罫線統計ラボ
あなたが見つけた「法則」は、
ランダムそのものの姿です
さっきのシューの罫線を入力してください。その形が純粋なランダムの中でどれくらいの頻度で出るかを計算します。
「バンカー」か「プレイヤー」を押して開始、または形を選んでください。
バンカー 50.68% / プレイヤー 49.32%（タイ除外）· 1シュー60ゲーム換算 · 厳密な漸化式による計算
Calculators
Three things you can verify yourself
All free, no signup. The numbers are whatever the math says they are.
Bet & side-bet EV calculator
Pick the rules (5% commission / no-commission / Super 6) and payouts, and see the true expected value of every bet on the table.
Open calculator →Risk-of-ruin calculator
Enter bankroll, unit size and session length to see your chance of busting before you walk away. Most people are surprised the first time.
Open calculator →Betting-system simulator
Martingale, Labouchère, 1-3-2-4, flat betting — ten thousand runs, the full outcome distribution, and why "works short term" and "loses long term" are both true.
Open simulator →Road Truths
Latest
Every conclusion can be re-checked with the tools above.
- How to read baccarat roads: bead plate, big road and the derived roadsBasics · 12 min
- How rare is a five-banker streak? Not rare at allOdds · 8 min
- Big eye boy, small road, cockroach road: what each one actually countsBasics · 15 min
- Why the banker bet pays 5% commission — house edge in one numberOdds · 6 min
- The math of Martingale: a 99% win rate that still goes to zeroSystems · 11 min
- Is no-commission baccarat really better? The two rule sets, side by sideOdds · 7 min
Methodology
Where these numbers come from
We are not asking you to take our word for it. Every step below can be recomputed.
Base probabilities
Standard eight-deck baccarat, enumerated over every dealing sequence: banker 45.86%, player 44.62%, tie 9.52%. With ties removed, banker 50.68%, player 49.32%. These figures are public and checkable.
Streak probabilities
Computed by exact recursion (dynamic programming) for "at least one run of length ≥ k within n hands" — not a Monte Carlo estimate, so the answer is identical every time. That is the algorithm running in the tool above.
What we do not do
We do not predict the next hand. We publish no "hit rate". We sell nothing that claims to raise your win rate. Independent random events contain no exploitable information.
Who writes this
Every article is bylined, dated, and shows its working. Found an error? Tell us and we will correct it with a revision note. Accuracy is the only asset this site has.
FAQ
Straight answers
So reading the road is completely useless?
As a prediction tool, yes. The road records what already happened, and after the shuffle each coup is independent of all of it. It has two real uses: tracking your own bets and results, and giving you a rhythm at the table. We support the first; be careful that the second is easily mistaken for a pattern.
Then why do I keep seeing long streaks?
Because random sequences produce streaks, far more often than intuition suggests. Try the five-banker preset above — the chance of seeing at least one in 60 hands is higher than most people guess. Human pattern-finding is very good at reading signal into noise.
Banker or player?
On expected value alone, banker’s house edge after the 5% commission is 1.06% versus 1.24% for player, so banker bleeds slower. Both are negative — "better" means losing more slowly, not winning. The tie bet carries a 14.36% edge and is one of the worst on the table.
What do you sell?
The calculators and road tools are free, permanently. The paid product is bet tracking and review: automatic settlement, your real P&L curve, and how far your actual results sit from mathematical expectation. We do not sell predictions, because there are none to sell.