How HiLo works: card odds, streaks and the house edge

HiLo is an instant card game with fast rounds and simple rules: you guess whether the next card drawn will be higher or lower than the current one. The strategy sits in a few ideas: how the deck is built, how card odds work, and three terms you will see in HiLo guides, card range compression, the banker's edge and predictive volatility. This page covers all of them.

HiLo shows one face-up card and asks for a higher or lower call on the next card. HiLo pays the multiplier printed for that card when the call is correct. HiLo ends the streak when the call is wrong, and the return built on that sequence goes back to zero. The cash-out control stores the current return and starts a new sequence.

The card on screen is the only reference for the next call. A 2 or a King leaves one side with almost every remaining rank. A 7 leaves the same number of ranks above and below. The multiplier on the likely call is smaller than the multiplier on the close call.

Deck structure and set-up

HiLo uses a standard 52-card deck without jokers, with the usual ranking from Ace (1) to King (13). The deck is shuffled at the start, and one card is dealt face up as the reference for your first call.

• Card ranking: Ace = 1, 2 to 10 by number, Jack = 11, Queen = 12, King = 13.
• Deck handling: in some versions the deck is used up card by card, so the cards left change your odds. Many online and provably fair versions instead draw each card independently, as if from a freshly shuffled deck every time. In those, past cards tell you nothing.

Check which type you are playing before you try to track cards. It is usually stated in the game rules.

A standard HiLo deck holds 52 cards and 13 ranks. Four suits share each rank, so each rank starts as four cards. A single-deck table removes each dealt card from the pool that can still appear. A fresh-deck table restores all 13 ranks before the next draw, so a card that has already appeared can appear again.

The rules screen names the deck model before the first stake. Card tracking changes the remaining counts only on a single-deck table. Card tracking changes nothing on a fresh-deck table, because every draw starts from a full set of ranks.

Provably fair HiLo versions usually publish a verification screen after each round. Check that screen before relying on it. The method typically combines a server seed, a player seed and a round counter to set the card drawn, and it lets you confirm the card was fixed before the call was made. That check covers how the card was produced. It does not test the payout table, which sets the multiplier for each call and so the banker's edge.

Each operator sets its own steps for these checks. Our guide to provably fair dice explains the general method, which helps when reading a HiLo verification screen. Confirm the exact steps on the screen itself, because the seed format can differ between operators.

Card probability basics

The chance of the next card being higher or lower depends on the card showing and on which cards can still come.

• Count the ranks: with a 7 showing, six ranks are lower (Ace to 6) and six are higher (8 to King).
• Work out the ratio: from a full deck, 24 of the 51 unseen cards are lower, 24 are higher and 3 are other 7s. So higher and lower are equally likely, at about 47% each. How a tie with the same rank is settled depends on the game.
• Adjust as cards go (single-deck games only): if two Kings have already appeared, fewer high cards remain, which tilts the odds towards lower. In games that draw from a fresh deck each time, the ratio stays the same.

A face-up 7 leaves Ace through 6 below it and 8 through King above it. Those two groups contain the same six ranks. From a full 52-card deck, 51 cards stay unseen after the opening card. Three of those 51 cards are the other sevens. Twenty-four of those 51 cards are lower. Twenty-four of those 51 cards are higher.

A same-rank result follows the rule on that table. The table may void the call, count the call as a loss, or count the call as a win. Read that rule before the first stake, because the three other cards of the showing rank are the tie cases.

A single-deck count changes after each deal. Two Kings already dealt leave fewer high cards, so a lower call becomes more likely. A fresh-deck draw ignores those earlier Kings and restores the original 24 and 24 split whenever a 7 is showing.

The 24 and 24 split assumes that the showing card is the only card removed from the deck. Every extra card removed in a single-deck session changes the counts, so a player who tracks cards needs a fresh count before each call. A table that draws from a fresh deck avoids that recount, and its split stays the same on every round.

A call decision has three steps. First, count the ranks higher and lower than the showing card. Second, read the same-rank rule on the table. Third, compare the multiplier paid on the likely call with the multiplier on the other call. The higher probability and the higher multiplier do not always point the same way, so the call with the better chance is not always the call with the better return.

Probability tells you which result is more likely, not which result will happen. A call that is more likely to win can still lose on the next card. The odds describe a long run of calls, and no single round carries a guaranteed outcome.

Building a streak

HiLo lets you string correct calls together for a bigger payout. Each correct call raises the multiplier.

• First call: look at the face-up card and choose Higher or Lower.
• Win and continue: after a correct call, the new card becomes the reference. You can call again or cash out.
• Risk compounds: a long streak can pay well, but one wrong call loses everything in that streak.

The question at every step is whether the extra multiplier is worth the risk of losing what you have built.

The first correct call replaces the face-up card. The next call uses that new card as the reference. Each added call multiplies the current return by the multiplier for the card now showing. One wrong call clears that return. Cash out stores the return and ends the sequence.

Each call in a streak uses the card on screen as its only reference. The return after a correct call equals the stake multiplied by every multiplier collected so far, so the payout rises faster than the number of calls. A player who has made three correct calls has passed through three separate multipliers, and the last multiplier alone does not set the final payout.

