The one drawback

After working for Gas Powered Games for 11 months I have found only one drawback to this job. While patching Supreme Commander 2, there have been many times when the game is so much better on my dev machine than the version installed at home. This was especially true back before the first AI patch was released. I wanted the patch to be released as much as any of the Supcom 2 fans just so I could have the improved AI on my home machine. The same is true of the upcoming patch. There are so many game improvements that I can't wait to be able to play with at home.

Recently @squareenixdave posted that some new news regarding Supreme Commander 2 should be released soon and I truly can't wait. We have been hard at work doing all sorts of things I wish I could post here. For me, the news can't come fast enough.

Over the past month I have been writing down ideas that I would like to implement for the Kings and Castles AI (assuming I am in charge of it). I have a habit of keeping a notebook nearby and just writing down any wild idea I get. Lately, I have been taking those ideas and trying to form them together into a cohesive plan. I still have a lot of refining to do before I actually put it all into a document.

In Supreme Commander 2 the addition of neural nets gave the AI some tactical thinking capability. It could decide if a single platoon should run or engage and, if it chose to engage, which target to focus fire on first. For Kings and Castles I not only want to improve upon that, I want to focus on the AI's strategy game. I don't want the AI to decide it has enough troops to send out and then send them somewhere solely based on threat values. I want the AI to keep track of enemy threats on the map and when it finds one that is a high priority I want the AI to grab nearby platoons, already in the field, and send them to take the threat out.

Having the AI oversee its platoons, rather than having the platoons basically doing their own thing as they are now, would allow the AI to react faster to high threats. If the AI found an group of enemy long range artillery in range of it's main base it could find all of its nearby platoons that are on lower priority missions and redirect them to this new threat. If the AI sent out a platoon on a low priority mission and suddenly found a large group of incoming enemy units, it could immediately bring that large platoon back to take up a defensive position.

I also don't want to the AI to see threats as single places on the map. I want to improve the AI's threat map by giving units a threat range. That way, those long range artillery (using SupCom 2 as an example) can not only be seen as a large static threat, but as a threat with a long range. Using the threat map for this purpose could give the AI a much better sense of safe and dangerous areas, threats that are very dangerous, and a better sense of who has the better map control. Best of all, it would be cheap to implement performance wise.

Next on my list is to give the AI the concept of front lines. This would also help give the AI a sense of map control as well as allow the AI to post units near the front lines in an attempt to hold them. I want to move away from the AI hording units in its base before sending them out onto the battlefield and instead have the AI build units, send them to forward position near the front line, and send attack waves from there. That way, the AI is not constantly fighting to retake land it had under its control just a minute ago. It would also allow the AI to easily implement advanced tactics such as PD or TML creeps (again, using SupCom 2 as an example) without having to code anything specific to handle it. Just defend the front line.

My goal is to make people say "Wow, I can't believe an AI just did that." I don't necessarily want the AI to wipe the floor with average joe player by default, that isn't any fun. But, I do want the option to make the AI very hard to be there for the hardcore players. It would make my month if Mike Swanson walked into my office and said that they need to give the normal AI a handicap to make it accessible by average joe player. Hell, it would probably make my year.

This is just a small sample of the ideas rattling around in my head. I hope to, over the next few months, refine more ideas and get a battle plan together for truly improving the AI system. Chris seems to be on board with it so far. I can't wait to see some of these ideas in action.


John said... / August 1, 2010 at 2:21 PM  

I'm looking forward to K&C and hearing your ideas for the AI is awesome! I hope you are the one who gets to work on it.

mocafrost said... / August 1, 2010 at 3:25 PM  

Sounds great. Keep it up.

Mazrix said... / August 2, 2010 at 1:53 AM  

Looking forward to eventually playing against the "Sentient AI" option.


Col. Jessep said... / August 2, 2010 at 8:09 AM  

I really hope they put you in charge of the AI in KnC Sorian! From what I have read on the German forum even the top players enjoy playing against your AI from time to time. Many of them would never consider playing against the AI in your average RTS.

Do keep in mind that it might be dangerous if your AI becomes self-aware... ;P

Harrier said... / August 2, 2010 at 10:55 AM  

Sorian, you're probably already aware of this, but maybe it's worth repeating it here: The dearth of patches is seriously hurting the SupCom2 community. The following forum threads are pretty representative for the current community sentiment:

Everyone realizes that GPG has limited resources and that testing/QA takes time, but no one understands why GPG essentially stopped balancing the game a few months ago or why the pre-order maps are still unavailable to most players.

There are some well-known balancing issues that ruin ranked gaming for a lot of people and that could be easily fixed by some simple changes to unit stats, but apparently such issues don't have any priority for GPG or SE.

Sorian said... / August 2, 2010 at 11:39 AM  

First off, GPG is not responsible for the patch schedule.

Second, the last patch was June 3rd, and this was after how many other updates? I don't think there is a dearth of patches at all. Steveb already told the community that the next patch would take some time.

If you truly want to get the game a balance patch then I suggest you contact Chris Taylor and Dave from SE (mostly Dave, since it would be SE's call).

Derek said... / August 4, 2010 at 5:43 AM  

I love SC2, but sorian, I do have 1 question,

Have you worked on the ability of the AI to engage the opponent when the opponent attacks from the water?

It just seems (regardless of AI setting) that if a player goes navy he can basically bombard the AI into submission,


Fingolfin said... / August 4, 2010 at 10:14 AM  

Well, let's cross our fingers for a rebalance of the above issues, some new units and more maps, novel pathfinding/UI tweaks, and some awesome GPG-esq surprises (such as a new UEF hunker-to-ambush unit called 'The Cavedog' – just for lols!)

