On Being A Commander

I believe what you are now talking about is an AI as part of a UI, which by the very definition is like a chain of command where you tell the AI what you want it to do for you by means of the UI...

So you say:
Go build a base and the AI goes and builds a base for you.
Go attack this guy and the AI goes off to engage the enemy.
Get more resources and the AI goes off an builds you a bunch more power plants and mass generators.

I can categorically say that I do not want this in the game. Period. I just cannot see how this appeals to anyone at all. It totally disconnects the player from the finer details of the game that made the previous titles so appealing to RTS players.

As far as the rational posts in this topic are concerned, we've already discussed that the existing controls available to a player are quite powerful and many of them do similar things to those which you are talking about, like building groups of buildings or area commands. This isn't AI but in the case of templates for example, simple player created macros.

In SC a building could be rebuilt from wreckage, dramatically reducing the resources required and build time. As far as I recall it could only be done by the SACU as explained in the link. A simple area command for this function given to an engineer would be more than adequate for PA. For more on area commands I think there is some required reading in the Ideas from Zero-K topic.


To summarise:
- Selecting a part of your base and toggling a 'rebuild' switch = Unit area command.
- Have 'smart' base templates = Just no. Let's go for player generated templates instead.
- Give the players a clear resource interface that allows them to select on a global level what their priorities in construction are going to be. Let the 'AI' then allocate resources to your priorities. Allow local individual overrides so it's easy for the player to focus their entire economy on a single project when necessary. = I think again that there is a priority function in Z-K that works a bit like this. AI is really not the right word here, it's just a priority equation...

Casually throwing the word AI around is not helping your cause. It's not really appropriate here and frankly, it will put people off your concept, which in itself is actually relevant.

Good call, that is a simpler solution - 100% behind the idea of area commands.

Ah, this is what I get for skim reading, I should have read more thoroughly. I saw "base template" and thought Templates from FA, which is where I would like this to go.

Ok so I guess this concept is arguing more for UI power than the thoughts are evoked with the word "AI"

I believe it should be possible to 'sub in' a computer AI for a missing player, but in that case it should be exactly the same as if you had another player helping you control your forces. It would be nice to be have an interface to give each other goals and targets, and thereby give your teammate indications on what you're doing (and thereby guiding an AI player as well). Generally though, I suspect an AI player sharing control with you may be more frustrating than helpful.

This may however open up a line of thinking on how best to communicate battle plans with your teammates - the more you know about your teammate's plans (be he an AI or a human) the better you'll be able to coordinate and be effective, and I believe a well fleshed out system would enhance team play considerably.

Just off the top of my head, what if you could collaborate on attack plans? Instead of just drawing lines the other player can see (ala TA Demo Recorder / Spring / ZeroK), why not allow players to 'draw' orders, and then allow different teammates to 'claim' a section of orders and assign units to carry out those orders?

Just saw the other two posts. Answering a suggestion with 'oh area commands will handle it' actually ignores the fact that what I'm suggesting is different from just simple area commands.

Wolfdogg, I think you completely missed my point about AI / UI. Giving out vague orders like 'go attack the enemy' and expecting an AI to do that for you is completely different than saying "Squad 10 attack from this direction, Squad 22 attack from the other direction, computer, figure out how to make sure they do it at the same time". And while ZeroK etc have great ideas, they aren't perfect, and it's the attention to detail that is going to make the difference on improving those interfaces.

A structure rebuild toggle, for example, is different in a slight but significant way. Namely, whether the player has to repeatedly issue a command in order for units to act on it. If someone bombs my radar tower, I may want it back up, and always want it back up. If I'm looking at another planet and someone bombs my radar tower on my home base planet, I should be able to expect to have that building replaced.

On 'Smart Templates' I'd like to point out that Supcom's version of base templates was experimental, integrated poorly with the interface, and was difficult for the user to set up and control. It's a good idea, I'm just suggesting once again that the player needs to be in better control, and to put the computer to work giving useful suggestions to the player. I think we're all on the same page with this one though. Player controlled, templates of some kind that allow you lay several buildings in a predetermined pattern.

So I'm suggesting something like being able to pick a template, and then drag a line out on the ground, and the template will repeat itself. Or to have a 'firebase' template that works like an area command in Zero-K - Click template, click on ground, drag out circle. Based on the size of the circle and the orientation of the cursor, the UI could pick from a number of different player-created templates.
Small Circle - LLT surrounded by DTs
Line drag - LLT surrounded by DT, spaced every 10 squares with DTs in between
Big Circle - Several LLTs in front, with a HLT behind them and a plasma cannon, facing in the direction indicated by the player.

