ストリーク計算機 - 連勝/連敗確率
無料ストリーク計算機。一定数のベットにおける連勝・連敗の確率を計算。
この計算機の使い方
- シングルベット勝率をパーセントで入力(例:55)
- 評価したい連勝の長さを入力
- 合計ベット数を入力
- 連勝確率と予想される最長ランを表示
公式
P(N勝の連勝)= p ^ N
P(N敗の連敗)=(1 − p)^ N
予想最長ラン(近似)= log(N ·(1 − p))/ log(1 / p)
P(Mベットで長さNの勝利連勝が≥1)≈ 1 −(1 − p^N)^(M − N + 1)
よくある質問
Why does my expected longest streak look so long?
Variance grows logarithmically with sample size. With 1000 coin flips you’ll typically see a streak of 9-10 heads. Long streaks feel surprising but are mathematically expected — most bettors mistake them for hot/cold periods rather than ordinary variance.
How does streak length affect bankroll management?
Even a 60% win rate produces 5+ losing streaks regularly. Bankroll management (Kelly fractions, flat staking) must absorb these without ruin. Use this calculator with a streak length of 5-7 to see how often you’ll see those losing runs and size your unit accordingly.
Are sports streaks predictive?
Mostly no. Independent events (coin-flip-like markets) produce streaks purely by chance. There can be small predictive effects (injury cascades, team morale) but they’re usually overstated. Treat past streaks as variance unless you have concrete model-based reasons to believe otherwise.
What's the math behind 'expected longest run'?
For independent Bernoulli trials with success probability p over N trials, the expected longest run of successes converges to log(N(1−p))/log(1/p). It’s a logarithmic approximation that’s accurate for large N and gives the typical longest streak you’d observe.