Author Topic: AIW2 v0.756 Released! "Retrieval of the Lost Arks" Plus an Essay, Apparently.  (Read 5484 times)

Offline TheVampire100

  • Master Member
  • *****
  • Posts: 1,382
  • Ordinary Vampire
I noticed that there are no tooltips in science and production tab. Is this a bug or was this removed when the UI changed?

Offline Ovalcircle

  • Full Member
  • ***
  • Posts: 133
  • Meh
I noticed that there are no tooltips in science and production tab. Is this a bug or was this removed when the UI changed?

It was a bug. A hotfix just got released to put them back in.
Winning a game of AI War provides me with a sense of pride and accomplishment.

 THE TRUE POWER OF THE AI

Offline BadgerBadger

  • Arcen Volunteer
  • Hero Member Mark III
  • *****
  • Posts: 1,229
  • BadgerBadgerBadgerBadger
Note that an affix system is also a great way to get flavour in.

Offline etheric42

  • Full Member
  • ***
  • Posts: 107
Okay, thanks.  I think with an affix or a "descriptive affix" system like you were saying it will help.

I will put out a minor counter to a couple of your points:

2. You don't just choose from some options, you use all of them and choose where to deploy them.

3. You're running a really long campaign (4-13 hours) with the same set of units, gradually increasing them, so you'll get used to them by trial and error at worst.

You do choose where your science goes, so if you leave a ship at Mk1 while another is at Mk4 you have effectively chosen not to use the first ship except in a minor way.  And since you can dump a lot of science into a single ship line fast (say double-bumping it to Mk3 early on) you may not have had the time to understand the ship (especially with all the chaos that happens in fights) to know that was a bad dump of science.

You had mentioned in the blog post that in PvE per-unit balance wasn't as important compared to fun factor.  I agree with the compared to fun factor bit, but I'm a fan of the last generation of D&D designer's philosophy: in a PvP environment like Magic, it is okay to have "trash cards" because part of the PvP experience is grading card power levels.  In a PvE experience, it is not fun to get saddled with an underperforming pick.  (of course another facet of that is choice and commitment, you have hundreds of choices in Magic and can change loadout easily between game nights, whereas respeccing in D&D is pretty limited, and you're usually only choosing an option from a list of 10 or less.)

Because of the way research works in AIW, it's possible to have trap picks for players that they can't respec out of, only hope to grow to get more research for a good pick (and hope AIP increases don't eliminate any gain they made from it).  I'm less concerned about the 4-hour games than the 13-hour-or-longer-for-people-who-pause-and-think-a-lot games.  In total war I can scrap a building chain and get my money refunded if the units I get from it aren't any good.  What can I do about research sunk into a bad ship line?

Anyway, don't get bogged down by this comment of mine.  I think random stats in some ways are very good (even traditional cRPGs like Divinity Original Sin randomized loot in chests to increase replayability and combat walkthroughing).  Just... keep it in mind.

(Oh, and the triangle ships and starter ship probably shouldn't have any randomization to combat players rerolling starts.)
« Last Edit: July 26, 2018, 03:39:35 pm by etheric42 »

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Points well taken.  I suppose one way to combat that would be to allow for a form of science refunding, although I'm sure that would be game-able in some fashion.  Though if it destroys all the ships relying on that tech, surely it can't be THAT bad...
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline BadgerBadger

  • Arcen Volunteer
  • Hero Member Mark III
  • *****
  • Posts: 1,229
  • BadgerBadgerBadgerBadger
To my mind, Science is like "Level up points" in an RPG. If I sink all my level up points carelessly then yeah, my game is going to suck. If you take a Wizard and max out his Strength and Dex, well, I'm not going to be a great spellcaster. And careless use of Science will hurt badly even without procedural stats; you could still hose yourself in AIWC.

That said, many RPGs do have a "respec" ability later in game. What if we seed a single "Respec" alien artifact on the map that you can use once to change your science allocations for some Price? Some sort of Devil's Bargain with a time-travelling alien species? That might make a fun thing for post 1.0.

Offline TheVampire100

  • Master Member
  • *****
  • Posts: 1,382
  • Ordinary Vampire
To my mind, Science is like "Level up points" in an RPG. If I sink all my level up points carelessly then yeah, my game is going to suck. If you take a Wizard and max out his Strength and Dex, well, I'm not going to be a great spellcaster. And careless use of Science will hurt badly even without procedural stats; you could still hose yourself in AIWC.

That said, many RPGs do have a "respec" ability later in game. What if we seed a single "Respec" alien artifact on the map that you can use once to change your science allocations for some Price? Some sort of Devil's Bargain with a time-travelling alien species? That might make a fun thing for post 1.0.
Problem is, what happens to all the unlocked ships and other stuff? Do they just go poof?

Offline BadgerBadger

  • Arcen Volunteer
  • Hero Member Mark III
  • *****
  • Posts: 1,229
  • BadgerBadgerBadgerBadger
I imagine if you lose the science then all those ships explode impressively

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
I like the idea of being able to find a respec thing.  Possibly some sort of "incinerator" structure, which lets you choose a ship line of yours to trash and get mega science back out of it.  That sounds like an idea for during Early Access.  I'll put that up on mantis. :)

edit: https://bugtracker.arcengames.com/view.php?id=19875
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Something I had not paid proper attention to were the nature of the levers required to balance units so far in the game.  It's extremely hard to understand, and I can't just use it directly in much of any fashion because it's being translated at runtime into one end product that is based on either AI War Classic data from a spreadsheet, or the data from AI War 2 balance xml files.

The complexity of this has led to a variety of funky things, like certain things costing nothing, or having movement speeds so high that they wrap back around into the negative numbers, or things of that nature.

