The Math of Snowballing
Marv framed it against fighting games. A character at 1% health hits exactly as hard as they do at 100%. Damage output and health sit in separate buckets, so a comeback stays on the table the whole match.
In an RTS like StarCraft 2, your unit count serves as both your health pool and your damage output. When five units fight three, the five focus down the three so fast that all five survive. Five minus three isn’t two.
That number was settled long before either army moved.
If your bot builds workers without dropping a beat and the opponent misses a cycle, you pull ahead by three workers. The income from those extra workers buys an expansion before the opponent can afford theirs. That earlier expansion is what funds your five units to their three. Every advantage you bank gets spent on generating the next advantage, and the interest keeps climbing.
The game rewards the economic engine that built the army. The micro of the units on the screen is secondary to that. Every game has a dominant structure like this that dictates what actually matters.
The Compounding Test: Three Questions
1. Does an early advantage make later advantages easier to get?
This is the litmus test. In an RTS, having more workers means you collect resources faster, which builds more workers. The advantage feeds itself. In a fighting game, landing a combo doesn’t make your next combo do more damage. The advantage resets.
Ask: “If I get ahead early, does the game help me stay ahead, or do I have to keep earning it?”
2. Does the game have resets, and where?
Every game has moments where the state equalizes. In a fighting game, it’s between rounds. In a MOBA, it’s death timers and respawns. In a card game, it’s the draw. You can be ahead on board and brick your next three draws. In chess, there are no resets. Every move permanently changes the board.
Resets are the brakes on compounding. The fewer resets, the more every decision echoes.
Ask: “Where does the game wipe the slate?”
3. What resource, if taken away, makes winning impossible?
Strip away micro, strategy, and luck. What’s left that you absolutely cannot win without? In an RTS, it’s economy. No minerals, no units, no game. In a 2D fighting game, it’s space. How much you control compared to your opponent decides how many options each of you gets. Pin them in the corner and their list of answers shrinks to whatever you allow.
The thing you can’t win without is the thing the game is actually about. Everything else is window dressing.
Ask: “If I had to win with only one advantage, what would it be?”
