Blackjack Card Counting

What Card Counting Actually Is

The biggest myth first: card counters do not memorise every card that comes out of the shoe. Nobody could. What counters do is much simpler. They assign every card a small value (+1, 0, or -1), add those values together as cards leave the shoe, and use the running total to estimate whether the remaining cards favour the player or the dealer.

The principle is straightforward. A deck rich in ten-value cards and aces is good for the player (more blackjacks, more successful doubles, more dealer busts on stiff hands). A deck rich in low cards is good for the dealer. By tracking which way the imbalance has shifted, the counter knows when to bet more (favourable deck) and when to bet less (unfavourable deck). The edge over time comes from betting bigger when the math is in your favour.

It is legal. You are using your brain, not a device. But casinos are private property and can refuse service or ban you for any reason, including the suspicion that you are counting. Most modern counters spend as much effort hiding the count as tracking it.

Why It Works: The Math

Basic strategy assumes a fresh, neutral shoe where every card is equally likely to come out next. In a multi-deck game that gets dealt out over many hands, the composition shifts. Once a chunk of low cards is out and the remainder is rich in tens and aces, the basic strategy assumption is no longer accurate, and the player’s expected value has shifted.

The shift per “true count” unit (more on what that means below) is roughly half a percent. So as the count climbs:

By betting more when the count is +2 or higher and less (or sitting out) when the count is negative, the counter shifts the average expected value of their wagered money from a loss to a small positive number. That small positive number, multiplied by enough hands, is the entire game.

The Hi-Lo System

Hi-Lo is the standard card counting system, popularised by Edward Thorp’s Beat the Dealer in 1962 and still the most widely taught. It is “balanced” (a complete deck counts to exactly zero), simple enough to learn in a week, and accurate enough that the more advanced systems do not offer enough additional edge to be worth the extra mental load for most counters.

The card values:

• Cards 2, 3, 4, 5, 6: count as +1 (low cards leaving the shoe is good for the player)
• Cards 7, 8, 9: count as 0 (neutral, ignore them)
• Cards 10, J, Q, K, A: count as -1 (high cards leaving the shoe is bad for the player)

That is the entire system to memorise. You start at zero when the shoe is shuffled, and you add or subtract one for every card you see, including the dealer’s up-card and the cards in every other player’s hand. The running total is called the running count.

From Running Count to True Count

In a single-deck game, the running count and the true count are the same thing. In a multi-deck game (6 or 8 decks, the standard online and live setup), the running count alone is misleading. A running count of +6 means very different things at the start of a shoe (six decks remaining, so the imbalance is spread thin) versus near the end of a shoe (one deck remaining, so the imbalance is concentrated).

The fix is the true count: the running count divided by the estimated number of decks remaining. So a running count of +6 with 3 decks left equals a true count of +2. A running count of +6 with 1 deck left equals a true count of +6, which is a far more favourable shoe.

Estimating decks remaining is the hardest mechanical skill. Counters learn to glance at the discard tray (the visible pile of dealt cards) and judge the remaining decks to the nearest half-deck. With practice this becomes automatic; without practice, the true count calculation is wrong every hand and the whole system breaks down.

Betting With the Count: Spreads

The true count tells you how big to bet. The ratio between your smallest and largest bet is called your bet spread, and it is the lever that determines both your edge and your visibility to the casino.

Bigger spread equals more earnings per hour. It also equals more attention from the floor. A 1-to-8 spread is openly aggressive and will get you noticed at most casinos within a session or two. A 1-to-4 spread is more sustainable but earns roughly half as much. Professional counters spread harder; recreational counters tend toward the conservative end and accept lower hourly EV in exchange for longer playing time before the heat arrives.

The “Wonging” technique (back-counting a table from outside until the count goes favourable, then sitting in) used to be a way to get an aggressive spread without showing the casino your minimum bet. Most casinos now block this with “no mid-shoe entry” rules.

Strategy Deviations (The Illustrious 18)

Counting also changes a small number of basic strategy decisions. When the true count is extreme, certain plays that were correct under basic strategy become wrong, and vice versa. These are called index plays or deviations, and they add roughly 0.1% to 0.2% to a counter’s edge on top of the betting advantage.

