Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Topics - keith.lamothe

Pages: [1] 2 3 ... 12
1
AI War / Private alpha of Vengeance expansion has started
« on: May 18, 2013, 12:56:17 PM »
FYI, last night I asked some of you if you'd be interested in participating in a private alpha of the next AIW expansion (Vengeance of the Machine) and today I've sent out instructions to those who responded positively (there are those who lack sufficient quantities of this substance known as "time"; I'll take your word for it; I've never had any).

If you didn't get contacted it's not a personal slight or whatever, I was mostly going off the top of my head.  There were also certain Skyward Collapse testers that Chris would kill me if I took them away before SC's release ;)  Also we don't want to wear anyone out and if you're already pouring substantial parts of your days into testing one of our games it might be a bit much to jump right into doing so with another.

Anyway, if you're interested in joining the private alpha please PM me and I'll keep you in mind for the next wave.

It's quite possible that we will skip the public beta phase and just go from private alpha to release (targeting something like June 10th or June 13th for 7.0 and Vengeance's official release), it depends on how stable things seem and such.  This expansion does add some fairly complex new stuff but it's probably easier to test than Fallen Spire or Champions were.  On the other hand we're not 100% decided against a public beta either, it's just a matter of seeing what makes sense.

Those of you in the private alpha, feel free to discuss it with whoever or wherever, just try to be clear if you're talking about something that's expansion-specific so people don't get too confused.

For reference, here's what's been implemented (and in testing) thus far:
- Dual AI Types.  You can give each AI player an additional AI type in the lobby now, if you want to.
- Core Turret Controllers.  Seeded similarly to Fabricators, in all games with this expansion enabled.
- New AI Plot: Hunter
- New AI Plot: Shark
- New Easy AI Type: Chivalric (I know, the adjective form is probably Chivalrous; will deal with that later)
- New Easy AI Type: Bouncer (may not be so easy, we'll see)
- New Medium AI Type: Artillerist
- New Medium AI Type: Exotic
- New Medium AI Type: Reservist
- New Hard AI Type: Vorpal
- New Hard AI Type: Vengeful
- New "Red Hard" AI Type: Vicious Exotic
- New Map Type: Wheel
- New Map Type: Honeycomb

2
AI War / Poll: Bonus Ship Type ideas for the Vengeance expansion
« on: May 16, 2013, 02:38:52 PM »
Nominations from here.

To clarify, a "Bonus Ship Type" is one you can get with your original homeworld(s) pick and from an Advanced Research Station.  Most are fleet ships, but in the last expansion we added the first Bonus Starship and some of the above would fall in that category.  Remember that the AI can get these too :)

So let us know which of the above community-submitted ideas you like best :)  We have some ideas of our own that may make it in to the expansion too, we'll see.

6 picks per voter, I'll probably go with the top 5 or 6, depending on feasibility.

Thanks for your feedback :)

3
Nominations from here.

To clarify, the AI gets about 8 total Core Guard Posts per AI HW.  Out of that 8, between 0 and 3 (difficulty dependant) are picked from the "Brutal" list.  The poll for the new Brutal types is in the other thread.  This is the poll for the non-Brutal ones.  So these should be nasty, and impact the endgame homeworld assaults, but they're not really intended to singlehandedly be gamechangers.

So let us know which of the above community-submitted ideas you like best :)  We have some ideas of our own that may make it in to the expansion too, we'll see.

3 picks per voter, I'll probably go with the top 2 or 3, depending on feasibility.

Thanks for your feedback :)

4
Nominations from here.

To clarify, the AI gets a very limited number of "Brutal" Core Guard Posts per AI HW.  On Diff 7 this is generally 1, on Diff 10 it's generally 3, and varying degrees between the two.  These are meant to significantly shape the endgame homeworld assaults.

So let us know which of the above community-submitted ideas you like best :)  We have some ideas of our own that may make it in to the expansion too, we'll see.

3 picks per voter, I'll probably go with the top 2 or 3, depending on feasibility.

Thanks for your feedback :)

5
AI War / CPA Timing
« on: May 14, 2013, 05:07:08 PM »
I'm looking at the CPA code at the moment (related to an expansion plot) and am rejiggering some very old timing code to be a bit more malleable (the current timings are fine for the normal case, just so I can tweak for plots, etc) and due to how the randomization is done I'm not 100% sure what the practical results of what I'm seeing would be.  So I wanted to check if my guess is actually what y'all generally see so I don't accidentally change the timing during the refactor.

From the code I can see that:

On Diff 7, the minimum time to the first CPA (and the minimum time between CPAs) is 3 hours, 12 minutes.

