📚 Coin Tossing Experiments: An IB Mathematical Investigation | 掷硬币实验：IB数学探究
Coin tossing is one of the simplest yet most profound experiments in probability theory. It serves as a gateway to understanding randomness, independence, distributions, and statistical inference. Whether you are flipping a fair coin to make a decision or exploring the mathematics behind repeated trials, the humble coin toss encapsulates core principles that underpin much of IB Mathematics. In this investigation, we will move from basic probability rules to advanced concepts such as the binomial distribution, the law of large numbers, conditional probability, runs and streaks, hypothesis testing, and simulation. By combining theoretical derivations with experimental thinking, you will deepen your appreciation of how mathematics models the real world.
掷硬币是概率论中最简单却也最深刻的实验之一。它是理解随机性、独立性、分布和统计推断的门户。无论你是通过抛一枚公平硬币来做决定，还是探索多次试验背后的数学原理，这个简单的掷硬币动作都浓缩了 IB 数学课程中许多核心概念。在本次探究中，我们将从基本的概率规则出发，逐步深入到二项分布、大数定律、条件概率、连续出现与连胜、假设检验以及模拟等高级主题。通过将理论推导与实验思维相结合，你将更深刻地体会数学如何为现实世界建模。
1. The Sample Space and Basic Probability | 样本空间与基本概率
te probability. If you toss the coin twice, the sample space expands to {HH, HT, TH, TT}, where each of the four outcomes has probability 1/4. The multiplication rule tells us that for independent events, the probability of a sequence is the product of the individual probabilities: P(H then T) = P(H) × P(T) = 1/2 × 1/2 = 1/4.
当你抛掷一枚公平硬币时，可能出现两种结果：正面（H）或反面（T）。样本空间 S 为 {H, T}。因为硬币是公平的，每个结果发生的可能性相等，因此得到正面的概率 P(H) = 1/2，得到反面的概率 P(T) = 1/2。这是离散概率的基础。如果你抛掷两次硬币，样本空间扩展为 {HH, HT, TH, TT}，其中每种结果的概率都是 1/4。乘法法则告诉我们，对于独立事件，序列发生的概率等于各事件概率的乘积：P(先正后反) = P(H) × P(T) = 1/2 × 1/2 = 1/4。
2. Independence and the Multiplication Rule | 独立性与乘法法则
This means that even if you have observed ten Heads in a row, the probability of Heads on the eleventh toss remains exactly 1/2— is fair. This is often misunderstood in everyday reasoning, a mistake known as the gambler' ing k successes (Heads) in n specific trials is pᵏ(1−p)ⁿ⁻ᵏ, where p = 1/2 for a fair coin.
掷硬币实验的一个关键假设是每次投掷都相互独立。一次投掷的结果不会影响下一次。这意味着，即便你已经连续观察到十次正面，第十一次投掷得到正面的概率仍然是 1/2——前提是硬币公平。这种独立性在日常生活中常常被误解，这种错误被称为赌徒谬误。根据独立事件的乘法法则，在 n 次指定的试验中获得 k 次成功（正面）的概率为 pᵏ(1−p)ⁿ⁻ᵏ，对于公平硬币 p = 1/2。
3. Introducing the Binomial Distribution | 二项分布入门
When the order of outcomes is not important and we only care about the total number of Heads in n tosses, we use the binomial distribution. Let X be the number of Heads in n independent tosses. The probability of getting exactly k Heads is given by the binomial probability formula:
P(X = k) = C(n, k) · pᵏ · (1−p)ⁿ⁻ᵏ
where C(n, k) = n! / (k!(n−k)! ibution is symmetric when p=0.5, and it becomes the basis for many inferential procedures.
当我们不关心结果出现的顺序、只关心 n 次投掷中正面的总次数时，就要用到二项分布。设 X 为 n 次独立投掷中正面出现的次数。恰好得到 k 次正面的概率由二项概率公式给出：
P(X = k) = C(n, k) · pᵏ · (1−p)ⁿ⁻ᵏ
其中 C(n, k) = n! / (k!(n−k)!) 是从 n 次试验中选出 k 次成功的方式数。例如，抛掷公平硬币 4 次，恰好得到 2 次正面的概率为 C(4,2) × (1/2)² × (1/2)² = 6 × 1/4 × 1/4 = 6/16 = 3/8。当 p=0.5 时，该分布是对称的，并成为许多推断方法的基础。
4. Expected Value and Variance | 期望值与方差
For a binomial random variable X ~ = np(1−p). For 100 tosses, Var(X) = 100 × 0.5 × 0.5 = 25, so the standard deviation is σ = √25 ions of 100-toss experiments, the number of Heads will typically fall within about 2 standard deviations of the mean, i.e. between 40 and 60. Understanding these parameters is essential for predicting long-term behaviour and testing hypotheses.