Right now there are also a ton of things that look like they're trying to auto-balance themselves, mainly because systems were created separately and then linked to ships.  I'm now having the system entries be separate still, but defined as children directly on the ship, meaning that there's less need to look between files and you can better understand what you're looking at directly in one file.

With that in mind, I'm also having battles with myself on which things to put into a lookup (for ease of knowing whether a number is high, low, or indifferent -- what the scale is, essentially), or to have it just be a raw number (for ease of understanding one ship versus another).  I'm handling that on a case by case basis.

What's telling, though, is that I'm having to massage and condense the data already, so even after I do this conversion the data isn't going to be the same as it was before.  Not that the game was well balanced before, so it's no great loss, but it's worth noting.

I'm also thinking ahead to my additions for the randomization, and so trying to make things not depend on so much else in order to calculate themselves.  And thus letting myself calculate strength dynamically, I think, before too long; that will be important with the changes to having things be random and thus the strength values needing to reflect that.

Overall the biggest challenge is just making it so that I can look at an xml entry and know pretty much what it means.  I already thought I could do that, but in the current version there are literally dozens of multipliers on most things, so my understanding of it was almost always flawed.

I know exactly why it was set up this way initially, but as the system has grown it has made it so that you pretty much need to understand most of the system in order to make even small changes.  I'm trying to basically decouple things so that's no longer true.

It's coming along, but I still have a long way to go.  Hopefully by the end of Monday on the non-randomized part of this.  It should fix bugs in a ton of ships, along with being more moddable, although it will probably introduce some new balance issues as well.
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline zeusalmighty

  • Full Member Mark II
  • ***
  • Posts: 151
I played good 10 hours this week (true test; no cheats) and have to say this update has made a huge leap forward in terms of getting ready for early access

Basically, it's fun. The factions interacting with each other is not only enjoyable to watch but it has great implications on how to play your given campaign. The bonus waves against the factions is particularly a significant step towards making the galaxy feel "busy in just the way I was hoping for  :)

Balance is still an issue of course and some of the units need special attention (artillery golem has a tiny range for instance)

I turned off the exo-waves from the hard golems because the player doesn't have access to anything that can effectively neutralize 30K of bombers bee-lining towards your command center. Also, these are very difficult to spot because they all clump up and look like one bomber until they get to your home planet and then holy cow, so many bombers  :o

Nanocaust can't capture their home planet if there's tractor turrets. I rescued them and they really took over (at 10 intensity)

Anyway, things are really looking up!

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
I have to say, that is by far the best news I've heard in a LONG time. :D

Man this has been a journey, and knowing that this is paying off is a huge relief.
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline BadgerBadger

  • Arcen Volunteer
  • Hero Member Mark III
  • *****
  • Posts: 1,229
  • BadgerBadgerBadgerBadger
Poor Nanocaust. Tractors really are its weakness.

Offline dfinlay

  • Newbie Mark III
  • *
  • Posts: 41
So, I haven't had the chance to play in a while so my thoughts may be somewhat out of whack with the current state of the game, but I do have a few thoughts about procedural stats. On the one hand, I'm excited as procedural content is awesome for the reasons you described, but I have a couple of fairly big reservations.

1) As already mentioned, legibility is going to be a huge problem. I realize you've streamlined the tooltips from last game, but this was already a huge problem in AIWC and the purely random stat approach is going to be much worse here. If you have to check every ship in the game and try to figure out how each one is going to perform based on a slew of numbers and then remember them for the rest of the game, that's just going to be too much. The affixes will help a bit, but only really if they always do the same thing. If a ship gets "destructive" when its damage is in the +10% to +25% range and every stat is randomized, it might help a bit to contextualize, but I still have to assess and memorize a ton of numbers. On the other hand, if it gets a "Destructive (+10% damage) affix and that's one of only one or two tweaks it gets this game, that's something I can handle. As such, I pretty strongly recommend a system where one or two (not a bunch) of random affixes get added to ships, which then change their stats.

2) Balance is still important for a few reasons. Sure, in a Roguelike, you don't immediately feel "well, this game is unbalanced" when you have a really strong or really weak run, but it's still not going to be fun to steamroll the AI because you have some ships which rolled high on every stat or to just feel ineffective because a bunch of your ships rolled stats that don't really let them do their jobs. Basically, it makes picking a difficulty setting that will generate the right challenge level that much harder on the players. The best roguelikes (in my opinion) are the ones that make sure that every run feels in the same ballpark for how strong it is, just in different ways (Brogue is really good at this). This problem is even worse in something like AI War. In a lot of games, if you have a good/bad run, it'll likely only last a couple hours anyway and then you try again. In AI War, games are a lot longer (at least for some playstyles) and so you are stuck with it for a much longer time. To that end, I recommend a fairly controlled system for how many bonuses and penalties a ship can have - something like "every ship gets one major positive affix or two minor ones" or "every ship gets one neutral affix or one positive and one negative". If you're going for the free floating stats model, then you might want to do something like how XCOM does random soldiers - they get their base stats and then a fixed amount of stat points spread around all the stats with some caps and some different conversion rates depending on how important the stats are.

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Definitely keeping it small, and also I agree on making it so that things stay within a certain realm of usefulness.

The main problem I want to avoid is having something that is ahead in DPS in certain situations in every game.  And things like the tachyon microfighters don't need to be their own unit, for example.

I want to get away from it being purely a numbers game on this as much as I can, though, I do have to say.  I have a fair bit of work to do, and we're already looking huge on the release notes: https://wiki.arcengames.com/index.php?title=AI_War_2:_The_Era_of_Discovery
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!