On Diff 8, it's 2 hours, 48 minutes.

On Diff 9, it's 2 hours, 24 minutes.

On Diff 10, it's 2 hours.

But that's only the theoretical minimum because once that threshold is reached it actually just gets to make a random roll every 90 seconds from (minimum) to (minimum*4), and if the time since the last CPA (or the beginning of the game, if no CPA has launched) is greater than the roll it announces the CPA (which launches when that visible timer runs out).

So on if the first roll is exactly on the minimum threshold there's actually zero chance of it announcing; exactly 90 seconds later it would get a second roll with some chance.  On Diff 7 it would be rolling a number between 11520 and 46080, and would only proceed if the result were less than 11610.  Or more simply it's rolling 0 to 34560 and hoping for less than 90.  That's roughly a 0.26% chance, I believe.  If it misses that, 90 seconds later it rolls again with a 0.52% chance, etc.

On diff 7, your cumulative chance (from the perspective of "before I make the first roll") of having a cpa is roughly 50% about 36 minutes after reaching the minimum, or a total interval of 3 hours, 48 minutes.
It's roughly 75% at 49 minutes after reaching the minimum, or a total interval of very nearly 4 hours.
It's roughly 90% at 63 minutes after the minimum, or a total interval of 4 hours, 15 minutes.

Those numbers squeeze closer together for the higher difficulties as the denominator on the rolls goes down (34560 on diff 7, 21600 on diff 10, linearly interpolated between I believe).


Anyway, my questions are:

1) Does this accurately describe what you're observing in the game?

2) Does this level of variation feel ok?  Would you prefer it to be less random?  More random?  (not that I'd really be ok with making it all that more, but may as well know what people think).

6
AI War / AI War Beta 6.028-6.030 "So I Rewired It" Released!
« on: May 14, 2013, 12:51:35 PM »
(Crossposted from this blog post)

This one contains a huge chunk of the "base game" (not part of the expansion) work going into 7.0. Following the expansion's theme we've just put a lot of work into updating the AI itself to better deal with the many additions and changes made to the game over the last few years, and also take advantage of new things we've learned in that time.

So fair WARNING: this release has some significant "foundational upheavel", mostly in the AI code, and may have some very strange and wonderful bugs (more strange than wonderful). That said, we've put many hours into testing it both in unit tests and just playing the game and it's been stable, smooth, and fun so I think it's ready for beta release. If/when you run into issues, please let us know! :) And please give us your feedback (positive or negative) on the changes in general. You may not even notice them very much depending on the situation, depending on how much time you spend actually looking at what the AI units are doing.

Anyway, a few highlights on the AI changes:

- The AI thread itself now runs much faster and thus the AI is now able to respond to changes in the situation more quickly.

- The AI's normal "threat" (as opposed to the "threatfleet") is now better at finding targets it both wants to reach and can reach and is now more responsive to changes in the situation.

- The AI's "wait" behavior (when it's sending units through a wormhole but realizes it shouldn't send them in yet or they'll just die) has been significantly improved to not wait right on the wormhole, to have a better understanding of the actual balance of power (and thus when to go through or not go through), and to "cut bait" if it's waiting too long and a better opportunity is available somewhere else.

- The AI's individual ship targeting priorities have been improved in several ways, including making some of its heavier combatants more serious about taking out the main human threats on the battlefield and making its units less concerned about secondary targets (and thus more concerned about primary targets) in major fleet actions.

- Various strange bugs and oddities that crept in over the years have been excised. There's almost certainly still some hiding out, and others introduced in what we've just done, but now we're in a better position to actually find and fix them. Notably, there are several new debug commands available to players who want to help us find out what's going on when they see some strange AI behavior (particularly those that we can't reproduce from a saved game).


And some other highlights of this release:

- Guardian seeding in a game's initial map generation is now much less painful because each planet won't start at full population, and guardian populations in general will be lower and less dependent on map type (because guardians will no longer spawn for Wormhole Guard Posts). But remember that alerting a planet will (as before) allow it to gradually reinforce up to maximum guardian population. That can be painful ;)

- Cross Planet Attacks now always deploy as carriers, and carriers auto-deploy much more readily in most combat conditions (and can now partially deploy, in those conditions). Also, we undid the recent change that made CPA ships start as threatfleet instead of normal threat, because the AI improvements have made normal threat much more immediately dangerous. Oh, CPAs can also syncrhonize with exos on higher difficulties under certain conditions now. That won't hurt a bit, really.

- Fixed a recent bug that was causing _way_ too many counterattack guard posts to seed during initial map generation.


