Ah, the life of an engineering student means making a post means taking time from other stuff...
anyways:
Vendetta wrote:Because, of course, this would be an option that the human player has set, in accordance with the strategy he has devised, and the AI would then follow that strategy without micromanagement, giving the impression that it is in fact an intelligent troop under player control.
I have no problem with player set behaviours. That said, it is still micromanagement in the same way that commanding peons are micromanagement. The only difference is that it is spread out over time as opposed to split second super inputs.
As for goliath targeting, it is critical for them to target the highest threat first (basic micro) and that depends on the tactical sistuation. Spreading out fire make things worst.
Darth Wong wrote:The style of RTS that PIGWANG favours is completely dominated by predictability and routines; nothing genuinely surprising can ever happen because it's a simple matter of following the same build order every time, doing the same basic things every time, and competing solely on the basis of how often you've practiced this sequence and how quick you are at executing it.
The problem with predictable "memorize build order = strategy" is due to something completely different from micromanagement and needs to be solved independently.
One problem is that developers can not balance an asymmetrical game without knowing almost all viable strategies before hand. If players come up with new and original and useful strategies it usually result in a broken game with unbalanced sides as it was not accounted for when all the unit statistics and costs are determined. If there is an original vision behind the game, it usually means it gets patched out of existence, and even if it does not, it simply become nerfed to balance and added to a finite pool of known strats.
The reason why strategy is partly memorized as opposed to created on the fly is also due to the relative processing power of the player base. An proven strategy has been scruntined by a huge part of the player community with all its kinks and flaws worked out. Even if the RTS community is not all that clever, evolution and mutation usually produce far better strats in the long run than even the most clever individual. Any clever trick an individual can come up with would be absorbed by the community quickly if there is merit. As an result, being relatively unoriginal is fine as long as they are at near the head of the curve as long as learning new strats is concerned, which takes as much as watching a few replays.
The other reason why strategy is memorized is simply because of the very predictable and consistant environment RTS games are played. Maps and units are all fixed and played for millions of games. This exhaustive search means most viable strats are discovered. The RTS community does not seem to like randomness much (especially when it unbalance things), unlike TBS gamers and nor do they like "meta-game" (unscoutable chance strategy) wins.
Finally, the game duration is not conductive to original thought as the average 20 minute session is not long enough for any serious computation. Most strategy ideas have to be developed before the game even starts as there is simply not that much time to hesitate. This is true even if micro is removed, simply due to the short duration of those sessions. Generally, there is alot of thought in RTS games. It just happens before games and is conditioned into the head.
The reason why TBS or table top gaming involves more originality is because start conditions is heavily randomized, greatly limiting pre-game planning at the cost of letting the worst player win when lucky. The reason why games like chess or go requires tons of thinking is because simple rote memory is insufficient to memorize everything. (however those games are completely unplayable without good memory)
Micromanagement, strangely enough, is a random variable that makes the game far less predictable as it is defined only by the other player as opposed to fixed unit stats. That one of the things where mere memorization is insufficient and flexibility required.
Fire Fly wrote:Judging by the latest game reviews the trend for RTS games is indeed for more unit intelligence, less micromanagement, a level of immersion, bigger battles, and a game that demands strategic thinking.
What needs to change even more is to break the mould of attrition warfare and the so called "Three Xs" and to include more elements of real modern warfare
Nice words, now back it up with analysis.
I agree that games are leaning towards immersion and unit "intelligence." Scale is more of an individual game design issue, however I do not see a reduction of micromangament or increase in strategic thinking.
Lets take supreme commander. Do you know what happens when you get a monkeylord behind another? Instant rape. You can bet that players will all struggle to micro to do this to the opponent. As for strategic thinking, what it probably means is massing tier 3 gunships.
I'd give it a few month before game strats are "solidified" for popular maps.
Just by tossing more units in, the age old problems have not died.
As for elements of "real warfare", I wonder what people are talking about. Attrition is an very real and very effective "strategy" in real warfare. It won as many wars as the fanciest strategy. (the other is close to being zomg tank rush kekekekekeke ^___^) If there is one thing that have not been modelled to satisfication in any game, it'd be logistics. That'd be the most annoying management issue ever, however.
Stark wrote:It is common in RTSs that an attack (of planes, say) will be attacked on detection by defences (SAMs and AA, say). However, there is no intelligence or coordination between defences - they simply target the closest unit and fire. Thus, you get many missiles/bullets flying at the same couple of front units, and the rear units get through unscathed. What PIGWANG doesn't understand is that the sort of autonomous AI we're discussing would allow the defences in this example to distribute their fire efficiently and spread it around all targets, instead of firing eight missiles at one target and none at others. I'd even like to see slow-loading weapons like SAMs hold fire entirely if all targets have missiles on their way, to better respond to other attacks.
If you are talking about standard RTS where air units have hundreds of hit points and take dozens of hits to take down, it make perfect sense. Distributing fire means you don't kill the enemy fast enough as every damaged enemy unit retains its full firepower. What you want is focus fire like in ground fire. There is occasionally wasted shots with super overkill if the projectile takes a long time to register a hit, but those are rare when air units are so tough.
If you are talking about RTS where air units are extremely fragile and anti-air is area-attack, than there is a need for hold fire and spread fire commands. Most games do not have this mechanic and does not need it.
However, if you want autonomous (as opposed to player set behaviour) AI to do this, you'll be meeting with a group of cursing gamers at the horribleness of the game AI when their SAM refuse to fire at those bombers that have just dropped bombs when there is no better target in range, or those fixed anti-aircraft guns that decided to spread fire at cheap intercepter as oppose to the nuclear bombing strategic bomber, especially when the bomber has been targeted once by the player already, but flew out of range of the turrets in its first run. (thus breaking "target lock" which usually means a reset of unit behaviour)
)
But that is one of the reasons I prefer TBS games to RTS, since usually they don't involve base management and the game it doesn't boil down to "I know fastest build time get ready for Zergling Rush kekekekeke!"