Subscribe

Get updates via e-mail:

Thursday
Oct232008

5) Divide and Conquer

Introducing Zileas

Although I grew up in the realm of fighting games, my community is a mirror image of nearly every gaming community. The real-time strategy game community even has a bizzaro-world version of me called Zileas (some call him Tom Cadwell). In an odd footnote of history, I was a senior at MIT when Zileas was a freshman, but I never actually met him there. We each knew the other through reputation in our respective game genres, and eventually crossed paths in the incestuous game industry.

Zileas wrote a great deal about how to divide and conquer the enemy and to concentrate firepower—and so did Sun Tzu.

Zileas was talking about StarCraft and Sun Tzu was talking actual war, but since real-time strategy games are (arguably) simulations of actual war, it’s not surprising that great minds thought alike here. What’s interesting is that while Sun Tzu wrote mostly about the large, macro scale, Zileas wrote about the very same concepts on the small, micro scale. Ironically, Zileas made his fame in the StarCraft world by developing and writing about his “new school” approach to the game where he focused on dividing and conquering and concentrating firepower on the micro level, rather than the “old school” approach of concentrating on the macro level. In case you missed the irony, it’s that the old school that Zileas argued against was just another interpretation of the very same concepts his own school was based on, all straight from The Art of War.

On the most zoomed out level, Sun Tzu tells us when to attack, based on the sheer size of the armies involved:

It is the rule in war: If our forces are ten to the enemy’s one, to surround him; if five to one, to attack him; if twice as numerous, to divide our army into two, one to meet the enemy in front, and one to fall upon his rear; if he replies to the frontal attack, he may be crushed from behind; if to the rearward attack, he may be crushed in front.

If equally matched, we can offer battle; if slightly inferior in numbers, we can avoid the enemy; if quite unequal in every way, we can flee from him. Though an obstinate fight may be made by a small force, in the end it must be captured by the larger force.

On the most zoomed out level, Zileas tells us when (in StarCraft) we are losing:

If your kill ratio multiplied by the ratio of your production to their production is less than 1, you are losing. If their economy is gaining speed, and yours is stationary and this number is close to but over 1, you are still probably losing. When I say kill ratio I do not mean units killed/units lost; I mean resources killed/resources lost both in terms of unit production, miscellaneous upkeep costs (scarabs) and building production/loss.

 

That last Zileas quote is pretty mystifying, but if you read it about twenty times, you might find some deep StarCraft insights!

Mismatched Forces

One of Sun Tzu’s main points is to attack an inferior force with a superior one. Even if both armies are of the same size and power, this can easily be done by looking at smaller pieces of the whole. If the enemy only defends one piece of his empire—and we know this—then the rest of his empire is wide open. We can send but a fraction of our troops to dismantle any number of his undefended spots. The more spots he defends, the weaker each spot becomes. If he defends all ten of his outposts equally and we concentrate the attack of but half our army at one spot, we outnumber him five to one! We have concentrated our firepower, while the enemy’s has been divided and weakened.

All of this rests upon the shoulders of secrecy and reconnaissance. Without these, Sun Tzu’s method of divide and conquer would not be possible.

The spot where we intend to fight must not be made known, for then the enemy will have to prepare against a possible attack at several different points; and his forces being thus distributed in many directions, the numbers we shall have to face at any given point will be proportionately few.

Numerical weakness comes from having to prepare against possible attacks; numerical strength from compelling our adversary to make these preparations against us. Knowing the place and time of the coming battle, we may concentrate from the greatest distances in order to fight. But if neither time nor place be known, then the left wing will be impotent to succor the right, the right equally impotent to succor the left, the van unable to relieve the rear, or the rear to support the van.

So Sun Tzu tells us to keep our own positions and intentions secret. He tells us to discover the positions and intentions of the enemy. Through this we can concentrate our firepower on the enemy’s weakest points, even at the expense of our own defense; if our weak points are secret from the enemy, he will not know where to attack and he will likely end up dividing his own forces. Our divided enemy thus conquers himself as he cannot hope to defend against our entire concentrated army with just a fraction of his own.

StarCraft: The “Old School”

Sun Tzu’s ways are the ways of the best StarCraft players in what Zileas calls the “old school.” These players strive to build a strong economy to finance overwhelming hordes of units. When they outnumber the enemy ten to one, they surround; five to one, they attack, you get the idea. Individual battles matter little to these players, since it’s more important to build a large mobile force capable of attacking the opponent’s weak spots.

Most of these players come from the days of Warcraft 2, StarCraft’s predecessor. Warcraft’s interface and units didn’t allow players to gain much benefit from micromanaging individual battles. Warcraft’s units were more homogeneous, meaning you didn’t see kill ratios of 50:1 like Templars and Reavers are capable of in StarCraft. In short, macromanagement was the only way to go. Build a large army. Divide the enemy’s army. Concentrate the firepower of your army.

StarCraft: The “New School”

And then there was Zileas. He came along and pointed out the amazing effects micromanagement of individual battles can have in StarCraft, and he preached the revolutionary ideas of divide and conquer and concentration of firepower—on the small scale, that is.