对于二项随机变量 X ~ B(n, p)，正面的期望次数为 E(X) = np。如果你抛掷公平硬币 100 次，预计会得到 100 × 0.5 = 50 次正面。方差衡量分布的离散程度，计算公式为 Var(X) = np(1−p)。在 100 次投掷中，Var(X) = 100 × 0.5 × 0.5 = 25，因此标准差为 σ = √25 = 5。这告诉我们，在多次重复的 100 次投掷实验中，正面次数通常会落在均值附近大约两个标准差的范围内，即 40 到 60 之间。理解这些参数对于预测长期行为和进行假设检验至关重要。
5. The Law of Large Numbers in Action | 大数定律的实际表现
bility p = 0.5 with ever higher certainty. This does not mean that the absolute difference |X − n/2| will shrink to zero; in fact, it tends to grow as √n. What shrinks is the relative error. For example, after 10 tosses you might observe 7 Heads (0.7), a 0.2 deviation from 0.5. After 1000 tosses, a deviation of 0.02 (e.g. 520 Heads, 0.52) is much more likely than a deviation of 0.2. This principle is the foundation of statistical regularity and is often demonstrated through simulation.
概率论中最著名的定理之一是大数定律。它表明，随着投掷次数 n 的增加，正面的相对频率（即比例 X/n）会以越来越高的确定性趋近于真实概率 p = 0.5。这并不意味着绝对差值 |X − n/2| 会缩小到零；事实上，绝对差值通常会随着 √n 增长。缩小的是相对误差。例如，在 10 次投掷后你可能观察到 7 次正面（0.7），与 0.5 的偏差为 0.2。而在 1000 次投掷后，偏差 0.02（如 520 次正面，0.52）比偏差 0.2 更有可能出现。这一原理是统计规律性的基础，常通过模拟加以演示。
6. Conditional Probability and Bayes' Theorem | 条件概率与贝叶斯定理
8). You pick one coin at random, toss it, and observe Heads. What is the probability that you picked the biased coin? Let B be the event of picking the biased coin, and H be observing Heads. Using Bayes' theorem:
P(B|H) = [P(H|B) × P(B)] / [P(H|B)P(B) + P(H|fair)P(fair)]
If P(B) = 0.5, P(H|B) = 0.8, P(H|fair) = 0.5, then P(B| o about 0.615. This simple example illustrates how evidence updates our beliefs about an unknown parameter.
掷硬币也为条件概率提供了自然的应用场景。假设你有两枚硬币：一枚是公平的（P(H)=0.5），另一枚是有偏的（P(H)=0.8）。你随机选取一枚硬币，抛掷一次，观察到正面。问这枚硬币是有偏硬币的概率是多少？设 B 为选中有偏硬币的事件，H 为观察到正面的事件。利用贝叶斯定理：
P(B|H) = [P(H|B) × P(B)] / [P(H|B)P(B) + P(H|公平)P(公平)]
若 P(B) = 0.5，P(H|B) = 0.8，P(H|公平) = 0.5，则 P(B|H) = (0.8 × 0.5) / (0.8×0.5 + 0.5×0.5) = 0.4 / (0.4+0.25) = 0.4/0.65 ≈ 0.615。因此，观察到一次正面后，硬币有偏的概率从 0.5 提高到了约 0.615。这个简单的例子说明了新证据如何更新我们对未知参数的信度。
7. Runs, Streaks, and Patterns | 连续出现、连胜与模式
nt, the human mind is attuned to patterns. A sequence of H, T, H, T, H, T might be seen as "alternating", while H, H, H, H, T, T appears to have a "run" a row are not at all unusual. Many gamblers are surprised by this and falsely believe that a long run must be "due" to break— Understanding the mathematics of runs helps prevent misinterpretation of random sequences.
尽管每次投掷都是独立的，但人脑天生就会寻找模式。序列 H, T, H, T, H, T 可能被视为“交替出现”，而 H, H, H, H, T, T 则显得有一段“连续”。在概率论中，n 次投掷中正面最长连续出现的期望长度与 n 的对数成正比。对于公平硬币，在 200 次投掷后，最长连续正面的期望长度约为 log₂(200) ≈ 7.6，因此连续出现 7 次或 8 次正面绝非罕见。许多赌徒对此感到惊讶，并错误地认为长连胜“必然”会中断——这正是常见的谬误。理解连续现象的数学原理有助于避免对随机序列的曲解。
8. Hypothesis Testing with Coin Tosses | 基于掷硬币的假设检验
alternative H₁ be p > 0.5. You observe 15 Heads. Is this result statistically significant? Under H₀, X ~ t H₀ and conclude the coin may be biased. This simple test mirrors the logic of inferential statistics.
