The Shrine

I've been reading an article on www.sirlin.net and it appears to me some guy was giving a lecture in UC Berkeley on the strategy of Starcraft. Although it reads like the generic 'the guy who won must be doing something deep because he won', I do like some of the concepts they discussed even though I'm sure the guy playing the game weren't even thinking about it. Now as a disclaimer I don't ever think of Starcraft as a strategy game. Not when a game whose skill is determined by your APM (Action Per Minute) rating. But a lot of the concepts that I think Starcraft got utterly wrong are interesting in the context of a real strategy game.

I was going to go over more than one but I realize just one wouold take a long time. So let's start with scouting/Fog of War. Fog of War has existed since the beginning of time for RTS games. It seems like it's a good idea. It's certainly a correct gaming concept if the goal is to have a game that can be finished with an hour, but it's a poor resemblance of reality. The moment any one of your unit sees one of the enemy units, every unit on your side magically knows where those enemy units are. If you have this kind of capability you already have some kind of global communication system anyway. To be an accurate simulation of war (which is what strategy games usually are), the unit seeing the enemy should only be able to relay this information if it successfully got back to your main base or some communication relay structure.

Obviously this doesn't work probably due to a limitation of gameplay when it comes to RTS games. It'd be really tedious if you have to send out a guy as 'scout' and he may or may not make it back and you got no idea what's going on, or if he's ever going to make it back (he could be killed). So what does a good scouting/fog of war look like? In ROTK 11, you can always see every unit on the map and the stat of every city. You can also see the location of every general, but NOT what they're doing. If it says Zhuge Liang is 'occupied' it could mean he was out this turn looking for money, or that he is currently participating in a 10 turn project and won't be able to command anything for the next 10 turns.

Since the generals are essentially the units, this means you know the location of every unit in the game, but not what they're currently doing. If I take my five best generals and move them all to city A for a surprise attack, to my opponent he will see my five best general in various city with the status 'unavailable' (because they're currently traveling to city A). Once the generals arrive he will see them in the new city, but by then it might be too late if he's not prepared.

In People's General, all units have a profile rating that determines how easy it is to see them. There is still a fog of war but scouting units have a huge scouting radius. Effectively this means unless you have no unit anywhere near an enemy cluster, you will see stuff like M3A1 battle tank or artillery from a mile away, well outside of their ability to hit you. But you might not be able to see a Stealth Tank because they've a very low profile rating. Sometimes you can't see them at all, and sometimes they'll just show up as ??? on your map. That is, you know there's a unit there but you can't be sure what it is. If you know the enemy has Stealth Tanks it's probably a Stealth Tank, but you can't be sure. It could be say a Light Infantry (those are very hard to see too), so the ??? unit can either be the strongest unit in the game or just cannon fodder. This means maybe your scouting unit now has to get close enough where he will get hit, but presumably finding out that ??? is a Stealth Tank is worth the risk. Of course the ??? could just be a Light Infantry too and then your recon unit has just died for nothing.

To make things even more interesting, you can even fire on the ??? units without knowing what they are. If your artillery fired on a Stealth Tank you probably ended up doing 0 damage since they got the best armor in the game too. If it's indeed a light infantry you'll inflict plenty of damage on them. But you can't see if you actually did any damage to a ??? unit, and even the weakest unit takes 2 or 3 shots to kill, so you really have no idea what you're firing at until you fired about 3-4 times.

Now you might ask why not just move a unit next to the ??? unit. After all you can discover the identity of any unit if you're next to it. The problem is that if that unit turns out to be a Stealth Tank then whatever you just moved next to it is now dead if it has to fight a Stealth Tank at a point blank range.

I should note that both People's General and ROTK 11 are turn-based game. In both game you have nearly perfect information and yet it's the last piece of the perfect information that you don't have that messes you up. In People's General you will not be able to easily tell what those last 2 ??? units are. In ROTK 11 you'll never be able to tell what your enemy generals are currently doing. You know they are either avaiable or they're out for 1 to a rather large number of turns, but you can't tell the difference beween a general that's out for 1 turn or 10 turns, or why they're out for that period of time.

