Card Counting in Blackjack: How It Works and How to Learn It

By Originally published Last updated

Card counting estimates how the cards already dealt change the composition of the cards still available. It does not predict the next card and it does not guarantee a winning session. A trained counter may obtain a small long-run advantage only when the rules, deck penetration, accuracy, bet spread and bankroll all support it.

On this page
  1. Important Reality Check Before You Learn
  2. How to Learn Card Counting Step by Step
    1. 1) Learn perfect basic strategy first
    2. 2) Pick one beginner-friendly count
    3. 3) Practice a running count at home
    4. 4) Learn deck estimation
    5. 5) Convert to true count
    6. 6) Add bet sizing discipline
    7. 7) Add strategy deviations later
  3. Popular Blackjack Card Counting Systems Compared
  4. Running Count vs True Count
  5. When to Raise Bets and When to Stay Small
  6. Common Card Counting Mistakes
    1. Trying to count before learning basic strategy
    2. Going too fast in practice
    3. Ignoring casino conditions
    4. Playing too long when tired
    5. Making obvious bet jumps
    6. Expecting instant results
  7. How to Practice Card Counting at Home
  8. Does Card Counting Still Work in 2026?
  9. Related Blackjack Resources
  10. How to Learn Card Counting Without Burning Out
  11. Risk Controls New Counters Ignore
  12. Penetration and Table Selection (Why Half the Battle Is Off the Felt)
  13. How Much Bankroll Do You Actually Need?
  14. The Illustrious 18: Strategy Deviations That Actually Earn
  15. Online Card Counting: Why It Mostly Doesn’t Work
  16. Frequently asked questions
  17. Card-counting system directory
  18. Sources and methodology

The arithmetic is learnable; reliable execution is harder. A player must maintain the count at live speed, estimate decks, convert to a true count where required, follow basic strategy and tolerate large swings without abandoning the plan.

Most counting systems assign simple values to cards as they are dealt. The most common version (Hi-Lo) looks like this:

  • 2 through 6 = +1
  • 7 through 9 = 0
  • 10, J, Q, K, A = -1

You keep a running count as cards come out. When the running count is positive, it usually means more high cards remain. When it is negative, more low cards remain. In multi-deck blackjack, players then convert the running count into a true count by dividing by the estimated decks remaining. The true count is what serious players use to size bets and make strategy adjustments.

That sounds simple on paper, but the real challenge is doing it accurately while playing at full speed, talking to a dealer, handling chips, and staying calm.

Important Reality Check Before You Learn

Card counting is legal in many places because you are using your brain, not a device. That said, casinos are private businesses and may still refuse service, limit your play, or back you off if they think you are an advantage player. If you want the short version of how casinos respond, read how casinos stop card counting and what a blackjack backoff looks like.

Also, counting works best in the right games. Bad rules can erase a lot of your edge. Always check the table rules first:

  • Prefer 3:2 blackjack over 6:5
  • Look for dealer stands on soft 17 (S17) when possible
  • Double after split (DAS) helps
  • Late surrender helps when available
  • Good penetration matters (how deep into the shoe the dealer deals)

If you are not sure what these rule changes do to the game, visit why casinos change blackjack rules and house edge in blackjack.

How to Learn Card Counting Step by Step

1) Learn perfect basic strategy first

Do this before anything else. Counting gives you a small edge only if your basic decisions are already strong. Use a chart or a blackjack strategy wizard and drill until your decisions feel automatic.

2) Pick one beginner-friendly count

For most players, Hi-Lo is the best starting system. It is not the fanciest count, but it is proven, practical, and easier to use in real games than many advanced systems.

3) Practice a running count at home

Take a single deck and count through it one card at a time. Your goal is to end at zero. At first, ignore speed. Accuracy comes first. Once you can consistently end at zero, start timing yourself and try to get faster without making mistakes.

4) Learn deck estimation

In shoe games, you need to estimate how many decks are left. This is the part beginners usually avoid, but it matters. A running count of +6 means something very different with six decks left versus one deck left.

5) Convert to true count

Practice quick division in your head. Example: running count +6 with about 2 decks left gives a true count of +3. You do not need perfect math to the decimal; close, consistent estimates are usually enough.

6) Add bet sizing discipline

Do not jump from one unit to a giant bet just because the count rises. Build a simple, repeatable betting ramp and stick to it. Counting works over many hours, not one hot shoe.

7) Add strategy deviations later

Once your count is accurate, then start learning index plays (like when to stand, hit, or take insurance differently than basic strategy). If you try to learn everything at once, you will mix things up and slow your progress.

If you want to train properly, use a card counting trainer and work on one skill at a time: count accuracy, deck estimation, true count conversion, and then deviations.