The canonical list is Don Schlesinger’s Illustrious 18: the 18 most valuable deviations to learn. The single most important is insurance at true count +3 or higher (the only situation where the insurance side bet is correct). Others include standing on hard 16 vs dealer 10 when the count is positive, doubling 10 vs 10 when the count is very high, and a handful of others. The full list and decision math lives in our Illustrious 18 guide.

Counters layer deviations on top of basic strategy as their baseline. The two work together; neither works without the other.

A Realistic Training Plan

Card counting is a skill that takes weeks of practice to develop and months to deploy reliably under pressure. A workable plan:

1. Weeks 1-2: Master basic strategy first. Without it, counting is worthless. Drill the chart until it is automatic, and you can play perfect basic strategy without thinking.
2. Week 3: Learn the Hi-Lo card values. Practise by counting through a single deck as fast as possible. Time yourself; the target is under 25 seconds for a full deck, with a final count of zero. This is the gateway skill.
3. Weeks 4-5: Practise the running count while playing through RNG hands online (with the count open beside you to check). Build up to maintaining accurate count across 50 hands without errors.
4. Week 6: Add true count conversion. Practise estimating discard-tray depth and dividing running count by decks remaining.
5. Week 7: Add a 1-to-4 bet spread tied to the true count.
6. Week 8 onward: Add the Illustrious 18 deviations one at a time, in order of value.
7. Live play: Only after weeks of error-free practice at home. Start at minimum-bet tables, single 1-hour sessions, with conservative bet spread. Build heat tolerance gradually.

The Heat Management Reality

Casino countermeasures are aggressive and well-developed. Floor staff are trained to spot bet variations that correlate with deck composition. Surveillance cameras record every session. Major properties subscribe to shared databases (the Oregon Surveillance Network, Griffin Investigations historically) that flag counters across multiple casinos.

Common countermeasures you will face:

• Continuous Shuffle Machines (CSMs). Cards returned and reshuffled after every hand. Counting impossible. Avoid these tables completely if you plan to count.
• Low penetration. The cut card placed high in the shoe, so the dealer reshuffles before reaching the favourable late-shoe counts. This kills counter EV without affecting basic strategy players.
• Mid-shoe entry bans. You can only join the table at the start of a fresh shoe, eliminating Wonging.
• Shuffle when you raise your bet. The dealer reshuffles whenever you increase your bet significantly. Direct counter-counter response.
• Back-off. Polite request to stop playing blackjack (often allowed to play other games). One step short of a formal trespass.
• Trespass / formal ban. Banned from the property; ID flagged in shared databases.

Counters use camouflage to delay this: not always increasing bets perfectly with the count, occasionally betting “wrong” on purpose, tipping the dealer (looking like a generous regular rather than a calculating one), drinking (or appearing to), making small talk, looking away from cards during meaningless hands. None of it is foolproof; all of it buys time. The more aggressively you spread, the faster the heat arrives.

Online Card Counting: Mostly Dead

RNG (digital) blackjack reshuffles after every hand, which makes the running count reset to zero every hand. There is nothing to count. Card counting in RNG games is impossible, regardless of what online “systems” might claim.

Live dealer blackjack uses real cards from a physical shoe and is theoretically countable. In practice the conditions are harsh: shoes are usually 8 decks, penetration is poor (typically 50%, sometimes worse), shuffles are frequent, and per-hand speed is slower than a live table (fewer counting opportunities per hour). Bet sizes are also typically capped lower than live VIP tables. The combined effect is an edge so thin it is rarely worth pursuing.

The honest picture: online blackjack is fundamentally a basic strategy game. If you are interested in counting, the only realistic venue is land-based casinos with real shoe games. Our recommended Malaysian casinos are evaluated on their main-game quality, not counting potential.

