the whole guide before starting but I’m still making mistakes like stocking up on five whole tools and then realizing the profession I want to train is 20 each. Each time I have to earn gold with the settlement I kick myself. Definitely working as intended early on, I think, and will keep me busy.

Why did you do that.

Well Tom’s stream went well. He got into quite a few systems. Though I am not sure it was a roundhouse recommendation… he played for over 2 hours. FYI.

WHERE IS TEH CLAN INFO SCREEN F8 DUZN"T WORK ON MY COMUPTER???SO MAD

-Tom