Thanks for the updates anyway Sorian.

Sorian said... / August 4, 2010 at 11:30 AM  

@Derek: Work on the KnC AI has not started yet. For SC2, I can't give out patch details quite yet.

Derek said... / August 4, 2010 at 1:34 PM  

Hello Again,

Thanks for getting back to me. Like yourself I love playing with or against AI and see how they adapt to different situations, I thought this might be interesting news to you. I ran a tournament between the AI series on Setons Clutch and I noticed a striking pattern,

The A.I. which is put on the top of Setons clutch vs an idential AI on the bottom, both in race and series wins nearly 100% of the time. I am not sure why this is, but AI on the top of the map always fairs better.

Also, even on Setons which seems to be one of the largest maps, AI AIR, AI BALANCED, AI NAVY always seem to lose to RUSH, or LAND. In fact based on my results, RUSH seems to win nearly 90% of every match.

If RUSH is more for smaller maps, why is it still dominating all other AI's even on the largest maps, which include naval testing range?

It seems if you want AI as a partner, or want a stiff fight, the obvious and easy answer is to select RUSH.

Even if you try to mix say a turtle and rush or turtle and balanced, it still isn't as good as two rush ai's together.

I thought perhaps a strange combination of turtle AI and maybe navy or air on naval test range would be able to counter 2 rush AI's, but sadly 2 rush ai's always win. I must say, each match is uniquely different, but even in games that last 40 minutes where tech should outstart to outperform numbers, rush still wins

Sorian said... / August 4, 2010 at 1:55 PM  

@Derek: Probably the biggest factor, regarding Rush vs other AIs, is the AI has not had any serious work done to its research priority list since release.

As far as map positions, I have no idea. Maybe the way the path nodes are laid out?

I have a nice long list of things I would like to fix in the AI. Hopefully I get to do the majority of it.

Derek said... / August 4, 2010 at 5:07 PM  

I know you are limited in what you can say regarding the next patch, but maybe you can answer this,

I know you have a laundry list of things you'd like to improve in the AI for SC2, have you been able to do any of these things in the next patch??? It would be awsome if you have.

Sorian said... / August 5, 2010 at 12:01 AM  

@Derek: Again, I can't comment. :(

Fingolfin said... / August 5, 2010 at 1:15 AM  

Have to admire his persistence! :)

Sorian said... / August 5, 2010 at 1:27 AM  

@Fingolfin: I don't blame him. There hasn't been much in the way of news as of late. I think the KnC blogs are spoiling people.

M said... / August 11, 2010 at 4:20 PM  
This comment has been removed by the author.
MarkTheRobotSlayer said... / August 11, 2010 at 4:22 PM  

There will be a new patch though, right?

Sorian said... / August 11, 2010 at 4:30 PM  


MarkTheRobotSlayer said... / August 11, 2010 at 6:49 PM  

Any idea when it will be released?

Derek said... / August 12, 2010 at 6:21 AM  

Darn! Lost a bet due to strange Overlord play

I bet my buddy he and some random player online couldn't beat 2 cheating AI's on Setons, one was turtle and the other was Air. I did this so they couldn't get bottlenecked in the middle...Anyway

Darn Overlord ability took both AI's and totally messed up their build orders and what not, cheating AI's got smoked.

I guess nothing is perfect, perhaps the odd game here and there the overlord actually takes away from the ability of the AI's to win?

You ever see this when you were testing? 99% good, 1% bad kinda deal?

Sorian said... / August 12, 2010 at 4:22 PM  

@Mark: No idea. That is all up to Square.

@Derek: AI Overlord does not change build orders, it was probably something else.

Derek said... / August 12, 2010 at 6:10 PM  

you would know better than anyone obviously, it was just weird, imagine this

2v2 game, 2 cheating ai vs 2 humans on setons, and even after 21 minutes the turtle AI didn't have a single artillery up or nuke and it was uef! i was shocked

Mazrix said... / August 20, 2010 at 12:39 PM  

I hereby charge you with the murder of one Jessica by internal teleportation.

How do you plead?


Sorian said... / August 20, 2010 at 12:55 PM  

Hey, I warned him!

Larry said... / August 25, 2010 at 1:40 PM  

Hi Sorian -

Quick suggestion for A.I. to add to your list if you don't have it.

Make offensive units "range aware" of defensive structures. For example, when a group of A.I. controlled tanks and mmls charge on Setons, they'll often encounter player-built clusters of point defense. The entire blob engages the PDs, placing the mmls within striking distance of the PDs. Perhaps in the Hard and Cheating AIs, you could add intelligence that keeps longer-ranged offensive units from stumbling into shorter-ranged defensive structures.

- Vax

Kirsten said... / August 31, 2010 at 10:52 PM  

amazing job on the new AI, its excellent

Derek said... / September 1, 2010 at 12:06 PM  

i think the ai is amazing, totally better than before

though i do have a question, is it a mistake or by design that the disruptor station does not increase in range with the tech upgrade to increase arty range? It seems that once upgraded, the regular long range arty has greater range than the disruptor station? Doesn't make a lot of sense?

Fingolfin said... / September 2, 2010 at 11:31 AM  

New patch is awesome! Top job on the AI too, massive improvement (and it was already the best RTS AI ever made).

Murray said... / September 7, 2010 at 2:02 AM  

New Patch is sweet

Additional ideas for Kings & Castles could be that you could have modified armour types for your unit that you could modify maybe before or in game to either suit your play style or aesthitic style

Having of course the famously good zoom and build queue features of almost all previous GPG games as well

Also add Golems and Hydras :D

Post a Comment