Does Card Counting Still Work?

The math has not changed

The arithmetic of counting is the same as it was decades ago: a shoe rich in tens and aces favors the player, and bets sized to that fact yield an edge. Nothing about the cards has changed. What changed is the games casinos choose to deal — and that is where the honest answer lives.

What casinos changed

Continuous shuffling machines erase the count entirely — a CSM game cannot be counted, full stop. Blackjack paying 6:5 instead of 3:2 taxes the game beyond what counting recovers in most spreads. Shallow penetration — cutting off two decks of six — starves the count of the depths where the edge concentrates. And table minimums, spread limits, and quick back-offs police the rest. Every one of these is a knob casinos turned specifically because counting works.

So the real skill is game selection

Counting in 2026 is two skills: the count, and finding games worth counting — 3:2 tables, human shuffles, decent penetration, rules that don't bleed you. They still exist; there are fewer of them, and they are unevenly distributed. Scouting — rules, penetration, pace, the room — is as much a part of the modern game as the arithmetic. The honest summary: the edge is real, small, and available exactly where conditions permit it, to players trained enough to hold it.

The honest answer, and what it depends on

The arithmetic has not changed and cannot. If the remaining cards are richer in tens and aces, the player's expectation is higher; that is a property of the rules of the game rather than a strategy anyone invented. What has changed is how many tables let a player act on it profitably.

So the real question is not whether counting works. It is whether the specific game in front of you is worth playing — and that has become a much more selective question than it once was. The skill has shifted from executing the count to finding conditions where executing it is worth the hours.

What actually kills a game

Three things do most of the damage, and none of them is surveillance. A shoe cut short, so the count rarely reaches a level worth betting into. A shorter payout on a natural, which is a permanent tax that no amount of correct play recovers. And continuous shuffling equipment, which removes the depletion the whole method depends on.

Any one of those can take a game from worth playing to not worth playing, and the third removes the possibility entirely rather than reducing it. This is why the useful skill is game selection first. A player counting perfectly at a bad table is doing careful work for nothing.

The scale of what is available

It is worth being blunt about size, because the gap between the honest figure and the imagined one is where most disappointment in this subject comes from. The advantage a good counter can obtain over a good game is small — a fraction of a percent, applied to the money at risk, realised only across a great many hands.

That is a real edge and it is worth having. An edge is not a paycheck: it does not arrive on a schedule, and it does not stop losing sessions from happening in long sequences. Anyone describing it in terms that sound like a wage is describing something else.

What has actually changed, and what has not

What has not changed: the mathematics, the mechanism, and the fact that a shoe dealt deeply enough will sometimes favour a player who is paying attention. Those are properties of the game.

What has changed is the supply of games worth the trouble. As of 2026, and on this desk's own reading of the tables it can observe: deeper cuts have become rarer and shorter payouts on a natural have spread. Shuffling equipment removes the effect entirely on the tables that use it, which is mechanical rather than a trend. The consequence is that the same skill produces far less than it once did in the same number of hours, not because the skill decayed but because the conditions did.

So the honest modern answer has two halves. Yes, it still works where the conditions permit it. And the hard part is no longer learning to count — it is finding, and being willing to travel to, the shrinking set of tables where counting is worth doing at all.

The question worth asking instead

"Does it still work" is not the useful question, because it invites a yes or a no about a thing that is entirely conditional. The better question is what a given table is worth, and that has a real answer.

It depends on how deep the deal goes, what the payouts are, how fast the hands come, and what spread the table will tolerate. Those are observable before sitting down, by anyone willing to stand and watch a shoe through before committing to it. A player who can read them is choosing games; a player who cannot is hoping, and hoping is expensive at these margins.

That reframing also settles the argument the original question keeps starting. People answering yes are usually describing a good game they have found, and people answering no are usually describing the ordinary ones they have seen. Both are reporting accurately about different tables, which is why the debate never resolves and why the general question was never the right one.

Questions

Is single-deck blackjack better for counting?

The count swings harder in fewer decks, but modern single-deck games usually pay 6:5, which poisons the well. The rules matter more than the deck count — a good six-deck game beats a bad single-deck game.