Is Card Counting Online Actually Worth It? An Honest Hourly-Earnings Breakdown

Most articles about card counting either promise a second income or declare it dead. Both are easier to write than the truth, which is a multiplication problem.

What you earn per hour is roughly:

average edge × average bet × hands per hour

Every part of that is smaller than people expect. This post walks through each one with realistic live-dealer numbers, then puts them together, then looks at the part nobody likes: how long it takes before the result is real.

The three numbers

1. Your average edge: small, and only on some hands

Counting does not give you an edge on every hand. It tells you which hands have one. Most of the shoe is still played at a small disadvantage. The money comes from betting more on the minority of hands where the count is high — that is what a bet spread is.

Averaged across every hand you play, a skilled counter on a good table typically ends up with something like half a percent to one and a half percent of their average bet. On a poor table it can be zero or negative even with perfect play.

The single biggest factor is penetration: how deep into the shoe the dealer deals before shuffling. A table that shuffles at half the shoe rarely lets the count climb high enough to matter. A table that deals three-quarters of the way through does. Which live tables qualify is its own guide.

2. Hands per hour: fewer than you think

Live-dealer tables have betting windows, dealing animations, and up to seven players making decisions. A Classic table typically manages somewhere around 40 to 60 rounds per hour. We will use 50.

That is much slower than an RNG game. It is also the only kind of online blackjack where counting works at all, so it is the number that matters.

3. Your average bet: set by your bankroll, not your ambition

Your betting unit should be roughly your bankroll divided by 1,000 to 1,500. Bet bigger than that and the risk of losing the whole bankroll before the edge shows up gets uncomfortable fast. So the size of your bankroll quietly caps what you can earn per hour.

Putting it together

Take a good table — 6 decks, cut around 75% — and a modest 1–8 spread (1 unit up to true count +1, then 2, 4, and 8 units). With rough but reasonable assumptions, that works out to about 0.74% of a unit per hand in expected value. At 50 hands per hour, that is about 0.37 units per hour.

Now attach dollars:

On a shallow table — 8 decks, cut around 50%, which is the most common live setup — the same player earns roughly a tenth of that, and with a narrower spread can be slightly negative. At a $25 unit, that is under a dollar an hour.

These are illustrative estimates, not simulation results. Different rules, a different ramp, or a different cut can move them a lot in either direction. But the order of magnitude is honest: card counting online is a low single-digit to low double-digit dollars-per-hour activity for most bankrolls. It is not a salary.

Variance: the part that decides how it feels

The expected value above is an average. What actually happens in any given session is dominated by luck.

With the good-table numbers above, the swing in a single hour is about 20 units in either direction — more than fifty times the expected win of 0.37 units. A few examples of what that means in practice:

• After 10 hours, you expect to be up about 4 units. The realistic range is roughly 60 units either way. Being down is completely normal.
• After 100 hours, you expect to be up about 37 units. There is still roughly a 40% chance you are behind.
• After 1,000 hours, you expect to be up about 370 units, and there is still something like a one-in-four chance you are behind.

Card counters call the point where skill reliably outweighs luck "N0." For this setup it is on the order of 140,000 hands — nearly 3,000 hours at a live-dealer pace.

That is the honest answer to "does it work?" It does, mathematically, on the right table. But a weekend, a week, or even a month of results tells you almost nothing about whether you are doing it right. Only the process does: did you track every card, bet by the true count, and play the right table?

When it is not worth it

Card counting online is probably not worth your time if:

• Your bankroll is small relative to the table minimum. If the minimum bet is more than about 1/1,000 of your bankroll, you are over-betting from the start.
• The tables you can access are shallow. If the shoe is shuffled at half-way, or it is an Infinite or continuous-shuffle table, the count barely moves. You are paying the house edge to watch.
• You need the money to show up this month. Variance will not respect that.
• You cannot keep up. A count that is off by a few cards is worse than no count, because it tells you to raise when you should not. Practice first.
• You enjoy side bets, chasing losses, or "hot" tables. Those habits cost more than counting can earn.

When it can be worth it

It can make sense if:

• You already play live-dealer blackjack. Then the real comparison is not counting versus a job — it is playing with a small edge versus playing with the house edge.
• You can find a deeper table and are disciplined enough to only bet up when the count says so.
• You treat it as a long-term project, measured in hundreds of hours, with a bankroll you can afford to lose.
• You like the puzzle. For a lot of players, the main payoff is playing correctly and understanding the game — with a small mathematical tailwind instead of a headwind.

The part people overlook: mistakes cost more than counting earns

For a recreational player, the bigger number is often not the edge you gain. It is the edge you stop giving away.

Players who go by feel — standing on 12 against a 2, never splitting 8s against a 10, taking insurance — commonly give up 1 to 2 percent of everything they bet compared with correct basic strategy. At 50 hands an hour and a flat $25 bet, a 1.5% leak is about $19 an hour. That is more than counting earns at the same stake on a good table.

So the first thing any tool should do for you is make every decision correct. The count comes second.

Does a calculator pay for itself?

The calculator does two jobs. It shows the best play for your hand given the cards that have actually come out, and it tracks the true count and your edge, so you know when to raise and when to sit at one unit.

A week pass costs about $20. Using the good-table numbers above, here is roughly how many hours of play it takes for the expected counting edge alone to cover that:

If you play at small stakes for a couple of hours a week, the counting edge alone will not cover it — be honest with yourself about that. If you play regularly at $25 or more per hand, or if you are currently making basic-strategy mistakes, it covers itself in an evening's worth of expected value. Whether that expected value actually shows up in any particular week is up to variance, as above.

Try it before you decide

You do not have to take any of this on trust. Open the calculator, sign in, and track one full live-dealer shoe in demo without betting. Watch three things:

1. How often the edge actually turns positive on that table.
2. Whether you can keep up with every card.
3. How many of your usual decisions the hand analysis would have changed.

One shoe will not tell you what you will earn. It will tell you whether the table can produce an edge, and whether you are the kind of player who can collect it.

Ready to size your bets? The bet spread and bankroll guide covers units, Kelly, and risk of ruin. For the math behind the edge, see expected value in blackjack, and for setting up a session, how to count a live-dealer shoe in real time.