There is no single “best” system for everyone. Systems trade off against each other on three axes: how easy they are to run accurately at full table speed, how well they correlate with actual bet sizing (betting efficiency, or BE), and how well they guide strategy deviations (playing efficiency, or PE). Execution errors can outweigh a theoretical improvement in system efficiency.

The table below covers every system we have a dedicated page for on Counting Edge. Click the system name to read the full breakdown, drills, and strategy-index tables for each.

System Level Balanced? BE PE Best for
Hi-Lo 1 Balanced 0.97 0.51 The default first system. Shoe games, solo play.
KO (Knock-Out) 1 Unbalanced 0.98 0.55 Beginners who want to skip true-count division.
Red 7 1 Unbalanced 0.98 0.54 A slightly simpler first system than Hi-Lo.
REKO 1 Unbalanced 0.98 0.55 Simplified KO — fastest learning curve.
Silver Fox 1 Balanced 0.98 0.53 Shoe games; often used as a Hi-Lo alternative.
Hi-Opt I 1 Balanced 0.88 0.61 Pitch (single/double-deck) games where PE matters.
Canfield Expert 1 Balanced 0.87 0.51 Players who want a simple system with strong PE in pitch games.
KISS II 1 Unbalanced 0.90 0.54 The simplest unbalanced count; casual counters.
KISS III 1 Unbalanced 0.98 0.56 A strong step up from KISS II without true-count math.
Zen Count 2 Balanced 0.96 0.63 Intermediate players targeting shoe games; better all-round than Hi-Lo.
Omega II 2 Balanced 0.92 0.67 Advanced players targeting pitch games; strong PE.
Revere Point Count 2 Balanced 0.99 0.55 Players who want the highest possible BE at a level-2 cost.
Hi-Opt II 2 Balanced 0.91 0.67 Advanced pitch-game counters; usually played with an ace side count.
Uston APM 2 Balanced 0.95 0.55 Intermediate pitch-game play.
Uston SS 2 Unbalanced 0.99 0.56 Very strong BE without true-count math.
Uston APC 3 Balanced 0.91 0.69 Professionals targeting pitch games with an ace side count.

Three honest takeaways: for a beginner in a shoe game, Hi-Lo or KO is the right answer — the BE/PE gap versus more complex systems is too small to justify the extra mental load. For someone targeting single- or double-deck pitch games, the jump to a level-2 balanced count like Zen or Omega II returns real EV because those games reward PE. And for everyone else, the “better” systems on the chart only matter if you can run them accurately at 80 hands per hour while a dealer is chatting and a drinks server is asking what you want — which you cannot, in most cases, without hundreds of hours of drills. A simpler count executed cleanly will almost always outperform a fancier count executed poorly.

Running Count vs True Count

This is one of the biggest beginner mistakes. Players learn the running count and think they are done. In multi-deck blackjack, you usually need the true count.

  • Running count: The raw total as cards are dealt
  • True count: Running count divided by decks remaining

Example:

  • Running count = +8
  • Decks remaining = 4
  • True count = +2

A true count of +2 is positive, but it is not the same as a true count of +8. That difference matters for bet sizing and deviations. If you want a deeper breakdown, read our page on true count conversion.

When to Raise Bets and When to Stay Small

A bet increase needs a game-specific estimate of advantage and a risk analysis of the complete betting plan. True count alone does not determine a safe stake. There is no universal betting ramp that guarantees a 1% risk of ruin.

Practice without money first. Evaluating a real plan requires the rules, penetration, frequency of each bet, variance and available bankroll.

Common Card Counting Mistakes

Trying to count before learning basic strategy

This is the most common mistake. Fix your foundation first.

Going too fast in practice

Speed without accuracy is wasted practice. Train accuracy first, then speed.

Ignoring casino conditions

A good count system does not fix bad rules, poor penetration, or a 6:5 game.

Playing too long when tired

Fatigue leads to counting errors and bad betting decisions. Short, focused sessions are better than marathon sessions.

Making obvious bet jumps

Huge bet swings can attract attention quickly. Learn to play within your bankroll and your comfort level.

Expecting instant results

Counting is a long-term edge. You can do everything right and still have losing sessions.

If you are making several of these mistakes, start with our common blackjack strategy mistakes page and tighten your fundamentals first.

How to Practice Card Counting at Home

A simple practice routine works better than random practice:

  1. 5 minutes: basic strategy flashcards or chart review
  2. 10 minutes: single-deck running count drill
  3. 10 minutes: shoe drill with deck estimation and true count conversion
  4. 5 minutes: review mistakes and repeat the weak spots

Keep a small notebook or spreadsheet. Track your time, error rate, and how quickly you can count a deck accurately. Progress is easier to see when you measure it.

Does Card Counting Still Work in 2026?

