A game of Starcraft is intense, back and forth war of keystrokes and heartbeats. Micro and macro go hand in hand, and the faster you play, the more you’re capable of achieving. There are dozens of combat types to select and potentially hundreds of builds and strategies to choose from. On the surface, it seems to have nothing in common with board games. Look a little deeper, though, and the differences quickly start to disappear.

What’s Go?

Go, also known as Igo and Baduk, is a turn-based strategy game with exactly one move: play a stone somewhere on the board. Your opponent does the same on their turn, and you alternate until the board is filled with stones. The points come from the simple capture system. Fill in all of the spaces around your opponent’s stones, and you capture them. Players posture to capture the most stones and secure the most territory – spaces where opponents can’t play without getting captured – and the game continues until no more moves are available. The player with the most stones and territory left on the board wins.

(a completed game)

Go is considered a beautiful game in the same way that Starcraft is considered a beautiful game. (Both games are popular in Korea as well; coincidence?) Both games are endlessly complex – so much so that Google had to build an AI to simulate the way that humans think about the game in order to win with a computer, whereas very deep games like Chess were conquered almost a decade ago through sheer brute force. Both games are about fighting over space in an enclosed area. And both games have something very special in common: each game is a negotiation.

The principle of negotiation is apparent in both games. In Starcraft, expansion first is an extremely common build for all races. However, expansion first is almost never right against a player that always makes aggressive structures and attacks immediately. But hyper-aggression often loses to simply making a defensive wall and a few guys, it’s very uncommon. This means that the greedy player is allowed to take the advantage and expand quickly, securing more game resources.

Now, say you’re the player who builds one unit-producing structure and then expands. You’re safe against a rush, but behind against the players that expand first! So you modify your build to include a scout, which gives information on fast expanders, and create a followup plan that allows you to pressure and attack the weaker defenses of those players, say by building a tank before expanding and making a small push.



This is the process of negotiation. Defensive players weed out the aggressors, and greedy players take an advantage against Defenders. Defensive players learn about this plan and tax the weaknesses of that advantage. So greedy players in turn look for ways to exploit that tax, say by making flying units to pick off our Terran player’s tanks. In a well-balanced game, aggression and defense are adjustable in minute quantities. Players are constantly asking the question: What can I get out of this?

This same process of negotiation is most apparent in simple, static board games with many viable styles. In Go, professionals play to eek out tiny corner territories or make winding snake-like constructions. However, your opponent can read your intent, and if they really want to, they can almost always stop you from executing your plan. Because Go is a game of exactly equal resources – the stones – negotiations often occur. The value of each stone is entirely in the judgement of both players.

Because your opponent can easily block your play, strategies quickly develop into neat-looking equal exchanges. Similarly to openings in Chess, Go has joseki, or series where both players mutually attack and defend to an even-ish result, to begin the game. These are memorized, and players with the initiative on the board can choose joseki and make adjustments based on the nuances of its possible results. Sound like build orders in Starcraft? You bet it does. Starcraft players scout early, and both players know how the opponent might develop. They choose their style based on that continuous review. When your opponent plays with Mutalisks or Tanks, it limits your possible reactions. Your opponent must then play within the limitations they’ve set, but you still have many ways to play. As the game develops, expansions are taken, income increases, and the possibilities increase as the players fight over important territory and resources.

Go players will recognize a lot of the ideas and language above. They will also recognize that if one player doesn’t understand the points of negotiation, they will fall apart and die.

Building Blocks of Negotiation: Decision Points, Arguments, Exchanges

So we’ve established that any strategy game comes down to negotiating with your opponent. The game proceeds along the path that both players have agreed upon. (If one player can dictate the entire path of the game to his favor no matter what, the game is imbalanced.) How can you use this to understand what direction you should take?

Games of Go exemplifies of how negotiation proceeds in a strategy game. Players tend to start the game selfishly. They establish basic foundations, called moyo, and execute a few joseki. The game seems tranquil. Peaceful play goes on until one player decides the direction of the game needs to change. It can be as early as move five or six or as late as move 50, but one player will slam a stone in the opponent’s supposed territory.

An attack in Go is a play that threatens to take away the opponents territory or capture their stones. The attacker says, “This looks like yours, but I think I will take it instead.” The defender now has a choice. They can say, “No, you can’t have it.” If they choose that, they attack the attacker and try to destroy their invasion. Or, they can say “That’s fine, but I’ll take this instead.” (The third option, “That’s fine, just take all my stuff” is rarely played…)



This is a decision point. The attacker asks for more than he’s entitled to. The defender can either start an argument or make an exchange. The argument comes down to each player’s relative strength.

If the Defender decides to exchange, they can play across the board or play stones near and around the fight for influence. The key to the exchange is that it’s minimally interactive and seeking a resolution.

