On July 16, 2025, Valve changed how trading works in CS2. Every item received in a trade became "Trade Protected" for 7 days: you can use it in game, but you can't sell it, trade it on or move it to a storage unit until the week is over. During that week the original owner can reverse the trade. It was sold as an anti-scam measure. For the market, it was the biggest rule change since the old trade holds.
A rule like this is easy to argue about and hard to measure. So we measured it. We pulled weekly sales and turnover for every category we track, plus the daily CS2 market cap index, and compared the weeks around July 16 with each other.
The week the volume fell off a cliff
In the week starting July 7, 2025, tracked marketplaces recorded 8.97 million CS2 sales worth $378 million. In the week starting July 14, the week the update went live, that dropped to 5.72 million sales and $258 million.
That is 36% fewer sales and 32% less money in a single week. The following week was even weaker on turnover: $229 million, 39% below the last full week before the change.
One thing to keep in mind: the week of July 7 was itself unusually busy, about 7% above the week before. Part of that was probably traders moving items before the new rules arrived. But even against a normal June week, the drop is far too large to be a coincidence. On a four-week average, sales after the update were 19% lower and turnover 24% lower than in the four weeks before.
You can see the overall market activity for today on our markets overview.
Cases took the hardest hit
Not every category reacted the same way. Here is the change in weekly sales, comparing the week of July 7 with the week of July 21:
Cases: -66% (from 559,262 to 191,456 sales)
Stickers: -48%
Gloves: -32%
Knives: -32%
Weapon skins: -32%
Other items: -19%
Cases make sense as the biggest loser. They are cheap, they are bought and resold in bulk, and a lot of case trading is pure flipping: buy low on one market, sell a bit higher on another. A 7-day lock kills that loop. If you can't resell for a week, a margin of a few cents is not worth the risk. Case turnover fell 66% in two weeks, from $1.54 million to $0.53 million.
Stickers, which are also cheap and traded in volume, behaved the same way. Knives and gloves fell too, but they came back faster, because people who buy a $500 knife usually don't plan to resell it within a week anyway. Glove sales were already above their pre-update level by mid-August, while case sales were still about 20% below it in September. If you follow the case market, our CS2 case prices page shows where each one trades today.
The market cap dropped 13% in eleven days
Volume is one side. Prices are the other. Our headline CS2 market cap index stood at $6.92 billion on July 15, 2025, the day before the update.
On July 16 it slipped to $6.75 billion. Then it kept falling, and on July 27 it hit $6.02 billion. That is a loss of about $900 million, or 13%, in eleven days. For a market that size, it is a very sharp move for a change that did not touch a single item's rarity or supply.
The logic behind the fall is simple. When an item can't be resold for 7 days, every buyer takes on a week of price risk. Buyers want a discount for that risk, and sellers who need liquidity accept lower prices. A market that loses speed also loses a bit of value. You can follow the index day by day on the CS2 market cap page.
How long did the recovery take?
This is where the data surprised us. The panic was real, but it did not last.
Market cap: back above its July 15 level on September 1, 2025, about six and a half weeks after the update.
Turnover: the week starting August 25 reached $406 million, already above the $378 million of the last week before the update. That took six weeks.
Number of sales: the slowest to come back. Weekly sales only passed the pre-update level in the week of September 8, with 9.92 million. That took eight weeks.
The order matters. Money came back before the number of trades did. In the eight weeks after the update, the average sale was worth $44.67, compared with $42.16 in the eight weeks before. Fewer small flips, slightly larger trades on average.
By the autumn the market was not just recovered, it was running at record levels. The week of September 8 crossed $470 million in turnover, and a few weeks later came the biggest trading week in CS2 history. The 7-day rule did not break the market. It changed what kind of trading made sense.
Who actually lost
If the overall numbers recovered within two months, it is fair to ask whether anyone really paid for the change. The data points to three groups:
Fast flippers. Anyone who earned money on quick resells of cheap items, cases and stickers above all, lost their main tool. Case sales were still about a fifth below their pre-update level two months later.
Sellers who needed money right away. In the weeks after July 16 prices were lower and buyers were fewer. Anyone forced to sell in that window sold at a discount of up to 13% compared to the week before.
Arbitrage between marketplaces. Buying on one platform and immediately selling on another became a 7-day bet instead of a same-day trade. Price gaps between markets are now a slower opportunity, which is exactly why it pays to compare prices across marketplaces before you buy rather than after.
Long-term holders, on the other hand, mostly lost nothing. If you bought in late July and held, you were back in profit by September.
What this tells us about the next rule change
Valve will change trading rules again. It always does. The July 2025 update gives a useful template for what happens next time:
Volume reacts first and hardest, often in the very first week.
Cheap, high-turnover items (cases, stickers) take the biggest and longest hit.
Prices fall more slowly than volume and bottom out about one to two weeks later.
If the change does not touch supply, the market tends to come back within one to two months.
The smartest move during the drop was not to panic sell. It was to watch the numbers. If you hold items you care about, a price alert tells you when they move, so you don't have to guess.
Explore the data
