I have been trading Counter-Strike skins since 2015. That means I was around for
the era when jackpot was CS:GO gambling — you threw skins into a pot, a wheel
spun, one person walked away with everything. If you played in 2016, you played
jackpot. There was not much else.
Last month I sat down and did something I had been putting off: I went through
every site I have an account on and tabulated what each one actually runs today,
game mode by game mode. Not what their landing page advertises — what is actually
in the lobby when you log in.
I expected jackpot to be down. I did not expect this.
Twenty-six percent. It is now rarer than plinko.
Here is the full breakdown, sorted by how many of the 42 sites run each mode:
| Game mode | Sites | Share |
|---|---|---|
| Coinflip | 27 | 64% |
| Roulette | 25 | 60% |
| Case opening | 24 | 57% |
| Crash | 23 | 55% |
| Raffles | 20 | 48% |
| Plinko | 18 | 43% |
| Case battles | 18 | 43% |
| Upgrader | 16 | 38% |
| Blackjack | 15 | 36% |
| Poker | 14 | 33% |
| Esports betting | 13 | 31% |
| Jackpot | 11 | 26% |
Below that it falls off a cliff — dice and mines on four sites each, keno and
towers on two, and baccarat, limbo and hi-lo on exactly one apiece.
The usual explanation is Valve's 2016 crackdown, and that is part of it. But
Valve hit every mode, not just jackpot, and coinflip survived at 64%.
The real reason is structural, and it is about liquidity.
Jackpot needs a crowd. A pot with two people in it is just a slow coinflip. A pot
with twenty needs twenty people online at the same moment, on the same site,
willing to wait for it to fill. That was fine in 2016 when four or five sites had
all the traffic.
Coinflip needs one other person. Crash needs nobody at all — the multiplier
climbs whether anyone else is playing or not, so a crash game with three players
online works exactly as well as one with three thousand.
When the market fragmented across dozens of sites, the modes that survived were
the ones that work with thin traffic. Jackpot did not fragment well. The handful
of sites still running it properly are mostly the ones with enough concurrent
users to fill a pot in seconds — and there are not many of those left.
I used to think mode count was a decent proxy for whether a site was serious.
More games, bigger operation, right?
The numbers killed that idea. The average site in my table runs 5.8 modes, but
four run ten or more and three run exactly one. And the single-mode sites are not
the bad ones. A site that only does case opening and does it properly is a more
honest proposition than one that bolted nine half-finished mini-games onto a
landing page to look like a casino.
Mode count is a menu. It says nothing about whether the food is any good.
After going through all of them, the order that matters turned out to be almost
the reverse of what gets advertised:
Can you withdraw? Depositing is frictionless everywhere — that is the
whole design. Getting money back out is where sites differ, and the verification
steps that only appear when you try to leave are the ones worth knowing about
before you deposit, not after.
House edge. This is the number that decides your result over time. On case
opening it is routinely above 10%, which is far worse than any other mode in the
table. Most sites do not publish it, and on several you have to derive it from
the drop odds yourself.
Can rounds be verified? Provably fair is checkable in about two minutes if
a site implements it properly. Copy the server seed hash before you play, rotate
the seed afterwards, run the revealed seed through SHA-256, compare. If it does
not match, the site changed it. That test is binary and there is no "close
enough".
The welcome bonus. Last, and by a distance. A large offer attached to a
40x wagering requirement is worth less than a small one you can actually
withdraw.
I keep the full table on my own site — 42 sites against 13 game types, sorted
alphabetically rather than ranked, because it is an index and not a top list:
https://www.csgopromocodes.com/cs2gamblingindex
It also names the twelve sites I do not have verified game-mode data for,
rather than filling those gaps in from their marketing pages. I would rather
publish a hole than a guess.
Disclosure, because it matters and Hive deserves it straight: that site earns
affiliate commission when someone signs up to a listed platform. It costs you
nothing extra and commission rate is not part of how I score sites, but you
should know it before you read anything I write about this subject. I am telling
you because every comparison site in this niche has the same incentive and most
of them bury it.
Source of the numbers: my own accounts on each platform, checked during
August and early September 2026. Where a site had changed its lobby since I last
logged in, I re-checked rather than trusting my notes.
18+ only. Every mode in that table has a house edge, which means the expected
result over enough bets is a loss. That is arithmetic, not an opinion, and
verifying a round does not change it — it only confirms the loss was honest.
Never deposit skins or money you cannot afford to lose. Free confidential help at
BeGambleAware.