This concept shows up in almost every strategy game. Your opponent is making some Marines. You’re Zerg, and you have a fresh Lair. If you make Lurkers, you’ll be able to defend all of the Marines. Your opponent will be able to hold off your Lurkers until they get tanks and science vessels, so you’ve essentially agreed to play a longer game. If you decide to defend with sunken colonies and make a spire, you’ll be able to attack their base, but won’t be able to contest the Marines very well and will lose some stuff. This game will play out in a different way.

That’s the essence of a decision point. It’s a point in the game where a player, presumably playing at the top of their game, may choose between uncertain futures. Decision points are worthy of study.

In Go, there are dozens of decision points each game – perhaps moreso than any other strategy game. It’s why computers have such a hard time with Go. Computers lack intuition and style. You cannot decide between roughly equal futures without these critical components.

Arguments occur when two players are trying to shape the game into different directions. An aggressor sees his opponent’s defenses and decides that he can do enough damage to make attacking a good thing. For instance, in Starcraft, when you show up at your opponent’s base with a bunch of Dragoons and get dinged by a tank on the high ground, you can usually decide to just go up the ramp and start lighting stuff up anyway. This is an argument. You’re saying that the opponent’s tank is not going to do enough to deter your dragoons. You can kill marines and supply depots and SCVs until you run out of plasma balls to throw.

Another common argument is Mutalisks against turrets. Mutas in groups can easily take down turrets with good micro. However, turrets are cheap and mutas are expensive. Zerg argues that the Mutas can do more damage than the turrets can prevent. Terran argues that the turrets are cheap and will do enough to hold them off until Marines arrive.

The winner of an argument is the player that takes an advantage. Arguments are worthy of study because the player with many small advantages tends to win the game.

Exchanges are also simple to understand. In Go, a complex fight can often be settled by losing a few stones to take a few stones. A similar example is a zealot drop. I drop four zealots, costing 400 minerals, and tax your APM and kill your drones. My zealots die. You get this, I get that. Easy. Exchanges are the opposite of arguments. Exchanges are agreements.

Interestingly, weaker players often won’t go for exchanges. Sometimes they don’t see them as possible, but often they don’t want to give up their precious stuff. Strong players, however, are voracious hunters for profits both small and large, and will happily throw four dropships full of dudes into a few missile turrets in exchange for economic damage and chaos.

So we have the building blocks of negotiation. The final game takes shape from the major decision points. The result of arguments and exchanges throughout the game sort out the outcome.

Chasing the Rabbit

It’s useful to think of decision points as seeds that grow into decision trees. A decision tree is a full game, from start to finish, laid out as the decisions of consequence that were made during the game. Chasing the rabbit is one of the best ways to learn about your strategy game.



Let’s say you want to do a Marine-Vulture-Wraith build and you decide to do it in advance. You might decide on one barracks, expand, two factory, two starport as your major build decisions. But then what do you do when your opponent makes a Scout? Scouts are good against vultures and your opponent can make corsairs against wraiths. What do you do when your opponent defends your shenanigans and makes big bad Ultralisks?



The end result is that this build is too rigid for a flexible game. The way you arrive at builds or strategies is not to just call your idea out in the air, but to build them on the basis of the early decision points and step forward. Here’s some sample steps:

Decide that we want to start with a Barracks into expand. This seems reasonable. Play out a few games with a Barracks expand and see what it works against and what you need to watch out for. Maybe instead of making a factory, just make low tier units on both sides to test out the idea. Our Barracks expand is solid and won’t die to anything except the most all-in insane builds. Now let’s add vultures. We have two decision points, so twice the decisions. Let’s see how fast we can play this and what we can get to in the endgame. It looks like it’s hard to defend big attacks from dragoons or harass from mutalisks with mostly vultures, so we add tanks or goliaths depending on the matchup to make use of our factories. Our opponent reacts by making some different units, and the game gets a bit crazy.

We decide that adding two starports, after trying it a few times, is just not viable with this build. By the time we get there, our opponent seems to have access to a lot of good answers like scourge or goliaths or dragoons. So as we solidify our scouting and the attacks that we can make with vultures, we can then make a better decision on what end-game units to make. This is called chasing the rabbit.

It’s important to not chase the rabbit too deeply. There are too many rabbits to chase. You can see that this starts to get really dodgy if you change even one building in the above example. But what we did discover is that we can get away with barracks expand.

Go magnifies this principle. There are dozens of fuseki openings alone, which then lead into hundreds of joseki exchanges, and then players play out unique sequences in the other corners. The principle to take away from Go is that the opening shapes the game to come, but in a deep enough game, many many decision branches remain and the puzzle is far from solved.