While everyone can see you have 330K troops defending Louyang , the biggest city in the game, (a decent total army size is generally less than a million even in a developed game), the other guy can't easily tell if you're just putting troops there to scare potential invasion commanded by your second string leaders, or if a bunch of top tier generals are about to arrive to launch a major attack. And sometimes you got to attack with no chance to win just to force the enemy to commit his resources here, instead of just scaring you away with large number of troops (which are relatively cheap) without a good leader (which are the real resources in this game).

I burned out on ROTK after I finally beat IV (?) and haven't looked back since. Used to love that series though. (I guess given that they're up to 11 it may have been awhile...)

Well most of the ROTKs have flaws. 11 got most of it right, but it's also time consuming. In the game I play against myself it's up to like the 1000th hour now and there's still no way to break a well defended position without relying on utter cheap tactics.

One thing I noticed is that stalemates are inevitable in a good strategy game. If you can't have a stalemate, then that means one side has a strategy that will always win at some point. In ROTK 11, stalemates are the norm, and the strategy is figuring out how to weaken your enemy by a little bit even in a stalemate. For example sieging Louyang is one of the toughest thing in the game since it's a huge city defended by gates in an easily fortifyable position and it can hold up to 330K troops while the max attacking force is only 200K if you sent out a max army attack force from 2 neighboring cities. So against an equally skilled opponent there is no chance you'll ever take the city in one shot. But if you can say secure the first gate with 100K troops, you might be able to prevent the enemy from using the superior geography to nail you next time. Perhaps you can try to cut his army off from his infrastructure, and then you'll be able to damage their economy so that they can't just continously pour out 330K troops at you. Then again, you can fail and you'll be on the receiving end of some truly viscious counterattacks. It's a calculated risk you got to take to try to take probably the best city in the game (it produces enough resources to be worth about 4 cities by itself).

Don wrote:Now as a disclaimer I don't ever think of Starcraft as a strategy game. Not when a game whose skill is determined by your APM (Action Per Minute) rating.

This is why I like to play StarCraft on slow. Sure, people can call me a pussy all you want, but at that point, it's more of a game of mental skill, instead of a game of spazzy mouse clicking.

Well there is a sheer gameplay element to the pace Starcraft plays at. If it didn't play that fast, then your average game might take an hour instead of 20 minutes, and a long game will be 3 hours instead of an hour. That's a lot of time to be playing and it's got to get pretty boring.

Of course, just because a game is fast paced or not, doesn't mean it can't minimize the advantages of clicking very fast.

I want a cheeseburger.

SineSwiper wrote:

Don wrote:Now as a disclaimer I don't ever think of Starcraft as a strategy game. Not when a game whose skill is determined by your APM (Action Per Minute) rating.

This is why I like to play StarCraft on slow. Sure, people can call me a pussy all you want, but at that point, it's more of a game of mental skill, instead of a game of spazzy mouse clicking.

The skill is always there with Blizzard RTS'es, it's just that you need skill AND spazzy clicking in order to succeed at normal speed levels or higher.

Don wrote:Well there is a sheer gameplay element to the pace Starcraft plays at. If it didn't play that fast, then your average game might take an hour instead of 20 minutes, and a long game will be 3 hours instead of an hour. That's a lot of time to be playing and it's got to get pretty boring.

Of course, just because a game is fast paced or not, doesn't mean it can't minimize the advantages of clicking very fast.

That's not how it ends up, though. Granted, the games are longer, but after you get past the first 30 minutes (which are kinda boring), you realize that you can micromanage a helluva lot better. You have the time to think about your actions, get stuff built, get armies where they need to go, scout the enemy, etc. in the same amount of time it would take to do ONE of those things playing on fast. In the end, the games are actually shorter than you think because all of that micromanagement gets you the better techs and armies at a faster rate.

Admittedly, I'd only got away with playing a few games this way, one of which was a tOWS game several years ago. (Four player game with Eric, Manshoon, I think, and two others.) It's hard to get people to play these kind of games.