July 16, 2026 · 6 min read · card counting · blackjack · strategy

Card Counting Systems Compared: Hi-Lo, KO, Hi-Opt, Omega II

Search for card counting strategies and you'll find dozens of named systems, each with its own book, its own fans, and its own claim to being the smart choice. The truth is less crowded: nearly every card counting method in serious use is a variation on one idea, and the four systems this site trains — Hi-Lo, KO, Hi-Opt I and Omega II — cover the whole meaningful range, from easiest to most powerful.

This is an honest comparison, which means it will say something most system reviews won't: the differences between these counts are small, and the difference between a well-drilled counter and a sloppy one is enormous. The right question isn't which system is strongest on paper. It's which one you can run flawlessly at full table speed — and the answer to that decides everything else.

What every card counting technique shares

Every system here works the same way underneath. Each card rank is assigned a tag — plus something, minus something, or zero — and you keep one running total as cards hit the felt. Low cards leaving the deck help you, so they carry positive tags; tens and aces leaving hurt you, so they carry negative tags. When the running total says the remaining cards are rich in tens and aces, you bet more. That's the entire trick, and it's the same trick in all four systems.

What differs is bookkeeping. How many distinct tag values are there? Does the count need converting before you can act on it? Are aces counted, ignored, or tracked separately? Those three questions generate every system on this page, and they define a real trade: each refinement buys a little accuracy and costs a lot of mental workload.

The two axes: level, and balanced vs unbalanced

Counters describe systems along two axes. The level is the largest tag value a system uses: a level-1 count only ever moves the total by one, while a level-2 count has some cards worth two, capturing finer information at the cost of harder arithmetic.

Balance is subtler and matters more day to day. In a balanced count the tags across a full deck sum to exactly zero, which is a built-in error check — count down a deck and you should land on zero. But a balanced running count means different things with five decks left versus one, so before betting you must divide by the decks remaining to get the true count. An unbalanced count deliberately breaks the zero-sum property so that the raw running count itself becomes actionable: no division, no deck estimation, just a threshold to watch for. That single design choice is the main reason beginners find some systems dramatically easier than others.

Hi-Lo: level 1, balanced — needs the true-count conversion.
KO: level 1, unbalanced — no conversion, bet off the running count.
Hi-Opt I: level 1, balanced — needs the conversion, plus an ace side count for full strength.
Omega II: level 2, balanced — the conversion, two-point tags, and a recommended ace side count.

Hi-Lo: the standard for a reason

Hi-Lo tags 2 through 6 as plus one, 7 through 9 as zero, and tens and aces as minus one. It was introduced by engineer Harvey Dubner, who presented it at the 1963 Fall Joint Computer Conference in Las Vegas as a practical simplification of Ed Thorp's ten-count. Julian Braun then refined it on IBM's computers, and the polished version appeared in the 1966 second edition of Beat the Dealer, with Stanford Wong's later work making it the reference count that most books, software and simulations are built on.

Sixty years on, it's still the default recommendation, because it sits at the sweet spot: a betting correlation of .97 — meaning it captures nearly all of the information that matters for bet sizing — from tags simple enough to run at full speed while holding a conversation. Its playing efficiency of .51 is the weakest of the four, which sounds damning until you remember that most of a counter's edge comes from betting, not playing deviations. Hi-Lo's real cost is the true-count conversion: estimating decks remaining and dividing under pressure is a genuine skill that takes drilling.

KO: the count that removes the division

The Knock-Out system, introduced by physicist Olaf Vancura and engineer Ken Fuchs in Knock-Out Blackjack (first published in 1996, revised in 1998), makes exactly one change to Hi-Lo's tags: the 7 counts as plus one instead of zero. That tiny edit unbalances the count — a full deck now sums to plus four instead of zero — and that imbalance is the entire point. Because the count drifts upward as the shoe depletes, the drift itself stands in for the deck adjustment, and you can act on the raw running count against a fixed key number. No estimating decks, no dividing, ever.

