The Real Expected Value of Opening CS2 Cases (With the Math)

The Real Expected Value of Opening CS2 Cases (With the Math)

The "is it worth opening this case" question gets the same answer every time: no, statistically. But the math behind that answer is more interesting than the conclusion. The expected value of a case opening is not a vibe — it is a sum of products you can compute on a napkin, and once you set it up correctly the entire CS2 unboxing economy makes sense.

This article walks the calculation on three cases that are actually open right now (May 2026 prices): Kilowatt, Fever, and Revolution. By the end you should be able to plug any case into the same template and get its real EV in five minutes. For the official drop rates that feed the math, see CS2 case opening odds explained; for the broader cases context, see the cases pillar guide.

The expected-value formula

Expected value (EV) of one case opening is:

EV = sum over all skins ( drop_chance(skin) × net_market_value(skin) )

Where:

• drop_chance(skin) = the tier probability divided by the number of skins in that tier.
• net_market_value(skin) = the average market price across exteriors, weighted by float clipping, after the Steam Market 15% fee, multiplied by (1 + StatTrak premium × 0.10) to fold in the StatTrak roll.

The drop_chance figures are fixed by Valve:

To get a comparable real return, subtract the cost of opening — case + key — from the EV. The Steam Mobile Authenticator key is $2.49, and the case itself is whatever the Steam Market median is at the moment.

Why most published "EV" numbers are wrong

Three mistakes show up in 90% of community EV charts:

1. They use Steam buy-now prices, not the 15% net. A skin listed at $10 nets you $8.50 if you list at the same number. The Steam Market takes 5% Steam fee + 10% game fee on every sale.
2. They ignore the StatTrak roll. A 10% chance of StatTrak adds roughly 5-15% to the average value of weapon drops. Skipping it understates the EV by a couple of cents per opening, which compounds at scale.
3. They compute the gold tier as "average knife price" without accounting for the dilution inside the gold tier (13 finishes × 5 exteriors × StatTrak roll = 130 distinct knife SKUs from one case, with Battle-Scarred Slaughter and Field-Tested non-StatTrak Sapphire wildly different prices).

Get those three right and the numbers tell a coherent story. Get them wrong and you get the inflated EV figures that float around YouTube thumbnails.

Worked example 1 — Kilowatt Case

Case price (Steam Market median, May 2026): roughly $2.20. Key price: $2.49. Total stake per opening: ~$4.69.

Drop pool by tier (full skin list):

*Average net price is the float-weighted average across exteriors, after 15% Steam fee, with the 10% StatTrak roll folded in.

EV calculation:

• Mil-Spec contribution: 79.92% × $0.18 = $0.144
• Restricted contribution: 15.98% × $0.85 = $0.136
• Classified contribution: 3.20% × $3.20 = $0.102
• Covert contribution: 0.64% × $45.00 = $0.288
• Knife contribution: 0.26% × $155 = $0.403

Total EV per opening ≈ $1.07, against a stake of $4.69. That is a −77% expected return per opening, or roughly $3.62 lost per opening on average. Over 1,000 openings, you would expect to lose around $3,600 — with a long tail of variance where one lucky StatTrak Karambit Fade pull could overshoot the mean.

Worked example 2 — Fever Case

Case price (Steam Market median, May 2026): $0.70. Key: $2.49. Stake: **$3.19**.

The Fever Case lineup runs heavier on mid-tier drops than Kilowatt does, with a Covert M4A1-S that carries most of the weapon-tier value:

EV: 0.7992 × 0.12 + 0.1598 × 0.55 + 0.0320 × 2.10 + 0.0064 × 28.00 + 0.0026 × 135 = ~$0.78 per opening. Stake $3.19. Return: −76%.

The headline knife pool average looks lower than Kilowatt because the Fever Case pulls from an older knife collection where the headline finishes (Marble Fade, Doppler) trade at a different price band than the Kilowatt-era Kukri lineup. Same expected loss percentage; different absolute dollar values.

Worked example 3 — Revolution Case

Case price: $0.45. Key: $2.49. Stake: **$2.94**.

The Revolution Case is the cheapest of the three by case cost, but the EV does not improve proportionally — the average skin value is also lower:

EV: 0.7992 × 0.10 + 0.1598 × 0.40 + 0.0320 × 1.50 + 0.0064 × 15.00 + 0.0026 × 110 = ~$0.57. Stake $2.94. Return: −81%.

Revolution is actually a worse return than Kilowatt despite the cheaper case price — the headline values across all tiers are smaller, and the knife pool dilution is heavier. This is the pattern across most of the cheap discontinued-or-near-it cases: low cost, low EV, similar negative percentage.

Recap table — the three cases side by side

The percentages cluster between −75% and −85% on every active case Valve has shipped in the last decade. The few cases that flirt with positive EV are all retired or near-retired — see the discontinued cases section below.

