Blackjack is a rare case for casinos: the composition of the deck changes after every hand, and that changes the game’s mathematics. If cards are not returned to the deck immediately and the dealer deals from a shoe (multiple decks) or from a single deck with incomplete shuffling, every ten, ace, or five that has come out affects the expectation of the next hand. Card counting is built on this: you track the deck’s “temperature” and make decisions—from bet sizing to deviations from basic strategy—based not on guesswork, but on numbers.
Why It Works: An Unbalanced Deck and a Shift in Expectation
When more “big” cards (tens and aces) remain in the deck, the player more often makes a 21 blackjack and wins more, while the dealer more often busts by pulling high values. An excess of small cards, on the contrary, helps the dealer draw safely. Counting is a compact way to measure this skew. You keep a running count, and in multi-shoe games convert it to a true count by dividing by the number of decks remaining. The higher the true count, the higher your expected value—and the more justified it is to increase your bet and adjust your play.
Ace–Five: Minimalism That Starts the Engine
Ace–Five (“ace–five”) is a starter technique for training attention and bet rhythm.
How to Count
- Start from zero.
- Each exposed 5: +1.
- Each exposed A (ace): −1.
- The higher the sum, the “richer” the deck is for the player.
How to Adjust Bets
- Double your bet when a threshold is reached (e.g., +5).
- Or tie bet size directly to specific count values.
Pros: barely interferes with decision flow; can be learned in an hour or two of practice.
Cons: small edge—more a training step than a fully battle-ready system.
| Card | Value |
|---|---|
| 5 | +1 |
| A | −1 |
Hi-Lo (“Plus–Minus”): The Gold Standard of Simplicity
The most famous method, popularized by the film “21.” Cards are divided into three groups; a high count is a signal to raise the bet and sometimes deviate from basic strategy (e.g., insurance or doubles in marginal spots).
| Card | Value |
|---|---|
| 2–6 | +1 |
| 7–9 | 0 |
| 10–A | −1 |
Pros: easy to learn; provides a noticeable EV boost with disciplined play.
Cons: coarse weighting (for example, a 5 affects EV more than a 2, but in Hi-Lo they are counted equally).
Hi-Opt I: Deliberate Caution
The distinctive feature of Hi-Opt I is that aces are ignored in the count, and weight focuses on the cards that most influence dealer busts.
| Card | Value |
|---|---|
| 3, 4, 5, 6 | +1 |
| 10 | −1 |
| All others (including A) | 0 |
Pros: simple arithmetic and a balanced risk profile—ideal for a careful style.
Cons: harder to apply aggressive bet spreads; some information about aces is lost.
Halves (“Half-Points”): When You Need Maximum Precision
This is a “sharpened” version of counting with fractional weights. In practice, it’s often presented with whole numbers (twice the original values) for convenience, but the idea is the same—weights are proportional to a card’s impact on expectation.
| Card | Value |
|---|---|
| 2 | +1 |
| 3, 4 | +2 |
| 5 | +3 |
| 6 | +2 |
| 7 | +1 |
| 8 | 0 |
| 9 | −1 |
| 10, J, Q, K, A | −2 |
Note: In Stanford Wong’s “Professional Blackjack” (1975), these weights are half as large (from −1 to +1.5).
Pros: high precision, especially for fine decisions on doubles/splits and insurance.
Cons: greater cognitive load; mistakes quickly erase the edge.
The Thorp System: Mathematics at the Limit
Edward Thorp is a blackjack legend and the author of “Beat the Dealer.” His weighting approach, which reflects the shift in expected value with pinpoint accuracy, remains theoretically exemplary but is extremely difficult in practice. J, Q, and K are not counted.
| Card | Value |
|---|---|
| 2 | +5 |
| 3 | +6 |
| 4 | +8 |
| 5 | +11 |
| 6 | +6 |
| 7 | +4 |
| 8 | 0 |
| 9 | −3 |
| 10 | −7 |
| A | −9 |
Pros: model-level accuracy.
Cons: doing it “in your head” is extremely hard; partially ignoring picture cards complicates live-table application.
From Count to Action: Bet Ramp and Index Plays
Counting is half the job. It’s decisions that make the money:
- Bet ramp. Define in advance how your stake grows with the true count. For example: TC ≤ 0 — minimum; TC +1 — 2× min; TC +2 — 4×, etc. Make the scale realistic for your bankroll.
- Index deviations. Basic strategy is optimal for a neutral deck. At high counts, answers to “borderline” questions change: whether to take insurance, whether to double A,8 against 6, etc. The index set depends on the system; learn at least the top 10 (insurance, 16 vs 10, 15 vs 10, 10 vs 10, 12 vs 3/2, A,7 vs 2/3, etc.).
- Deck estimation. In multi-shoe games convert the running count to a true count (divide by decks remaining). Without this, decision accuracy drops.
How Much to Practice: Less Myth, More Practice
The common belief that learning to count takes “many months” is true only for complex systems like Halves at high speeds. For Ace–Five or Hi-Lo, a few weeks of regular short sessions is enough:
- Pure arithmetic. Turn over a deck and keep the count, checking the final sum (for balanced systems, it should be zero).
- Speed. Stopwatch + one deck until you reliably hit 25–30 seconds without errors.
- Table simulation. Count alongside basic strategy and your bet ramp; say decisions out loud.
- Noise and distractions. Music, conversation, and “multitasking” to approximate casino conditions.
The key is discipline. Even perfect counting loses to impulsive decisions and bankroll-management violations.
How Casinos Reduce Your Edge
The industry knows no less about counters than you do, so it implements countermeasures:
- More decks and shallower penetration. The earlier the dealer shuffles, the less time you have to “harvest” the edge.
- Bet-spread control. Max limits, bans on sharp jumps, and “no mid-shoe” entries.
- Shuffle machines and CSMs. Continuous shuffling makes counting meaningless.
- Behavioral analytics. Observers and software track the correlation between your bet and the count. If detected—warnings, rule changes, a “friendly” back-off, or a ban.
The takeaway is simple: the more visible your bet spread and your basic-strategy deviations at the “right” moments, the higher the risk of being noticed. Balance yield and camouflage.
Online vs. Live: Where Counting Actually Matters
In typical RNG-based online blackjack, the deck is effectively refreshed after every hand—the shoe is virtually reshuffled instantly. Counting is useless in such conditions. It’s a different story with a live dealer in a live casino: cards are often partially shuffled there, and you can truly count. Important:
- track the cut-card depth (how much is burned before the shuffle);
- know local table rules (S17/H17, doubles after split, surrender, number of decks);
- remember the platform rules: obvious bet upscaling is unwelcome and your account may be limited.
What to Bring to the Table: A Quick Guide to Choosing a System
- If you’re starting out and want to feel the mechanics: use Ace–Five as a warm-up, then switch to Hi-Lo.
- If your style is cautious and you value simplicity and a low risk profile—try Hi-Opt I.
- If you’re ready for a high cognitive load for extra precision—learn Halves (until it’s automatic).
- Leave the Thorp system as an intellectual exercise; for most real games it’s overly cumbersome.
And remember: counting is a tool, not magic. It doesn’t erase variance or turn every session into a cash-out; it simply helps you choose the moment, justify risk, and make your betting logical. The right system, learned indices, a suitable bankroll, and a calm mindset are the four pillars of profitable blackjack.
