Supply and Demand in CS:GO — A Case Study

CSGO is a decade-old game at this point, with a stable playerbase and a vibrant market economy. As such, it serves as a great example of what a mature game marketplace looks like, a model example showcasing classic economic effects that apply as expected in games. Examining game changes from 2015 (when this article was originally written), we can see demand shocks, substitution, and income effects in the wild.

Way back in September 2015, CSGO updated the game to tweak gun stats and introduce a new case (loot box) containing new gun skins. This had predictable economic effects:

• [Demand Shock] M4A1-S rifle was nerfed in power
• [Substitution] New gun skins were released, distributed via…
• [Income] Cases, which were given freely to active players

Demand Shocks — Nerfs and Buffs

Counter-Terrorist teams in CSGO have access to two main rifles — the M4A1-S, and M4A4. Historically, the M4A1-S was the clear favorite, so CSGO nerfed its power in September 2015. As a result, players shifted their usage towards the M4A4. Market demand for their respective gun skins shifted as well.

Popular M4A4 skin prices shot up overnight, nearly doubling in price in a hard step function. (ex: Asiimov, $25 → $40)

M4A1-S skins dropped accordingly. (ex: Cyrex, $25 → $17)

Curiously, the Cyrex price drop was not as sharp as the Asiimov, with a gradual decline over the following weeks. This suggests that players are quick to buy new gun skins for active use, but slow to sell off unused inventory.

This type of demand shock is analogous to government policy changes. Whenever the government announces a new tax code or subsidy, stocks in affected industries respond accordingly.

Substitution — New Skins Cannibalize the Old

The iconic AK-47 is canonized in media as the wild weapon wielded by the “bad guys”. CSGO is no exception, as the Terrorists’ AK-47 is arguably the best gun in the game. When new gun skins are released, players pay particular attention to new AK-47s.

But attention is a limited resource, and new skins draw demand away from old skins. This substitution effect can be clearly seen on the prices for old skins, who fall out of fashion and drop in desirability. (ex: Aquamarine Revenge, $40 → $30)

In fixed-price marketplaces such as League of Legends, this substitution effect manifests as reduced sales volume on older content. But in floating price marketplaces, we can observe this directly in the market price.

Income — Play-to-Earn?

Not all AK-47 skin prices dropped at the same time. Counter-intuitively, low-end skins (ex: Emerald Pinstripe, $1.40 → $1.80) actually increased in price! While the substitution between premium and budget skins is likely weaker, that doesn’t quite explain the reversal. There is another effect at play here — income from case drops.

The Shadow Case, a loot box containing desirable new gun skins, was added on September 17, 2015. Cases are readily earned, distributed to players as they play. As the cases proliferate, their value drops to be near-worthless. Most cases sell for 3¢, and the marketplace minimum fee is 2¢ (the real cost of opening a case is the $2.50 key). However, when new cases are introduced, they can sell for upwards of $1 for several weeks while supply is still tight. This creates a broad income effect, as players earn a few dollars in the wake of a new case. With a few dollars in hand, players look for skins to buy within their price range. Cheap guns like Emerald Pinstripe see a surge of purchases. This is a textbook example of income effects pushing up demand.

A quirk of game marketplaces like CSGO is that many players are not willing to deposit or withdraw any money, but are eager to trade assets earned in-game. Earning Shadow Cases by playing effectively brought many players’ net worth from $0 to $2, driving a lot of purchases of gun skins under $2. When analyzing game economies, it is important to consider discrete behavior at the individual level rather than only looking at aggregates.

Closing Thoughts

Game economies have seen a surge of interest lately, with the rise of user-generated content platforms like Roblox and oft-hyped web3 games like Axie. While MMOs are often touted as the classic example of game economies, their systems are far more complex and their data far less transparent. CS:GO serves as my go-to example for game economy phenomena when consulting on new designs.