Counting requires the composition of the remaining cards to persist between rounds. Check the actual table rules, cut point and shuffle procedure. A continuous shuffler is different from a machine that shuffles an entire shoe between shoes.

Restrictions and casino responses vary. This guide does not verify current table availability or promise that a betting spread avoids detection.

How to Learn Card Counting Without Burning Out

Most people quit card counting because they skip steps. A better progression is: (1) flawless basic strategy, (2) running count accuracy at low speed, (3) true-count conversion drills, (4) bankroll and bet spread rules, and only then (5) live practice. This reduces frustration and prevents expensive early mistakes.

Use short, repeatable training blocks. Ten focused minutes of count drills every day beats one long weekend binge. Track your error rate, not just your speed, because fast wrong counts are useless in live play.

Risk Controls New Counters Ignore

  • Never raise stakes just because one shoe went well.
  • Set a max session loss before first bet and respect it.
  • Avoid alcohol and distractions during count-dependent sessions.
  • Log every session so you can spot leaks in discipline and spread.

Penetration and Table Selection (Why Half the Battle Is Off the Felt)

Penetration is how deep the dealer deals into the shoe before the shuffle. Shallow penetration reduces the time a strong true count can develop, which means fewer high-value betting opportunities. That is why two tables with identical rules can still produce very different results for a counter—one may shuffle early enough that your skill never gets enough “room to breathe.”

Pair penetration awareness with rule shopping: 3:2, reasonable double rules, and favorable surrender options still matter. If you want a structured comparison mindset beyond single metrics, read about SCORE after you are comfortable with true count and spread basics.

How Much Bankroll Do You Actually Need?

Hypothetical example: with a 1% player advantage, an average bet of 100 and 100 hands, expected profit is 0.01 × 100 × 100 = 100 units. This is a mathematical average, not a session forecast. A 1% house advantage would instead mean an expected loss of 100 units.

Fluctuation depends on variance and the betting pattern. A fixed multiple of the largest bet cannot establish a 1% risk of ruin. A session loss limit also does not change expectation or guarantee protection of the entire bankroll.

Use the risk simulator to explore its stated assumptions; it does not replace a game-specific simulation of counting with variable bets.

The Illustrious 18: Strategy Deviations That Actually Earn

Indices identify when deck composition can change the best decision. They depend on the counting system, rules and convention for converting and rounding true counts. Do not mix systems or apply a hit/stand index when surrender is the appropriate decision.

Don Schlesinger’s Illustrious 18 is a well-known study selection. Its contribution to expected value depends on the game and betting ramp; this page cannot promise a universal percentage. Learn indices from a source that specifies those assumptions, after mastering basic strategy.

Online Card Counting: Why It Mostly Doesn’t Work

If a game resets a complete deck between rounds, cards from previous rounds provide no counting advantage. Online games do not all use the same procedure: read the rules for the specific version.

Live-dealer blackjack may preserve a physical shoe, but penetration, rules, pace and restrictions determine whether counting is useful. A streak of results does not establish an advantage. We do not assume universal penetration percentages, dealing speeds or detection methods across providers.

For practice without wagering, use the counting trainer.

Frequently asked questions

Is card counting illegal?

In many places, no. Using your brain to track cards is generally not illegal. Casinos can still refuse service or back you off, so always follow local laws and casino rules.

Can you count cards online?

Usually not in standard RNG blackjack because each hand is independent and shuffled by software. In live dealer games, conditions vary, and many casinos use countermeasures that make counting difficult.

What is the easiest card counting system to learn?

Hi-Lo is the most common beginner system because it is simple, practical, and well documented.

Do I need to learn true count conversion?

Yes, for most multi-deck games. The running count alone is not enough to estimate advantage accurately in shoe blackjack.

Can I learn card counting without a trainer?

Yes, but a trainer speeds things up because it helps you practice accuracy, speed, and true count conversion in a more structured way.

Can card counting guarantee a profit every session?

No. Counting improves long-run expectation, but short-term variance is still large.

What should I master before trying live casino counting?

Near-perfect basic strategy, consistent running count accuracy, and true-count conversion under a realistic pace and distractions—plus a conservative betting ramp that matches your bankroll.

Why does deck penetration matter so much?

Deeper penetration gives the count more time to reach useful true-count territory before the shuffle, which increases the number of high-edge hands where larger bets are justified.

Do I need a team to beat blackjack?

No. Many players study and play solo. Teams can scale bankroll and coverage, but they add coordination and heat considerations. Start with solid individual fundamentals first.


Card-counting system directory

Use this directory to move from the overview to a specific method, person, or ruleset.

Sources and methodology

Expected-value examples on CountingEdge are educational estimates. Actual results depend on the exact game and execution; no count guarantees profit.