Yes, to everything Pawz just wrote.

Seconded. Never used the template function in SupCom but the concept sounds pretty good.

I think OP has an interesting idea but don't feel it would fit what this game was pitched as. and despite the scale of this game it could be a very dry slow experience if executed poorly. Also I think that the abilitys of the AI has only recently become clever enough to achieve some of these ideas. I think the intel/counter intel ideas could be rather good in this game.

Just looked up Commander on wikipedia and it sounds like a commander is in charge of One big ship, small shore activity or an aviation squadron. not quite as grand as what you implied.

Command and Conquer: Tiberian sun had a Waypoints system that was useful but did crowd the map a bit. You would pot waypoints on the map independent from units and when you sent a unit(s) to the first point they would follow the points. I thought this had a lot of potential and meant you could concentrate on your base as waves of units moved in to war or a staging ground.

I am interested in seeing the implementation of helper AIs. I like the idea of the attack or defend split as building the base has always been one of my favorite parts of RTS games. but as a turtler and a staunch single-player only kind of gamer I'd be uninclined to expand beyond what I can control and would just set the AI so I could beat it playing how I want.


* haven't quite finished reading all this thread but only about a page to go

This.

Call it "helper AI", "assistant AI", "smart template", "squad AI", "advanced UI", "area orders" or what you want, the point is to have your units do what you want with less clicks and focus time. So yes, it's basically UI power.

That's the meaning of the word in a few of today's Navies ; the word means and meant many other things across History. The original commanders, for example, actually controlled some land. Without going that far, military grades don't always have the same meaning depending on the time period, nation or army branch. As such, Commander can mean what we want, from ship, installation or unit commanding officer to chief of staff.
SupCom actually takes another road, by having Commander being a function name, independent from the actual pilot's rank (apart from the Cybran who don't use ranks anyway).
Here, a Commander would be a mix between a warlord, a paratrooper field officer and a ronin somehow...

While I am not much in support of this for units, with the exception of allowing standing orders (ie, your transports won't stop ferrying units because the transport that created the route was destroyed), I do like the idea of integrating units with the base template system.

Create a base with an assigned resource income for AI controlled options (ie, any unit/building construction or replacements made by the base AI), and assign it a certain amount of engineers, and it will maintain itself, while you focus on building and managing your armies, and fighting the battles. If this was integrated with the unit order system, then you could do something such as create a patrol route for combat units to defend the base, and assign, say, 30 ASFs to it, and they could be replaced when destroyed when it fits within the bases budget. Perhaps the could be queued with additional, nonessential tasks that would be performed merely to maintain exactly the allotted resource usage, such as intelligently assigning engineers to assist.

The resource limits would allow the base micro to be reduced to the point where the player only has to maintain the edgeline construction to keep from wasting resources, as the bases could maintain the baseline usage.

And of course, all of this behavior should be togglable within the template system, and nonexistant without.

I don't agree with AI taking over high-level functionality and decision-making, such as base design, overall troop movements, etc. This is the realm of player control, and the priority should be focused on making this as easy to manage as possible.

With base design, there are three areas of focus: Building the initial base, maintaining the base, and rebuilding the base. Building the base can be cumbersome because for the player, it's just a lot of extra clicks to do a repeatable process at the beginning of each game. (Or every time they expand, especially to a new planet.) We already have a solution for this, however, that being building templates, which were already implemented in FA. There's definitely a lot of improvement that can be done to make them smoother to implement and more mutable depending on the terrain and other factors. A strong template system, or base planner should solve the first area of focus.

Maintaining the base involves expanding, upgrading, controlling economy, and repairing minor damage. Expanding is very situational, and an AI couldn't accurately make the decision on how to extend their base. However, this doesn't require a lot of extra clicks on the players part, and the worst of it can again be solved by templates. Upgrades will probably be more limited than FA, and if it follows the TA style, it will be more along the lines of extending the base anyway. Good planning will be required by the player for good building placement, but this is a lot less cumbersome than initial base creation. Economy control consists of being able to control which factories are producing, how much the engineers are assisting, building priorities (if implemented), and controlling metal makers/other major resource drains. Here, the best solution I can recommend is one I've already mentioned, that being something like the Empire Tree from Sins of a Solar Empire. By putting all important structures on the planet in an easily organized display, activating the buildings, changing build orders, and jumping to buildings of concern would take a lot less searching. It also allows for monitoring on an extra-planetary scale, since each planet can be organized as a branch of the tree.

