Skip to content
LAB

Hi-Lo

Cards arrive one at a time. Call each one higher or lower. Every correct call grows the multiplier — and one wrong call ends it. Cash out whenever you like.

Hi-Lo

Guess higher or lower than the current card. Multiplier accumulates with each correct guess. Cash out anytime.

Bankroll
Bets0
Win %
Net P/L
Streak
Biggest win
Current multiplier 1.00×
Place a bet to draw the first card.
Presets
Auto-bet
How dishonest operators rig this game 3 documented tricks
01 Push-rate inflation

Mechanism. Tie cards (push) are weighted higher than 1/13 — operator wants you to "guess again" because each push slightly drains your edge and morale.

Red flag. Track push rate over 200+ rounds. Expected 1/13 ≈ 7.7%. Significantly higher = deck-weighting.

02 Multiplier ceiling cap

Mechanism. After ~10 correct guesses, operator displays multiplier 50× but caps cashout at 30× silently. Players don't notice until withdrawal.

Red flag. Read cashout terms before chasing long streaks. The displayed multiplier MUST be the actual payout.

03 Card-RNG seed reuse

Mechanism. Some operators use the same HMAC float for both "current card" and "next card" — guessing higher/lower is then non-trivial to lose.

Red flag. Each card must use a NEW float from the stream (different cursor). If your "next card" rank is deterministic given current card, the RNG is broken.

For the full compendium across all games, see The Book of Casino Dirty Tricks.

Server seed hash
Server seed (revealed after rotation)— pending rotation —
Client seed
Next nonce

The rules and the math behind each guess

Mathematical Audit Benchmark

How Does This Compare to 0% House Edge Protocols?

Audited 100% RTP

Instead of standard 1.00% to 4.00% casino house margins, compare with mathematically verified 100% RTP Originals (0.00% House Edge) on Duel Casino:

Duel 100% RTP Dice
100% RTP Dice: Exact 1:1 mathematical return (0.00% edge).
Duel VIP Rakeback
Instant VIP Rakeback: 0x wagering requirements with code Vip.

Hi-Lo House Edge Comparison
Hi-Lo Probability Bounds: Calculating expected card draw odds and mitigating house margin.
card_index  = floor( HMAC_float · 52 )         // 0..51
rank        = card_index mod 13               // 0=Ace, 12=King
suit        = card_index ÷ 13                 // 0=♠, 1=♥, 2=♦, 3=♣

A round opens with one face-up card. Your job: predict whether the next card ranks higher or lower. The payout for a correct call follows directly from that call’s probability:

if guess = "higher":   p = (12 − rank) / 12
if guess = "lower":    p = rank / 12
step_multiplier  =  (1 − HE) / p

Same-rank outcomes are pushes — the revealed card simply becomes your new reference point and the round continues with your stake untouched. Correct guesses stack multiplicatively, so ten right calls produce far more than ten times the base multiplier. Quit while you’re ahead; misjudge once and the streak is gone.

Why every card needs its own HMAC float

A properly built provably-fair Hi-Lo pulls each card from a distinct cursor position in the HMAC stream — one nonce, cursor advancing with every reveal. A fifty-card streak is therefore fully reconstructible from the original (server, client, nonce) triple plus cursor indices. Shady implementations sometimes feed the identical float into both the current card and the next one, or run a broken stream — either flaw makes outcomes predictable or statistically dependent, and either is detectable in verification.

Manipulations operators have been caught using

Sourced from the book:

  • Push-rate inflation. The true push chance is 4/52 = 7.7%. Versions engineered toward 12% or more bleed players quietly — a push never shows as a loss, but it burns attention and demands another decision you didn’t plan for.
  • Multiplier ceiling cap. The screen advertises 50×; the cashier quietly refuses anything above 30×.
  • Card-RNG seed reuse. Sharing one float between current and next card hands a fully predictable round to anyone who checks.

Strategy: where the edge actually lives

Hi-Lo is unusual among in-house games in that your choices visibly change outcomes — though the effect is small and exists only relative to the operator’s edge. The obvious play is the right one: take “higher” on low cards, “lower” on high cards. That maximises win probability per guess. The house edge applies on top of that probability regardless, so the long-run expectation stays negative no matter how you call it.

Three places where play style compounds:

  1. Cashout discipline. A single guess on a strongly biased card wins roughly 92% of the time. After N straight calls your multiplier approximates (0.99/0.92)^N — but your bust probability compounds on the same schedule. The streak looks invincible right up until it isn’t.
  2. Starting card selection. Rounds that open on a 7 are the toughest in the deck — neither “higher” nor “lower” has a meaningful edge.
  3. Tie handling. Pushes cost nothing financially, yet each one spends a decision and a slice of focus. Long streaks almost always contain at least one.

Frequently asked questions

Which starting card gives the best position?

Ace, 2, or 3 favours “higher” calls — near-certain wins, thin multipliers. Jack, Queen, or King mirrors that for “lower.” A 7 sits at the bottom of the list: it’s close to a coin flip in both directions.

Can I cash out after every correct guess?

Yes, and the design encourages it. The cashout button unlocks after your first successful call. Bank small multipliers whenever you want — or push toward 100×+ knowing one wrong guess returns the lot to the house.

How do I verify a Hi-Lo round was honest?

Once the seed rotates, derive the cards yourself. HMAC-SHA256(revealed_server, “{client}:{nonce}:0”) returns bytes; every 4-byte chunk maps to one card_index. The reconstructed sequence must match the round you actually played. Our Hi-Lo verifier does this for you.