CS2 Gambling Games: What Each Mode Actually Costs
Longer bar means more of your stake goes to the house over time. Quoted verbatim from skinrave.gg/en/fairness, read 2 September 2026. These are one operator's figures for its own games, not an industry average, and no other site we rank publishes a comparable table.
- 0.5%
- Cheapest mode
- 6.66%
- Dearest mode
- 4-10%
- Cases
- 15
- Games with a printed edge
21 and 21 Multi, on SkinRave's own published table
Bait, on the same table. Mines and plinko sit at 6.5%
The widest band, because a case is priced per case
SkinRave publishes one per game. Read 2 Sep 2026
Where these modes run
Ranked partners, with the part of the board each one is actually built around. Offers and codes are read from our affiliate registry, not typed in.
100% deposit bonus
Plus rakeback and a $25K monthly race
Code SKINSVAULT
Free coins
Sign-up bonus, no deposit
Code SKINSVAULT

3 free cases
+ 200% deposit bonus
Code SKINSVAULT

Promo boost
Roll, crash, and case modes
Code SKINSVAULT
In brief
The modes you meet on a CS2 site, and the house edge each one carries. SkinRave publishes a figure for all fifteen of its games. That table is the most useful thing on this page.
The one number that decides which mode to play
Each mode on a CS2 gambling site pays back less than it takes in. The size of that gap is the house edge, and it is what separates a game you can sit with for an evening from one that clears a balance in twenty minutes.
Almost nobody prints it. SkinRave does, for all fifteen of its games, on its own fairness page. The figures below are quoted exactly as that page states them, read on 2 September 2026. Treat them as one operator's numbers rather than an industry standard, because that is what they are.
Grouped by the figure each mode carries. Quoted from skinrave.gg/en/fairness, read 2 September 2026.
Case opening and case battles
You pay a fixed price, a weighted draw picks an item, and the item is worth whatever it is worth. SkinRave lists cases at a 4 to 10 percent house edge, the widest band on its own table, because a case is priced per case rather than to a single house number.
Case battles are the same draw wrapped in a race: two or more players open the same cases at the same time and the highest total takes everything. The draw does not change, the variance does. You either double your stake or lose it.
A battle does not change the draw
Case battles use the same weighted draw as opening the case alone. What changes is variance: instead of a small loss most rounds, you take the whole pot or none of it. Same edge, bigger swings.
Crash, roll and limbo
A multiplier climbs from 1x and stops at a point fixed before the round began. Cash out before it stops and you keep the multiplier. Wait too long and the stake is gone.
Crash and roll both sit at 4 percent on SkinRave's table, limbo at 5 percent. The appeal is that you choose your own risk: a 1.2x target wins often and pays little, a 50x target almost never lands. Neither choice changes the edge. It is the same 4 percent whichever target you pick, which is the part most crash guides get wrong.
Coinflip, double and the wheel
The cheapest games on the board. SkinRave lists coinflip, double and the wheel at 2 percent each. Coinflip is one player against one player with the site taking a cut of the pot; double and the wheel are a segmented spin where colour decides the payout.
CSGOEmpire's roulette is the best known example of the format and is what its front page opens on.

Upgrader, mines and plinko
These three cost the most. Upgrader is 6 percent, mines 6.5 percent, plinko about 6.5 percent, on the same published table.
An upgrader trades an item you hold for a percentage chance at a more expensive one. Mines is a grid where you reveal tiles for a rising multiplier until you hit a bomb. Plinko drops a ball through pegs into a payout slot. All three are built to feel like decisions, and in two of the three the only real decision is when to stop.
The three that cost the most
Upgrader, mines and plinko carry the highest edges on the published table apart from cases. They are also the three built to feel most like skill. The interface asks you to make choices; the arithmetic does not care which one you make.
Blackjack is the outlier
SkinRave lists 21 and 21 Multi at roughly 0.5 percent, which is an order of magnitude below anything else on its board and close to what a land casino offers on the same game. Blackjack has a fixed, well understood optimal strategy, and an operator that deals it by the standard rules cannot charge much for it.
If you are going to play something on a skin site and you want the smallest edge against you, on this operator's own numbers it is blackjack, not cases.
Where each mode actually runs
Mode lineups differ. CSGOEmpire is a marketplace with roulette and match betting attached. Clash.gg, Farmskins and Cases.gg are case-first. CSGORoll runs roll, crash and case battles. SkinRave carries the widest board of the sites we rank, fourteen games listed in its own menu.
Payout speed is a separate question from mode choice, and it is one we do not publish a figure for: we have not run withdrawals ourselves, so any number here would be the operator's own claim. For how the draw itself is verified, read how provably fair works.

Ranked partners running the modes above. Each links to its own review.

SkinRave
100% deposit bonus Plus rakeback and a $25K monthly race, code SKINSVAULT
The only site we rank that publishes a house edge for every one of its games.
Frequently asked questions
Which CS2 gambling game has the best odds?
Does picking a higher crash multiplier change the house edge?
Are case battles better value than opening cases alone?
Why do most CS2 sites not publish a house edge?
What is the difference between an upgrader and a case?
Sources
Every legal, tax and regulator claim on this page traces to one of these 4 references.
- 1Per-mode house edge for all fifteen SkinRave games, quoted verbatim from the operator's own fairness page. skinrave.gg“House Edge. 21 Multi: ≈0.5% House Edge. 21: ≈0.5% House Edge. Wheel: 2% House Edge. Coinflip: 2% House Edge. Double: 2% House Edge. Roll: 4% House Edge. Crash: 4% House Edge. Hilo: 4% House Edge. Cases: 4-10% House Edge. Limbo: 5% House Edge. Keno: 5% House Edge. Upgrader: 6% House Edge. Mines: 6.5% House Edge. Plinko: ≈6.5% House Edge. Bait: 6.66% House Edge”
- 2SkinRave's game menu, listing the fourteen modes shown in the capture above. skinrave.gg“Battles, Double, Cases, 21 Multi, Keno, Upgrader, Mines, Roll, Plinko, 21, Hilo, Limbo, Coinflip, Crash”
- 3SkinRave's operator of record, for the figures above. skinrave.gg“SkinRave is owned and operated by RUNITUP LTD located at Avlonos, 1 Maria house 1075, Nicosia, Cyprus.”
- 4How a provably fair draw is constructed on a site that publishes its method, for the distinction between verification and pricing. skinrave.gg“For every battle server seed is generated at the time of creation, then waiting for future EOS Block Id once all slots are filled. Server seed combined with EOS Block Id, slot number and round number use the code to generate 0-9999 tickets (total 10000).”
More Education


Three of these sites give you a case before you deposit anything. 10 run the offer in total, with what each one actually hands over.Free cases Related Articles
How Provably Fair CS2 Sites Work, With a Real Example
Provably fair is the cryptographic protocol every CS2 gambling site claims to support. We walk through what it actually means, the three values you need (server seed, client seed, nonce), and a real worked example with HMAC-SHA256 you can verify yourself in 30 seconds.
Is CS2 Gambling Legal in 2026? Country-by-Country Guide
A real, sourced answer to whether CS2 skin gambling is legal where you live. Covers the New York AG lawsuit against Valve, Belgium and the Netherlands' case-opening bans, Norway's outright ban, the EU's fragmented landscape, and Valve's 2025–2026 crackdown on gambling sponsorships.
How We Test CS2 Gambling Sites: Our Full Methodology
The exact test behind every ranking on this site: what we weight, what we deliberately ignore, when it was last run, and how many of the sites we list have actually been through it.