Update: 6.029 hotfix for a bug from a few versions ago where auto-loading with the ships in a control group with the new "control group always group moves" toggle would cause unhandled errors. Also fixed a bug (probably from the last version, not 100% sure) where threat could get stuck on a planet where the only human unit was a captured core warhead interceptor.

Update: 6.030 hotfix to fix a number of issues found in the recently started private alpha of the expansion.

Enjoy!

This is a standard update that you can download through the in-game updater itself, if you already have 4.000 or later. When you launch the game, you'll see the notice of the update having been found if you're connected to the Internet at the time. If you don't have 4.000 or later, you can download that here.

7
Just looking for a list of bonus ship types that make you guys ragequit or whatever when the AI gets them.  I know some of you never back out for such a reason, and some of you would probably ban half of the types if you could, but I wanted to ask more directly.

8
(Crossposted from this blog post)

This one contains many visible changes but mainly a lot of behind-the-scenes changes in preparation for the upcoming "Vengeance of the Machine" expansion (more details to come on that fairly soon).

Some of those changes are peeking out from behind the curtain: the AI's Guardians are now becoming another source of variety between games (i.e. Replay Value). So we're revising the existing guardians to be more notable so that you'll care whether the AI rolled Laser Guardians or Artillery Guardians or Special Forces Guardians (or all three) or whatever. Right now there are only 19 guardian types to choose from (Tachyon Guardians are separate) so there's something like a 40% chance in any given game that one guardian type will be in the initial 3 picks for both fo the AI players. That's not too bad, as far as variety goes, but one goal of the expansion is to add substantially to that guardian variety. We'll see!

Other notable changes include some significant follow-up buffs to the Enclave after its recent re-imagination, fixes for some longstanding (mainly edge-case) AI behavior quirks, and an increase in the number of rolls of tissue paper used to armor Advanced Factories.

Update: 6.027 hotfix to get some for-next-expansion stuff back behind the scenes (sounds of something being hit with a cane). "Shoo! Shoo!"

Enjoy!

This is a standard update that you can download through the in-game updater itself, if you already have 4.000 or later. When you launch the game, you'll see the notice of the update having been found if you're connected to the Internet at the time. If you don't have 4.000 or later, you can download that here.

9
For this expansion I'm hoping to add 3 new "brutal" core post types and 3 new "non-brutal" core post types. 

To clarify, "Core Guard Posts" are the ones that go on the AI Homeworlds and (in most cases) must be destroyed before you can kill the home command stations themselves.  The "brutal" ones are like the Core Raid Engine or Wrath Lance or Core Eyes where the AI gets a small number of them per homeworld (generally 1 each on diff 7, and 3 each on diff 10, etc).  The non-brutal ones are used for the remaining "slots" (8 minus the number of brutals) on each AI HW, and are still quite dangerous but not likely to completely wipe your fleet or directly cause you to lose the game.

Every nominated idea should have a mantis record and indeed that's a good place to find other suggestions for these though there's no specific category for them (iirc).  But this thread is a good place for an "all-together" discussion so we don't all have to keep up with the comment thread on every single linked mantis record, etc.

Anyway, I'll let this thread go for a few days before actually starting a poll.

Thanks for the feedback :)

10
So we're gearing up for this.  Right now I'm planning on 6 new bonus ship types though there's some play in that number if we get more really awesome ideas than that (though to some extent we want to save some for later, too).

To clarify, by "Bonus Ship" I mean a type of ship that can be built by the human player at the Space Dock or the Starship Constructor, but can only be built if it was granted by the homeworld(s) initially chosen in the lobby or obtained by capturing an Advanced Research Station.

Every nominated idea should have a mantis record and indeed that's an excellent place to find bonus ship suggestions to nominate (just filter by Category = "Suggestion - New Unit Ideas - Bonus Ships").  But this thread is a good place for an "all-together" discussion so we don't all have to keep up with the comment thread on every single linked mantis record, etc.

Anyway, I'll probably let this thread go for a few days before actually starting a poll, not needing to rush this :)

11
This is just preliminary announcement so no images or release dates or preorder dates or even tons of detail, but since we've had to change timeframes on the Exodus project I'm now gearing up to work on the next AI War expansion: Vengeance of the Machine.

You puny human players will of course get some new toys (actually, considerably more than a few), but the theme here is making the AI more interesting and more varied an opponent.  That includes "make the game harder" stuff (the new "Shark" AI Plot, as I'm currently calling it) but focuses on "make the game more interesting/varied" by adding more variation in what each AI planet's "subcommander" does (can pick different command stations, control guard post selection and placement, etc) and what guardians each AI gets in a specific game, stuff like that.