A wrong call ends the sequence and clears the whole return built so far, including every earlier correct call. Four correct calls followed by one wrong call leave nothing from that sequence. The table keeps no partial credit, so each decision to continue is a decision about the entire return on screen.

Cashing out follows a simple test. Compare the return on screen with the chance of losing it on the next call. When the next call is close to even and the return is large, cashing out protects what is already built. When the card showing leaves one side with almost every remaining rank, the likely call pays a small multiplier, so the extra return rarely covers the risk of losing the whole sequence.

This test has limits. It depends on the payout table and the odds being what the rules state. It does not change the expected result of each call, and a cash-out rule can only control how often a large return is lost. Over many rounds the banker's edge still applies to every call made, whether the player cashes out early or late.

Key strategic terms

Card range compression

In single-deck versions, the spread of ranks left in the deck narrows as cards are dealt. Early on all 13 ranks are in play and calls are uncertain. Late in the deck, only a few ranks may be left, and some calls become far more likely to win.

• Early in the deck: all 13 ranks are possible, so uncertainty is high.
• Late in the deck: with only a handful of ranks left, you can make higher-probability calls based on what remains.

This only applies when the deck is actually used up. In fresh-deck versions there is no compression.

Banker's edge

The banker's edge is the house advantage. It comes from paying slightly less than the true odds. For example, a 1 to 1 payout on a call that wins 49% of the time costs you 2% of your stake on average.

• Read the payout table: compare what each call pays with its real chance of winning.
• Keep streaks short: the edge applies to every call, so long streaks pay it again and again.

Predictive volatility

Predictive volatility describes how uncertain a call is. It is highest around the middle cards (7 to 9), where higher and lower are close to even.

• High-volatility cards: middle cards give close to 50/50 calls, so results swing.
• Low-volatility cards: low cards (2 to 4) and high cards (Jack to King) give lopsided odds and steadier results, with smaller multipliers on the likely call.

Some players lower their stake or cash out when a middle card comes up, to protect their bankroll.

Card range compression is a single-deck effect. Early in that deck all 13 ranks can still appear. Late in that deck only the ranks that have not been dealt can appear. A call on a remaining edge rank is then much more likely to win. A fresh-deck table redraws from all 13 ranks every time, so compression never starts.

The banker's edge is the gap between the true chance of a call and the multiplier the table pays. A 1 to 1 payout on a call that wins 49 percent of the time returns less than a fair price. That shortfall is the small gap described above, paid again on each call. The same shortfall applies again on every later call in the streak, which is why a short streak pays the edge fewer times.

Predictive volatility is the uncertainty of the next call. Cards 7, 8 and 9 leave higher and lower close to even, so the result swings. Cards 2, 3 and 4, and cards Jack, Queen and King, leave one side with far more ranks. The likely call on those edge cards pays a smaller multiplier. Lower the stake or cash out when a middle card appears if the goal is to protect the return already built.

These three terms describe how the game works. None of them predicts the next card. Card range compression describes the ranks left in a single deck, not the rank that will come next. The banker's edge is a property of the payout table, and it stays the same whatever call is made. Predictive volatility measures how widely results can spread, so it tells you how large a swing to expect, not which way the next result will go.

A reader who uses these terms should be clear about what each one can change. Compression can change the odds on a single-deck table, the banker's edge changes the expected return of every call, and volatility changes the size of the swings around that return. Those three effects work at different levels, so one term does not replace another.

Practical tips

• Bet small early: in single-deck games, start with smaller bets and only raise them late in the deck, when the odds are clearer.
• Do not treat a result as due: track your wins and losses, but remember each call depends only on the cards that can still come, never on your last result.
• Split your bankroll: keep separate amounts for short, safe streaks and for longer, riskier ones, so one bad run cannot empty everything.
• Set limits: decide a loss limit and a session length before you play. Our responsible gaming page lists the tools to use.

Set the stake before the first card, not after a win. A single-deck session can use a smaller stake while all 13 ranks remain, then a larger stake only when few ranks remain. A fresh-deck session has no late-deck change, so the stake stays the size chosen at the start.

Keep a separate amount for short streaks and a separate amount for longer streaks. One wrong call on a long streak takes the whole return of that sequence. A loss limit and a session length, chosen before play, stop a run of wrong calls from using the rest of the bankroll. The tools for those limits are listed on the responsible gaming page.

None of these tips changes the table's edge. A bankroll rule controls how losses are spread over time and stake sizes. It does not lower the expected loss on each call. A loss limit caps how much is lost in one session, and a session length caps how many rounds are played, but neither makes a call with a house edge profitable over many rounds.

Use the tips as limits on your own behavior, not as a method to win. A player who keeps stakes small, sets a loss limit and stops after a set time has a more predictable session. The result still depends on the table, the rules and the calls made.

Similar instant games

HiLo shares its core with other instant games such as Dice, Plinko, Mines and Limbo: you choose your risk, the odds are visible, and crypto casino versions are provably fair. If you enjoy HiLo's blend of probability tracking and fast rounds, these titles offer varied formats built on the same foundation of instant, decision-driven gameplay. For ways to keep HiLo sessions steadier, see our HiLo risk reduction strategies.