How Hi-Lo Card Counting Actually Works
The Hi-Lo system is the most widely used advantage play technique in blackjack. It was developed by Ed Thorp in the 1960s and refined by Bryce Carlson into what gamblers now call the "standard" count. The core idea is simple: track whether the remaining cards are rich in high values or low values, then adjust your bet size accordingly. When the deck runs hot with tens and aces, you bet more. When it runs cold with low cards, you bet less. That's the entire premise. Each card gets a point value. The low cards from 2 through 6 are worth +1. The middle cards, 7 through 9, are worth 0. And the ten-value cards — 10, Jack, Queen, King, plus the Ace — are all worth -1. You maintain a running count by assigning these values as each card hits the table. That's literally all the system requires you to do. Here's where most people trip up. The running count by itself means nothing without context. A running count of +6 in a single-deck game is significant. A running count of +6 after three decks have been dealt in an eight-deck shoe is basically noise. You need to convert the running count into a true count by dividing by the estimated number of decks remaining. So a running count of +12 with roughly four decks left is a true count of +3, which translates to roughly a +0.75% edge over the house. At that point, you increase your bet. A true count of +1 or below means you're playing at or near the house edge, so you make the table minimum.
I spent about two years working this system in various casinos before I stopped going for personal reasons. One specific problem I ran into that most guides never mention is shuffle tracking and the "cut card effect." In most casino games, the cut card is placed about six decks into an eight-deck shoe. When that card is hit, the dealer shuffles early. This means you rarely see the final deck or so, which is often where the most favorable counts concentrate because the removal of low cards from the earlier portion skews the remainder. I learned to stop playing past what I calculated as 75% penetration because my edge was evaporating in those final messy hands. It cost me a few sessions where I could have squeezed out another quarter of my expected value, but it also kept me from making sloppy bets when the shoes were thoroughly confused. There's a common misconception that Hi-Lo is somehow outdated and unreliable. It is outdated in the sense that every casino knows about it and monitors for it. But it is not ineffective. The math still works exactly as Thorp derived it. What changed is not the system — it's the countermeasures. Modern casinos use continuous shuffling machines in many low-limit games, which completely neutralize any counting system. They also employ back-rounding on true count conversions, shuffle the shoe more frequently, and flag consistent bet spreaders for surveillance. If you're playing at a table with a CSM, counting is mathematically pointless regardless of your skill level. You need a manual shuffle game with at least seven-deep penetration to make Hi-Lo viable. Another thing beginners consistently get wrong is bet spreading. The theory says you should bet somewhere between one and twelve units depending on the true count. In practice, betting patterns that aggressive draw attention faster than anything else. I found that a three-to-five unit spread was about the maximum I could sustain before the pit dropped me. That doesn't make the system worthless — a three-to-five spread still nets you positive expectation over thousands of hands — but it does mean your hourly return is lower than the math suggests if you're betting full spread. You trade raw edge for longevity. Most people who get caught counting aren't caught because their counts were wrong. They're caught because they went from a five-dollar bet to a two-hundred-dollar bet on the same hand and the floor manager happened to be watching.
If you're just starting out, don't begin at a casino table. Get a deck of cards, shuffle it, and run through a full deck while calling out the running count and true count out loud. Do it until you can clear a deck in under thirty seconds without mistakes. Then practice with two decks. Then three. Once you're comfortable with the mechanical aspect, introduce basic strategy decisions that vary based on the count — things like insurance when the true count reaches +3 or standing on 16 against a dealer ten when the deck is rich. The count-only approach gets you partway there. Combining it with index plays pushes you into genuine territory. The main bottleneck with Hi-Lo is that it only tracks one dimension of the deck. It treats all ten-value cards as identical, but an ace is far more valuable to a player than a nine. Some advanced players layer in a side count for aces to capture that extra information, but that requires significantly more mental overhead and the marginal gain is small. For most practical purposes, the basic Hi-Lo count with correct bet spreading and basic strategy deviations is sufficient to establish a measurable edge at compliant tables.
Get the Full Details
