I watched you and Nuetrino discuss the AI. I know you will be doing a lot of cool things. See what you can do to make the Commander much smarter. I really think a specific neural network should be implemented for the commander if it will make the game more fun. I think this is really important, because unlike in Star Craft or Dawn of War or any other type of RTS Annihilation games have had a very special role for the commander as a super weapon as well as your hearth stone (heart, etc) which must be defended to prevent a defeat. However, we still haven't had an AI with a commander intelligence that looks smart in any of the Annihilation games. Example: Commanders don't dodge projectiles and they don't have fancy attack moves. They should even if they aren't that good at it. When commanders are low on health they should retreat and troops should move the intercept attacking units and other units that are close to be able to attack the commander. I think you will want to work with the rest of the Uber team to make sure the commander has options (other units, upgrades, high energy cost abilities and facilities) that make providing AI for the commander late game important to having a fulfilling skirmish experience that doesn't feel under played because the AI doesn't fully recognize the importance of the Commander, how to use it or how to protect it. Love the pod casts. Bast
None of this is a good idea. The function of the commander is to do three things. 1) To kick the game off by seeding you with resources and build capability. 2) To provide a concrete end state to the game when killed. And 3) To prevent early rushing by having both sides begin the game with a powerful fighter relative to super-fast rushes. However in the mid and late game, the commander should be outclassed by things it indirectly produced. You need to protect the commander. It is like the King in chess- a somewhat powerful piece in its own right, but one which must be protected at all costs. Consequently the King cannot be powerful enough to be worth using as a primary combat asset. The commander is not a superweapon, except that it's pretty amazing at building a giant army, or building actual superweapons, in a big hurry when it shows up on a planet. All the AI capabilities you suggest, such as dodging projectiles, retreating, etc. are either a function of the player giving orders, or are unnecessary to begin with. The commander retreats when you order it, and not before. Expecting an AI to make that decision for you is wrong. Additionally, bullets and lasers are fast. Big objects are generally much slower, including commanders. So either you can't really dodge them anyway, or if they are slow and dodgeable guns, then player micro should govern instead of an automated shot-dodging algorithm- otherwise why make them slow? Now, regarding your "special moves" idea. Were you thinking karate chop? The canonical special weapon of the commander is the D-Gun. And the reason why this weapon suits the Commander is that it kills any target, regardless of individual strength. This means a super-fast rush to T2 and showing up with just one powerful unit is very ineffective against the commander. A swarm of smaller units will perform much better. Additionally, the Dgun has limited capability to splash onto multiple targets, meaning skilled and judicious use of it can get you out of a tight spot. However due to the high value of the commander, it is seldom worthwhile to rely on the Dgun. The commander is the ultimate anti-cheese mechanic. You can rush to harass and disrupt economy and production, but a super-early rush will not actually kill the other player and end the game right there. Secondly, the commander needs to be weak compared to mid and late game weaponry, and must be protected. It can have a Dgun, but it does not need shot-dodging AI or "fancy attack moves." A gun and a nanolathe, standard issue, please.
It already kind of is isn't it? D gun is a lot more potent than most of what you'd consider a King able to do (if it gets one). But, early game, do we want people to have full on Queen pieces though? In Chess you have to deliberately invest a number of moves in order to even activate your Queen, and even then its generally considered less than advantageous outside of those novelty strategies. So I think the answer is, no, no Queen. Invest in your queen, be happy with your provided potent king.
Consider the design of an upgrade to the commander of this kind. Sure, it sounds sort of cool to increase the strength of a unit like the commander. However, what you are really doing is increasing the strength of the commander relative to other units. It takes more hardware to bring down the commander, and thus win the game. It has to be possible for the other player to make units in order to kill this upgraded commander, otherwise the game is going to last literally forever. So, your unit design system must contain units that are more powerful than the UPGRADED commander, which can end the game. So now we have to have a T3. And if there are more commander upgrades.... and so on. It's just like in an RPG where the player gains levels, and the game has to make the monsters tougher just to keep up. Only the other player has to spend resources on the monsters, and more expensive monsters means fewer of them. Overall, it's simpler and better to just have non-scaling commanders and avoid this endless upward cost progression of units, which only results in delay and fewer pieces on the board. This also introduces the dangerous possibility of a player opting to prioritize this power increase for offensive purposes. Especially early in the game when there will be few other assets. In commander-versus-commander battles, the upgraded one will virtually always win. This is also extremely boring, stat-based, deterministic combat. Rather than increasing the HP or damage of the actual commander, it is better game design to construct more units, more base facilities, more bases, and gain more control over the map around your commander. This makes your commander more difficult to kill in a much more interesting way than just doubling its HP or what have you. In my opinion SupCom ACU's are far too tough. They can take ridiculous amounts of punishment, and regenerate insanely quickly. SupCom ACU's are so tough that they just take FOREVER to kill, even if you make tremendous mistakes, just by sheer durability your ACU might take so long for the enemy to kill that reinforcements arrive, or maybe the ACU just pewpews his way out himself. And then T3 bombers kill him in three hits anyway, because that's the kind of power differential needed to actually end the game, and that possibility creates tension. TA commanders are about right, perhaps a bit on the squishy side. You can use your commander as a weapon, but it's risky, and you should bring as large an escort as you can. If the enemy has any nasty surprises in store, he could spring them on you and win the game if you're exposing your commander like this. Rather than upgrading your commander's HP and damage, you build decoy commanders essentially giving your real commander more "lives" as your opponent might commit a lot to try and assassinate a decoy.
With PA being more focused on the smaller units then experimental's , I figured that the game could use at least 1 queen unit.
This. If we aren't getting experimentals as per Supcom/TA then we should at least be able to make our Commander into a rightous badass of robotic destruction...
Ihmo, a player should be able to use the commander offensively as a risky mean to stop a slippery slope and to put him back in the game. Commander vs tech1 units in TA was a good balance. ( perhaps a bit too strong ). But against T2 he was really to weak to be interesting anymore.
You forgot 4) To seed new army growth across the galaxy Having early game expansion dependent on your Commander would make it the most strategically important unit in the game. Edit: There has been a lot of discussion on the function, abilities, and roles of Commanders. Reading through "ACU snipes" is a good idea. viewtopic.php?f=61&t=41336&p=618490
@ ledarsi Bastilean is specifically referring to AI controlled players, not an AI assisted human player. The AI in past games have controlled the ACU in a rather unweildy fashion that would be nice if it were improved. It was easy to bait the unit out if you knew what to exploit.
ALL HAIL THE MACHINE GOD!!!111 (Sorry, couldn't resist). It was more or less answered within livesteam already - "- What about ACU? - Such differences should be probably hardcoded". So, they seems to be are aware of the problem.
I play TA, not SupCom, so have no input on their different actions. However, there have been many games (Of TA) where my attention has been elsewhere on a battlefield, and when (IF!!) I get a warning that my commander is under attack, occasionally I have only just selected him as he goes BOOM! I was wondering if it was possible or fair to introduce a different in-built standard response to being attacked (assuming the Commander isn't under direct player control at the time), so that instead of just standing there and firing at the units attacking him (Or her!), he automatically retreats away from the attacking forces for a few seconds (still firing at enemy troops, of course), giving a player at least a few seconds in which to try and salvage the game. If this is deemed unfair, then can the game be made slightly more reliable than TA when it comes to getting an audible warning of a commander being attacked?
Madcook, the issue in TA is that the Commander Under Attack warning didn't go off until your commander had lost a decent amount of health, and by then it tended to be too late. An earlier warning would probably be a good solution. Also, features like ZK's retreat zones might help out as well.