Lesson 1: Shift queue to concentrate firepower. When enemy forces engage, say ten marines versus ten marines, they will fire at each other in a mostly random distribution, so units will only start dying toward the end of the battle. The better player will select all his marines and concentrate their firepower on a single enemy marine, then (hold the shift key to) queue the next command to concentrate firepower on the second enemy marine, and so forth. All ten of the first player’s marines will kill one of the enemy’s units right away, reducing his firepower. The ten marines will then automatically (through shift queuing) concentrate their fire on the next enemy unit, then the next one, and so on. The enemy is dividing his own fire but the better player concentrates it. If you use this technique but your opponent doesn’t, you’ll probably end the fight with four marines left when he is down to none.

Lesson 2: Use formation to concentrate firepower. When two enemy forces engage, say ten marines versus ten marines, formation can be everything. If one player marches his single file line of marines into a horizontal line of enemy marines, the horizontal line formation will be able to concentrate its fire on the first marine in the single file line, then the second, and so on. The last marines in the single file line won’t even be close enough to fire until all their friends are dead. Even better than a horizontal line is “shallow encirclement,” a crescent-shaped formation that maximizes the firepower one can apply to a point.

Lesson 3: Use choke points (narrow passes) to divide the enemy’s units. When a large enemy force must pass through a narrow choke point (either naturally created by terrain or artificially created by your buildings) he is dividing his own force for you. You can concentrate your firepower on each unit as it passes by.

There are more lessons, but his point is to focus on the concentration of firepower on the small scale of an individual battle. I cannot leave out Zileas’s most extreme and signature use of concentration of firepower: his “Doom Drop.”

Zileas is known for playing the Protoss race, the race smallest in numbers and most powerful in punch. Notice that they are already concentrated before he even got a hold of them. A so-called Doom Drop is when you fill about four shuttles (flying transports that carry other units) with amazingly powerful Protoss attack units such as Reavers, Templars, and Archons. (Heavily armored air units (Scouts) must sometimes accompany the shuttles.) This superabundance of force—this concentration of firepower—is enough to overwhelm nearly anything so long as it is applied instantly at a single point. When one Archon, three Reavers, four Zealots, and three Templars suddenly appear in the middle of your base, the sheer force of it all applied to your surely badly positioned units is usually too much.

Even more devastating is what Zileas calls his “Extra Crispy with Slaw” version of the Doom Drop, where he uses hallucinated (illusionary) units to draw fire. Flying four shuttles into an enemy base is not an easy task, because they’ll probably be shot down by whatever anti-air happens to be scattered about. Four Shuttles accompanied by, say, five Scouts is another matter. Now the anti-air fire has been divided among more targets. Better still if all these targets are accompanied by, say, ten illusionary Scouts. The illusions can’t attack, but they draw enemy fire giving the real units more time to act. In effect, the illusions divide and conquer the enemy’s anti-air fire. Deception at its best.

Micro and Macro

Why not apply Sun Tzu’s teachings of divide and conquer and concentration of firepower on the large scale as well as the small? Must one choose one over the other? The answer in StarCraft, realistically, is yes. One only has so much attention that must be divided between micro Extra Crispy with Slaw Doom Drops and macro economy horde-building.

The Third Resource: Concentration

Zileas explains:

Minerals and Gas are the resources that most players think in terms of. Although these are central to the game, you also need to think in terms of concentration. I define concentration as the time that a player has to spend focusing on a task during the game. Expanding is a high concentration task, especially if you are Protoss. Attacking certainly has a high concentration level, and the more concentration you put into an attack, the higher the effect. Even scouting carries a high associated cost. One big difference between ‘someone who is really good’ and someone who is #1 is knowing when you need to watch a battle and when you don’t, and recognizing that your opponent also has a finite amount of concentration to draw from. There are a number of techniques for minimizing concentration costs (i.e. hot-keying buildings, using magic spell hot keys, queuing attacks, etc.), but everything you do has some intangible concentration cost. I would argue that as you get better at StarCraft, you go into a match with a larger innate concentration income/second. It is very possible when doing multiple coordinated attacks at different locations to use your superior concentration reserve (if you have it) to decimate an enemy who is tied with you in terms of unit control and tangible resources. Although I’m sorry to say this, concentration is basically talent. Playing a lot of games slowly raises it, but it’s something some people have a lot of and some people don’t. It’s kinda like fast sprint ability in running: you can train up and become a great long distance runner, but for sprinting, there’s always that talent based barrier—you can slowly improve it, but everyone has a limit. I’m sure that someone will push me off #1 who has more innate talent, along with the same skills.

The best way to train concentration is to play 2 on 1s and 3 on 1s (multiple opponents vs. you). I can often pull 3 on 1s, and certainly 2 on 1s, and really the only reason I can do this is my ability to multitask. Also, team melee is an interesting game as it involves doubled concentration reserves on both sides . . . well almost doubled since its not one mind thinking at once and they have to communicate.

Whether you, as a player, spend your concentration resources on the large scale or the small depends on which game is at hand and your personal style. In either case, the same principles are at work. On one level or another, thou shalt concentrate thy fire and divide and conquer thine enemy!