Then there's the question of base repair. When individual buildings get damaged or destroyed, the player should be able to get it all put back together with minimum effort. Fortunately, we already have the idea of self-repairing buildings, which will take care of repairs. For faster repairs, repair turrets would be handy on the front lines, and patrolling engineers would be handy as well, though since they are harder to keep track of, they shouldn't be the prevalent method. As for destroyed buildings, I maintain that the idea that was previously suggested, that of using the commands-as-entities structure would be sufficient to keep this process from becoming cumbersome. In other words, when a building is destroyed, its original build order is reactivated, so the building can be quickly replaced exactly as it was before. As was mentioned, the major issue with this is that if uncontrolled, this can become a major drain on the economy. However, a solution presents itself by using the Empire Tree. Destroyed buildings (or rather, their associated build commands) are displayed as "grayed out" items, and if the player wishes them to be built, a simple meta-click can activate that command for rebuilding (normal click zooms to the location, right click cancels the command permanently, etc). This gives the player ENORMOUS control over the maintenance of their base with regards to rebuilding.

The final matter revolves around rebuilding the base. This is reserved for the catastrophic, major attack, where many buildings have been destroyed. But I think the solution here is exactly the same as the solution I presented for minor rebuilding above (using entity-commands and an Empire Tree). In order to present some order to the rebuild process, shift clicking on the items would add the rebuild command in sequence, allowing for certain buildings to take priority and to minimize the economic drain to only a few buildings at a time.

In summary, with a proper UI, NONE of the facets of base maintenance should need any sort of AI assistance. Therefore on this level, AI shouldn't be necessary.

I'm out of time, so I'll post my thoughts on troop movements and AI later. Suffice it to say, my argument will be that on a lower level (unit micro), AI can be perfectly acceptable (though PA can be designed without needing it at all), but on any higher level (tactical and strategic), player control is absolutely necessary. Issues regarding multiple fronts/planets, super long term games, etc, once again boils down to UI. (No doubt I'll pimp the Empire Tree some more.)

That was a really well thought out post sylvesterink. I really like all these concepts, even outside of the whole AI debate it would present a great deal of management power to PA.

I think the idea of taking some UI concepts from SoaSE is a great in terms of managing a solar system. I would like to see the Empire tree get an overhaul however as it could be a bit dense and quite a task on its own to manage; regardless I think a PA implementation of such a system would be fantastic, particularly where base management and repair is concerned.

Definitely. Whenever I refer to the Empire Tree, I mean more as an abstract idea, based loosely on what was presented in SoaSE. However, I do like the amount of power their implementation presents to the player. It's almost a very abstract minimap.

Agreed. A well rounded post and a logical suggestion on how to make sure the player gets the best (and least frustrating) experience from the game possible.

As I already said, the misuse of the term AI will cause a knee-jerk reaction in many people reading the OP. Sylvesterink describes advanced 'UI tools' exactly how I imagine. I also concur with the initial statement:

*ahem* The original poster actually doesn't have 'AI' anywhere in his original post.

Sylvesterink, I think you've got a good idea there - essentially, be able to separate your planet into 'bases' (they do this in the video already), but then give a strong set of more abstract summary information to the player to show them what's in their base, and give them a higher level of control over the base. This could very easily include information like whether a factory is on repeat, what priority the factory has in the base, and whether or not the building has a automatic rebuild turned on.

I like it - it's the kind of evolution of the whole 'selected units' panel that PA needs to help manage 20,000 units. And the concept would probably scale fairly up to the planetary level as well, and down to individually selected units.

Really?

Out of context, maybe. But I couldn't resist.

If you read what I'm saying, I basically agree with your original premise anyway. I think the commander does need more powerful tools, units should be smarter and there should be more information.

The problem is that I don't want the game to play itself and I don't want to be bombarded with so many bar graphs, numbers and icons that I can't see the wood for the trees.

This is such a different and creative idea! This is why I pledged so much to the cause!

As a turtle I would love to build beautifully symmetric bases and have an AI defend/attack for me. Although it is quite possible that I would get fed up with the AI just throwing my units into the grinder. As always if you can add it to the game why not give the players the chance to make that choice themselves? If you get a working prototype let us try it in alpha or beta?

Dammit wolfdogg.. *edits first post*...

lol.

This is getting a whole section in the UI requests list.
It's not just good for the player, but the casters as well.

How often have you seen them get "surprised" but an experimental or nuke they didn't spot that has already been built?

The president = Commander in Chief.
The Joint Chiefs answer to him.

He's probably in charge of the largest military in all of human history.
In manpower, logistics, infrastructure and most definitely killing power.

The biggest problem Supreme Commander ,was been made from future.
Again i say sry ,cant stay quiet and see players that not now FA ,and not now the quality off that game,talk without Know.
Iff i have to choose a perfect Rts game i choose FA(F)orever