Sente: I Go First

Perhaps the most valuable principle from Go is sente, or initiative. Sometimes a player plays a move that threatens to do something really severe, like capture a group of stones or break up the opponent’s territory (among other things), the opponent must respond immediately. If not, they’ll lose. So the opponent does. Now it’s the first player’s move again. If they have another scary move, the opponent must respond again.

In Go, it’s very common to see a half-dozen sente attacks in a row. The attacker uses the attacks to restrict the opponent while developing his own strong shapes on the board. It’s an interesting concept that applies to many strategy games, including card games like Magic and Hearthstone and real-time strategy games like Warcraft and Starcraft. While it turns out that it’s very hard to kill stones when the opponent is playing stubbornly, it’s very easy to poke around and make an nice shape for yourself by threatening to kill things. To give you an idea, professionals will say that sente in some cases is worth 15-30 stones.

Going back to Starcraft, there are all in builds that win or lose on the spot based on the success of their attack. These builds are not sente. They do not give you initiative for a follow up. Sente attacks are often hard to spot, but the key is usually a conservative offensive maneuver. In Starcraft II, for instance, making several reapers and poking around can be sente. You are forcing your opponent to make enough army to push them out and avoid economic damage. The exchange is that you yourself can expand behind them. Your investment is small, and your return is large.



Now, suppose you’ve made your reapers and expanded. Expanding is another decision point, and that decision is gote, the opposite of sente. Now your opponent has dealt with your reapers, and you’ve sunk money into expanding. Your opponent probably gets the next sente move because their army is larger at the moment. Again, they can all-in here or they can try a poke or they can even maybe just expand themselves. It’s their choice.

Sente also falls into the principles of negotiation. Sente can be exchanged for points in Go by using forcing moves to build up territory and capture stones. However, your opponent can say, OK, take that small thing, and I’ll take sente. Players can argue over sente. Many joseki can be pushed in a certain direction so that a player gives up space to and regains sente at the end of a sequence.

The law of sente in game design is that if the opponent responds well, it doesn’t last forever. Eventually, you run out of threatening stones. Eventually, there is no number of reapers that you can make to threaten your opponent. Eventually, the creatures player in Magic gets his board wiped and runs out of cards. So it’s important to think of initiative as an exchange. One player pushes forward, the other pulls back to avoid dying. Now the player that pushed has some game space to do something else of their choosing. They can attack again, or they can develop some other resource to greater benefit. But if they choose to keep attacking, the weaknesses of being aggressive and not developing start to become apparent.

The Play-Review-Study Method

If the key to winning at Starcraft is skill and hand speed, the key to not losing is builds and game knowledge. Go is the same – winning comes from calculated insight and talent. Avoiding defeat comes from knowledge of possibilities.

In the modern era, players of any game have a rich base of knowledge to draw from. People have already chased most of the rabbits very far down the hole. But books alone do not a Senpai make. Players get better by playing. Playing develops instinct and habit. In Go, it’s common for players to read out 10 or 20 moves ahead. In Starcraft, players develop prediction and positioning skills by playing.

But with play only, a player gets bad habits. The above vulture-wraith build can probably work all the way up to fairly high ranks if you master it, but against great players, they’ll destroy it. What happens when a bad idea becomes ingrained? It worked once, it can work again, right? Maybe you just need to play it again and fix what you did wrong!

This turns out to be too slow for a top player. Opportunities are missed. Bias takes over. A player needs to review their games to uproot mistakes quickly. In Starcraft, good players almost always watch their practice replays, often with their opponent. In League, coaches record matches and review them with the team on the big screen in the living room. In Go, players replay the game and examine possible moves, called variations, in detail. This is an extremely important process, as it opens up new branches of play that the player might only be able to see in hindsight.



Play and review are a cyclical pair. You play, you see where you can improve, you play again. Improvements are internalized. But in order to see the game in a whole new light, you need to study the players who have gone farther than you. You need to zoom out and take stock of how little you actually know.



And so we see the loop. Play and review. Study new ideas. Play and review. Study new ideas. Think about decisions. Have big arguments. Make exchanges. This is learning, and in a sufficiently complex game, you can keep it up forever.

Wrapping Up

There is a lot to be said for Go as a strategy game, but I think the biggest point to learning Go is that it’s the simplest strategy game available. There is only one move. Place a stone. But the principles that result from placing stones with a few simple rules become very complex. Thus, strategy. Thus, decisions, argument, exchanges.

Go strips the nonessential from the strategy game. There is only move, countermove. All moves are equal. There is no question of balance or power level. It’s a reduction. You get all of the strategy flavor with none of the decision fatigue from choosing a race or character or deck. Go is the strategy game that exposes the true genius and beauty of strategy games, and I encourage you to learn it at least a little. I promise that the lessons you learn will carry over to all of the other games that you play.