Other Counting Systems

Hi-Lo is the standard, but it is not the only system. The major alternatives:

• KO (Knock-Out): An “unbalanced” system that skips the true count conversion (you bet based on the running count alone). Easier to learn; slightly less accurate. A common gateway system for new counters.
• Hi-Opt I and Hi-Opt II: More accurate than Hi-Lo, but require side-counting aces separately. The accuracy gain is small; the mental load is substantial. Most pros stick with Hi-Lo.
• Wong Halves: Uses fractional values (e.g., 5 counts as +1.5, 9 counts as -0.5). Marginally more accurate than Hi-Lo. The fractions make tracking much harder; few people use it in live play.
• Red 7 / Zen Count: Various other systems with their advocates. All variations on the same principle.

The honest summary: for almost every counter, the right system is Hi-Lo. The added complexity of more accurate systems delivers tiny additional edge in exchange for substantially higher mental fatigue, which usually costs more in errors than the system’s accuracy buys.

Bankroll and Risk of Ruin

Even with a real 1% edge, blackjack variance is high. The math of risk of ruin says how likely you are to lose your entire bankroll before your edge turns it into long-term profit:

• 100 unit bankroll: Roughly 75% risk of ruin. Way too small. You will almost certainly bust before variance smooths out.
• 500 unit bankroll: Roughly 25% risk of ruin. Still risky.
• 1,000 unit bankroll: Roughly 10% risk of ruin. Workable for serious recreational counting.
• 5,000+ unit bankroll: Near-zero risk of ruin. Professional-level.

“Unit” here means your minimum bet. At RM25 minimum, a 1,000-unit bankroll is RM25,000. This is one of the major reasons MIT and similar teams pooled bankrolls: it let individual players bet meaningfully without each one risking ruin.

The Realistic Earn Rate

Honest numbers. A skilled counter playing a good 6-deck S17 DAS game with reasonable penetration, using a 1-to-8 spread, can expect roughly a 0.75% to 1.25% edge over the long run. At 80 hands per hour with a $25 average bet ($2,000 wagered per hour), that is $15 to $25 expected per hour. Variance around that average is large; individual sessions can swing thousands of dollars in either direction.

A more conservative 1-to-4 spread to delay heat earns roughly half as much per hour. A counter playing 20 hours a week with conservative spread might expect $200 to $400 per week before variance, before travel costs, and before the inevitable backoffs that eventually shrink the available playing venues. It is real money. It is not movie money.

Team Play and the MIT Era

Counters figured out early that team play multiplies the edge in two ways. First, by pooling bankrolls, individual players can bet larger without exceeding their personal risk of ruin. Second, by splitting roles (one player “spotting” a favourable count from minimum bets, signalling a “big player” to sit in only when the count is high), the team disguises the count from the casino’s perspective.

The MIT teams of the late 1980s through the early 2000s, led at various times by Bill Kaplan, J.P. Massar, and others, ran this approach professionally and reportedly cleared substantial sums before counter-surveillance caught up. They were not the first (Ken Uston ran similar teams in the 1970s), and they were not the only ones, but they became the most famous through journalist Ben Mezrich’s book and the 2008 film “21.” Both took significant liberties with the facts. The full story is in our blackjack history guide.

Common Myths

• “You need a photographic memory.” No. You only track plus-one and minus-one changes to a running total. The mental arithmetic is grade-school level; the skill is maintaining accuracy under distraction.
• “Counting means winning every hand.” No. You still lose individual hands at normal frequency. The edge comes from betting bigger on the favourable hands and smaller on the unfavourable ones, over many sessions.
• “Casinos can have you arrested for counting.” No. Counting is legal everywhere. Casinos can refuse service, ask you to leave, or formally trespass you. They cannot have you charged.
• “Counting works online.” Mostly no. RNG games reshuffle every hand; live dealer games have poor penetration and frequent shuffles. The remaining edge is marginal at best.
• “Hi-Lo is outdated; the pros use better systems.” No. Most professional counters still use Hi-Lo because the marginal accuracy gain of more advanced systems does not justify the added mental load.
• “I can spot a counter at a table.” Without surveillance training, very unlikely. Skilled counters camouflage their play deliberately. Floor staff who think they can spot counters by feel usually cannot.

Frequently Asked Questions