I'm going to be asking for some feedback soon, notably about what kind of new Bonus Ship Types, Guardian Types, AI Core Guard Post Types, etc you'd like to see.

Anyway, I figured y'all would like to know ;)

12
Folks,

Thanks for your enthusiasm about the Exodus of the Machine project, and we certainly look forward to bringing it to you this year, but there's been a delay for external reasons (it's still coming out this year, but we don't currently know when during that range).  Nothing we can really control or talk about, unfortunately.  We appreciate your understanding in all this.


But the good news is that this means we can shift our next planned AIW expansion forward and get to work on it soon :)

13
Off Topic / Leviathan: Warships
« on: May 01, 2013, 10:50:24 AM »
http://store.steampowered.com/app/202270/

Anyone tried it?

Is it any good?

The idea is very compelling to me as I love turn-based tactics and I love configuring sci-fi ships (surface navy in this case, but I think I saw some energy weapons in there).  So I really want this to be good.

But it's paradox-published, and I'm wondering if you can see the game through the bugs yet :)

14
AI War / Taking a look at Guardians
« on: April 29, 2013, 01:34:36 PM »
I'd like to take a look at the stats of all the AI's guardians with a view to a few things:

1) Making the AI not always have every guardian type in every game, but rather having a certain (preferably small) "core set" that's always there and a number of "unlocked" guardian types at the start of the game, with additional unlocks coming as AIP goes up (and/or, potentially, hacking-antagonism).
- But in order for this to work, each optional guardian type needs to be at least roughly equivalent to the others in danger.  Or have a different "unlock category" so that it doesn't think Warp Gate Guardians are in any way comparable to Raider Guardians, etc.  It may be that "core" and "unlockable" are all we need in terms of those categories, we'll see.

2) Having some good balance targets and categorization system for some new guardian types to be added in the next expansion.

3) Improving the current numeric balance and bringing these a bit "up to date", for example:
- making the bonuses even for each guardian type (so no unit having x2 vs one type and x5 vs another or whatever)
- make sure the hull bonus types each has make sense
- making armor and armor-piercing more rare and more deliberate rather than "everything has it", as that seems the most likely near-term improvement to the armor situation.

So, the current stats (of mkI guardians, on epic combat style; high caps though that doesn't matter as guardians don't scale; may be some odd rounding/urounding of certain numbers compared to actual in-game):

ObjectTypeHealthBaseDPSBonusDPSHull TypeArmorMovement
Speed
Attack
Range
Armor
Piercing
Shot
Power
Targets
Hit
Per
Shot
Shots
Per
Salvo
Seconds
Per
Salvo
Highest
Bonus
Bonuses
Even?
BonusesSpecial
Ability
Range
AIGuardianArtillery4000001333366666Artillery200281800032000016000011125FALSEHeavy;Turret;UltraHeavy0
AIGuardianBeam11199217425074250Medium3002810000500330009141TRUE0
AIGuardianCarrier800000018605580Medium3005060000600131103FALSEHeavy0
AIGuardianEMP111992118605580Medium3007060000600131103FALSEHeavy0
AIGuardianFlak11199211375068750Medium1002840000550055105FALSECloseCombat;Refractive;Composite0
AIGuardianGravity800000666720000Artillery30024100000400001163FALSEHeavy;Neutron;UltraHeavy7000
AIGuardianLaser8798821360068000Medium30034800030034001825TRUELight;Refractive;Polycrystal0
AIGuardianLightning11199216666766667Medium60022400099999960002001181TRUE0
AIGuardianRaider4000001375068750UltraLight3759290000110001545TRUECommandGrade;UltraHeavy;Structural0
AIGuardianSelfDestruction200002000000060000000UltraLight092-1000999999100000200103FALSEArtillery;Swarmer;Turret0
AIGuardianSniper4000001583363333UltraLight15022999999000999999190001564FALSEMedium;Polycrystal0
AIGuardianSpecialForcesRally80000010001000Medium032600001000011101TRUE0
AIGuardianSpider4000001000040000UltraLight150229999990000300012064FALSEMedium;Polycrystal0
AIGuardianSpireImplosion8000000.25%0.25%Medium040800099999911111TRUE0
AIGuardianStarshipDisassembler240000020006000Medium0386000800040001123FALSEHeavy;UltraHeavy0
AIGuardianTachyon80000020002000Medium150060000400015101TRUE6500
AIGuardianTractor11199211750052500Medium45022600003500015103FALSETurret1700
AIGuardianVampire800000533316000Medium505810000040001433FALSEMedium;Neutron;Polycrystal0
AIGuardianWarpGate1600000080008000UltraHeavy0226000040001421TRUE0
AIGuardianZombie2666666000060000Medium30032100000600001221TRUE0

Thoughts?

