Skip to main content
CS2 Cases
$---.--

How We Calculate Odds and Expected Value

Last reviewed: September 30, 2026

Every number on cs2cases.cc comes from one of three places: drop rates published by Valve, market prices from the Steam Community Market, and the simulator's own record of simulated openings. This page explains each formula, says which figures are official and which are our estimates, and lists the known limits. Data sources are detailed on the Sources page.

Weapon case drop rates (official)

For weapon cases, Valve published the rarity odds in 2017 for the Chinese market. Every simulated opening of a weapon case uses exactly these tier probabilities:

  • Mil-Spec (blue): 79.92%
  • Restricted (purple): 15.98%
  • Classified (pink): 3.20%
  • Covert (red): 0.64%
  • Knife or gloves (gold): 0.26%, about 1 in 385

Inside a tier, every item has the same chance: the tier rate divided by the number of items in that tier. A case with 3 Covert skins gives each of them 0.64% / 3. Knives and gloves share the gold rate the same way, finish by finish. Doppler and Gamma Doppler phases are listed in the data as separate entries, so each phase gets its own equal share of the gold rate.

Souvenirs and capsules (estimates, ×5 rule)

Valve does not publish odds for souvenir packages, sticker capsules, autograph capsules, patch, pin, graffiti or music kit containers. For these we apply the rule the official case odds follow: each tier is five times rarer than the one below it. Souvenir packages have six tiers and no gold items, which gives Consumer 80.0264%, Industrial 15.98%, Mil-Spec 3.2%, Restricted 0.64%, Classified 0.128% and Covert 0.0256%. For capsules, only the tiers actually present in the container are used and the rates are renormalised to 100%. These rates are estimates, not official figures, and we label them as such.

Float values and StatTrak

The float (wear value) of a skin is drawn uniformly between that skin's own minimum and maximum float. The wear name follows the standard ranges: Factory New 0.00–0.07, Minimal Wear 0.07–0.15, Field-Tested 0.15–0.38, Well-Worn 0.38–0.45, Battle-Scarred 0.45–1.00. A skin limited to 0.00–0.08 can therefore only drop Factory New or Minimal Wear. Skins that exist in StatTrak have a 10% chance of dropping as StatTrak; souvenir items never are.

Expected value and return per dollar

The expected value (EV) of one opening is the probability-weighted average price of what can drop: EV = Σ P(item) × price(item), where P(item) = tier rate / items in the tier. The cost of one opening is the case price plus a key at $2.49 for weapon cases (souvenir packages and capsules need no key). The return per dollar is EV / cost: 0.40 means that, on average, every dollar spent comes back as 40 cents of items. The best cases ranking sorts all weapon cases by this ratio.

Known limits of this EV, all of which tend to make it approximate rather than exact:

  • Each skin is priced at a single wear (Field-Tested first when it exists), not weighted by the probability of each wear.
  • StatTrak premiums and Doppler phase premiums are not included.
  • Items without a market price are left out, which lowers the EV.
  • Steam Market prices include Valve's 15% fee on the buyer side; selling returns less.

Statistics from simulated openings

The statistics page compares the tiers the simulator actually produced with the theoretical rates above. These are simulated openings made on this site, not real openings by CS2 players. A tier rate is only shown when the sample expects at least 10 hits (n × p ≥ 10), always with a 95% Wilson confidence interval. A knife rate for a single case needs about 37,000 openings before it means anything, so we do not publish one below that. Under 500 openings, a case only shows its opening count, best drop and latest drop.

Try it

Open the Kilowatt Case, the Dreams & Nightmares Case or any other container from the weapon case list: each page shows its drop table, its expected value and the prices used.