Why the math is so brutal

The 80% Mil-Spec floor is the structural problem. Four out of five openings give you a sub-dollar skin. To break even on a $4.69 stake, the remaining 20% of openings need to average $23.50 net. Because gold drops happen 0.26% of the time, the gold tier alone needs to carry ~$1,800 of average net value to break even — which would require an average knife net price of ~$1,800. The actual average knife net price across most cases is $100-200.

Restated: the cases pay out at roughly 20-25 cents on the dollar. There is no case in CS2 where that ratio is meaningfully better at standard market prices. If a case appears to have positive EV in a community calculator, almost always one of the three mistakes from earlier in this article is the cause.

Where the math flips: discontinued cases

The above is true for cases in the active drop pool. For cases that have left the drop pool — they no longer drop after weekly care packages — the EV calculation changes because:

• The case itself appreciates over time. The CS:GO Weapon Case (the original) trades at $80+ on Steam Market in 2026, up from a sub-$1 base.
• The skins inside also appreciate as the supply stops growing.
• But your stake also goes up by exactly the case's appreciation, so opening a discontinued case at current market prices is rarely profitable; the profit is in holding the case unopened.

The cases worth opening for a non-negative return are typically the ones where (case + key) is below the EV of the drop pool — and that condition is briefly satisfied during transition windows (a case has just left the drop pool, the unboxing demand has not yet pushed up the case price, the key cost is fixed at $2.49). Those windows are short. For the current snapshot of which cases have already appreciated, the top expensive cases roundup is the reference; a dedicated /blog/are-discontinued-cases-worth-buying-2026 deep-dive ships in a later week of this cocoon.

Direct buy vs case opening

If your goal is to own a specific skin, opening cases is the worst path. A worked comparison for the AK-47 Inheritance from the Kilowatt Case:

• Direct buy on Steam Market: ~$60 for a Field-Tested copy.
• Case opening to land an Inheritance: the AK Inheritance is one of two Coverts at 0.32% chance. Expected number of openings to hit one: 1 / 0.0032 = ~313 openings. Stake: 313 × $4.69 = $1,468.

That is more than 24× the direct buy cost in expectation. The variance only widens the gap — half the time you would need more than 217 openings to hit the first Inheritance, and about 5% of the time you would still be opening at attempt 938. The opening path is thrilling, but if you actually want the skin, just buy it. For where to buy efficiently, the marketplaces guide covers fees, payout speeds, and which platforms to use depending on the skin.

When opening cases makes sense anyway

The only honest answer: when the entertainment value of opening exceeds the expected loss. If a $4.69 stake gets you 30 seconds of suspense and you would have spent that on coffee anyway, the EV is irrelevant — you are paying for the experience, not the items. The trap is treating it as an investment, which the math says it is not.

For the people who do want to optimize: keep the openings rare, target cases with the strongest Covert lineup relative to the case + key cost, and never set up auto-key purchases. The math does not budge, but the ceiling on the variance is set by your stake size, not your hopes.

FAQ

What is the average return on opening a CS2 case? Across active cases at May 2026 prices, the expected return per opening is between −75% and −85%. You get back roughly 15-25 cents per dollar on average.

Which CS2 case has the best expected value? Among currently-active cases, the Kilowatt and Fever cases sit at roughly the same −76% return — neither is meaningfully better. Discontinued cases sometimes show positive EV briefly, but the case price moves to absorb the difference within weeks of leaving the drop pool.

Is it worth opening cases for the chance of a knife? Statistically, no. At 0.26% per opening, you would expect to spend ~$1,800-1,900 in stakes to land one knife on average, and the average knife from any case is worth $100-200 net. The EV of "spend until I get a knife" is also strongly negative.

Does the StatTrak roll affect the EV calculation? Slightly. StatTrak is an independent 10% roll on top of the rarity tier; folding it in adds 5-15% to the weapon contribution depending on the StatTrak premium of the specific skin. It does not change the negative-return conclusion, but it is the difference between a sloppy EV calculation and a correct one.

What about Souvenir packages — same EV math? Conceptually yes, but the inputs are different. Souvenir packages drop during Majors (free, no stake other than your watching time), and the value is dominated by the chance of a Souvenir Dragon Lore or similar headline drop. A dedicated breakdown lands in /blog/souvenir-skin-mechanics-and-pricing in a later wave of the cocoon.

How accurate are community case-opening calculators? Quality varies. Most use Steam buy-now prices instead of the 15% net, which inflates EV by 17.6%. The good ones (CSGOLuck's calculator, a few maintained Reddit calculators) get the fee right and disclose their methodology. The bad ones produce numbers that look 10-20% better than the real EV.

The simplest way to know what your inventory is actually worth — including the cases you have not opened and the skins you already own — is the inventory calculator on this site. One click, multi-marketplace prices, no math.