What does that convenience cost? Less than you'd expect. KO's betting correlation is .98 — on that metric it actually edges out Hi-Lo — and its playing efficiency of .55 is respectable. The honest fine print: an unbalanced count is least precise exactly when precision pays most, at extreme counts and for fine-grained playing deviations, and you lose the count-down-to-zero error check. For a recreational player who keeps fumbling the true-count math, that trade is usually worth making without regret.

Hi-Opt I: accuracy for the hands you play

Hi-Opt I comes from a different lineage. Charles Einstein proposed the underlying count in the late 1960s, and Lance Humble and Carl Cooper refined it into the version published in The World's Greatest Blackjack Book in 1980, with playing indexes computed by Julian Braun — the same programmer behind Hi-Lo's refinement. Its tags are narrower than Hi-Lo's: only 3 through 6 count plus one, only tens count minus one, and everything else — including the ace — is zero.

Ignoring the ace is the deliberate move. Aces behave like high cards for betting but roughly like small cards for playing decisions, so leaving them out of the main count makes strategy deviations more accurate: Hi-Opt I's playing efficiency is .61 to Hi-Lo's .51. The catch is the betting side, where aces matter enormously — the raw count's betting correlation is just .88, so serious Hi-Opt I players run a separate ace side count and adjust their bets with it. That's two numbers in your head instead of one, which is why this system suits players who already own Hi-Lo cold and want sharper playing decisions.

Omega II: the omega 2 card counting system

Omega II is the heavyweight of the four, developed by professional player Bryce Carlson and published in his book Blackjack for Blood, first released in 1992 and revised since. It's a balanced level-2 count: 4, 5 and 6 are worth plus two, 2, 3 and 7 plus one, the 9 minus one, tens minus two, with the 8 and ace neutral. The two-tier tags weight each card closer to its true effect on your expectation, and the numbers show it — a playing efficiency of .67, the best here, with a betting correlation of .92 that an ace side count improves further.

The cost is everything else. You're adding and subtracting twos at dealer speed, converting to a true count, and — for full strength — tracking aces on the side. Every one of those steps is a place to make an error, and errors don't just shrink a counter's thin theoretical edge, they can erase it entirely. Omega II was popular with professionals in the 1990s, but the community has since drifted toward simpler counts for precisely this reason: the extra power on paper is only real if you can execute it flawlessly for hours.

Which card counting strategy should you learn first?

Hi-Lo, first, almost regardless of your ambitions — that's this site's standing advice, and it isn't contrarian, because it's what most serious books and teams recommend too. It teaches every skill the other systems assume: tag arithmetic, deck estimation, the true-count conversion. Master it and KO becomes an afternoon's adjustment, Hi-Opt I a change of tags, Omega II an upgrade you can actually evaluate. Start with Omega II and you'll drill three skills at once, badly.

The exception is honest self-knowledge: if you already know the division under pressure is where you'll break, start with KO and lose almost nothing. And whichever you choose, remember the ranking that actually matters. A perfectly executed Hi-Lo beats a sloppy Omega II every single session, because in this game accuracy compounds and errors are expensive.

Each of the four systems has its own full explainer on this site — tag charts, worked examples, and the specific skills to drill — plus a trainer that deals real shoes and checks your count at every checkpoint, in any system you pick.

Hi-Lo — the default. Best all-round power-to-effort ratio; learn it first.
KO — the pragmatic pick. Nearly Hi-Lo's strength with no true-count math.
Hi-Opt I — the specialist. Sharper playing decisions for counters ready to add an ace side count.
Omega II — the maximalist. Top accuracy on paper, unforgiving in practice.

Reading about counting techniques is the easy part — the systems only separate winners from dreamers at speed, under pressure, with a dealer who doesn't wait. The trainer deals real shoes and checks your running count and true count in all four systems.

Train any of the four systems free →

Share this post — the link unfurls with its own card.