15
(Crossposted from this blog post)

This one ... does a lot.

In this release we're mainly we're answering a recent community poll by making its top results more attractive places to spend Knowledge (though the Raid Starship actually got its treatment for this in 6.023). Specifically:

1) Turrets in general. The combat stats are something of a sidegrade, though on the buff side of that line. The main thing here is bringing their K costs down (as well as making them more regular across different turret types).

2) Command station upgrades (like the MkII Economic Command Station, etc) now also provide a hefty bonus to your homeworld in the form of a foldout that provides the same benefit as the unlocked station. These are cumulative, so unlocking mkII and mkIII military command stations and mkII logistical command stations will put both a mkII military, mkIII military, and mkII logistics foldout on your homeworld. Incidentally, that gives everything on the planet a +100% to firepower and completely negates AI cloaking on that planet (both from the mil III) as well as slowing down enemies (from the log II) so that might not be a bad idea. If it didn't cost an arm and a leg in Knowledge, that is.

3) The Mobile Repair Station has taken over the Enclave Starship's role of fleetship production in the field (which incidentally makes that a base-game ability now, for the two of you who don't have CoN), been officially renamed the Mobile Space Dock, and come down from 4000 to 3000 Knowledge.

4) Speaking of the Enclave Starship, it needs to stop winning these polls ;) So it's been reinvented a bit. As mentioned above it no longer builds fleet ships in the field. It also no longer builds its unique-to-the-Enclave drones in the normal fashion, but instead generates them automatically much like a Spire Blade Spawner or Tackle Drone Launcher. The result is something much easier to use for just getting stuff dead and something much closer to the original "carrier" inspiration behind the Enclave Starship. K costs have also been adjusted to match most of the other starships.

5) The MkII Transport is now the Assault Transport. Basically the MkI Transport does such an exemplary job in the "Ship Bus" role that there wasn't much reason to spend 4000 Knowledge on a better transport. So the Assault Transport now provides several new tactical options through: cloaking, guns (that get stronger the more ships are loaded in the transport), instantaneous unload, more durability, and not delaying the first shot of unloaded ships as long as the MkI does. We'd also toyed with the idea of it dropping a shortlived forcefield on death, but that seemed a bit too gimmicky and a bit too easy to cheese (without still more gimmicks to prevent the cheese). Knowledge cost is now 3000.

A pretty good haul of buffs for the human side, eh?

Well the AI isn't happy. Not a bit.

A few key improvements to carrier/cpa logic should help make those more "interesting", and the AI's response to Knowledge Hacking has been... augmented, you could say. You won't notice much difference in the first K hack, and maybe a minor difference in the second hack, but from there on it can get considerably more dicey. None of this "9 Knowledge Hacks on Diff 10" shenaniga-tomfoolery, to use an example which is obviously hypothetical and not at all related to a recent AAR win.

Also we've tuned the recent Starship buffs back a bit as they were very much getting into the territory of overshadowing fleet ships despite the lower "total" available. This set of changes is relatively mild and some people think more tuning back is necessary, but some don't, so we'll see.

Based on a lot of thought and player feedback that the game was simply being too nice on this one point, Advanced Research Stations no longer grant mkII of their bonus type for free. Instead they give 500 bonus Knowledge immediately upon capture. Also, though not for strictly-related reasons (having more to do with not saddling mkIII price balance with the adv-factory question) mkIII fleet ship cost has come down from 6000 to 4000. The result is that if you normally unlock mkIII of every ARS bonus type you'll see absolutely no difference in total K expenditure (-500 + 2500 + 4000 vs 6000), but that if you don't it's something of a nerf.

The release contains a few other changes as well, notably the addition of a couple control toggles that should make life more pleasant for people wanting to "automatically dump any resources over 1.5M metal+crystal into mercenary production (or building that fort or trader toy, or whatever)" or "don't let cloaked or immobile AI units prevent automatic harvester rebuilding".

Update: 6.025 hotfix to fix a bug preventing most foldouts from spawning in multiplayer.

Enjoy!

This is a standard update that you can download through the in-game updater itself, if you already have 4.000 or later. When you launch the game, you'll see the notice of the update having been found if you're connected to the Internet at the time. If you don't have 4.000 or later, you can download that here.

Pages: [1] 2 3 ... 12