假设一位朋友声称他有一枚公平硬币，但你怀疑这枚硬币偏向正面。你决定进行一个实验：抛掷硬币 20 次，记录正面的次数。设原假设 H₀ 为 p = 0.5，备择假设 H₁ 为 p > 0.5。你观察到 15 次正面。这个结果是否统计显著？在 H₀ 下，X ~ B(20, 0.5)。p 值为 P(X ≥ 15) = 1 − P(X ≤ 14)。利用二项概率或正态近似，该 p 值约为 0.021。若我们采用显著性水平 α = 0.05，则结果显著，意味着我们会拒绝 H₀ 并推断该硬币可能是有偏的。这一简单检验反映了推断统计学的逻辑。
9. Simulation of Coin Tossing Experiments | 掷硬币实验的模拟
riment many times yields an empirical distribution that can be compared with the theoretical binomial distribution. The table below shows the theoretical and simulated probabilities for X = 0 to 5 Heads in 10 tosses of a fair coin, based on 1000 repetitions of the simulation.
| k Heads | Theoretical P(X=k) | Simulated Relative Frequency |
|---|---|---|
| 0 | 0.0010 | 0.0012 |
| 1 | 0.0098 | 0.0095 |
| 2 | 0.0439 | 0.0428 |
| 3 | 0.1172 | 0.1146 |
| 4 | 0.2051 | 0.2089 |
| 5 | 0.2461 | 0.2510 |
Simulation reinforces the connection between theoretical probability and empirical observation. Small discrepancies are due to sampling variability, and as the number of repetitions increases, the simulated frequencies converge toward the theoretical probabilities—another demonstration of the law of large numbers.
在数字时代，我们可以利用随机数生成器瞬间模拟数千次投掷。例如，生成一个介于 0 和 1 之间的均匀随机数 U，若 U < 0.5 则计为正面，否则为反面。对 n 次投掷重复这一过程，我们便得到正面的模拟次数。将整个实验重复许多次，就能获得一个经验分布，可与理论二项分布进行比较。下表展示了抛掷公平硬币 10 次，正面次数 X 从 0 到 5 的理论概率与基于 1000 次模拟重复的相对频率。
模拟强化了理论概率与经验观察之间的联系。微小的差异源于抽样变异性，随着重复次数的增加，模拟频率会向理论概率收敛——这又一次证明了大数定律。
10. The Gambler's Fallacy and Psychological Biases | 赌徒谬误与心理偏差
The gambler' t past independent events affect future outcomes. If a coin shows Tails five times in a row, many individuals feel that Heads is "due" r coin, P(Heads) remains 0.5 regardless of history. This bias can lead to poor decision-making in gambling and other areas of life. Mathematically, we can demonstrate that the coin has no memory: the conditional probability P(H₆|T₁∩T₂∩T₃∩T₄∩ .5. Understanding independence sharply differentiates statistical thinking from intuition.
赌徒谬误是指错误地认为过去的独立事件会影响未来的结果。如果一枚硬币连续出现五次反面，许多人会感觉下一次“该出”正面了。实际上，对于公平硬币，无论历史如何，P(正面) 始终是 0.5。这种偏差可能导致赌博和生活中其他领域的不当决策。数学上，我们可以证明硬币没有记忆：条件概率 P(H₆|T₁∩T₂∩T₃∩T₄∩T₅) = P(H) = 0.5。理解独立性使统计思维与直觉截然不同。
11. Real-World Applications and Extensions | 实际应用与延伸
Coin tossing models extend far beyond games of chance. In cryptography, key generation often relies on truly random bit sequences, which can be thought of as long series of coin tosses. In clinical trials, randomization of patients into treatment and control groups mimics a fair coin to avoid bias. Moreover, the binomial distribution is used in quality control, genetics, and survey sampling. The simple act of tossing a coin thus connects to powerful mathematical frameworks that support decision-making under uncertainty. Further investigations might explore how to detect a biased coin using sequential analysis, or how the Central Limit Theorem allows the binomial distribution to be approximated by a normal curve for large n.
掷硬币模型远远超出了机会游戏的范围。在密码学中，密钥生成通常依赖于真正的随机比特序列，这可以视为一长串的掷硬币过程。在临床试验中，通过类似于抛公平硬币的方式将患者随机分配到治疗组和对照组以避免偏差。此外，二项分布还广泛应用于质量控制、遗传学和调查抽样。因此，投掷硬币这一简单行为连接着强大的数学框架，支持在不确定性下做出决策。进一步的探究可以探讨如何利用序贯分析检测有偏硬币，或者中心极限定理如何使得二项分布在大 n 时可用正态曲线近似。
12. Reflection and Further Investigation | 反思与延伸研究
Through this investigation, we have seen how a single coin toss branches into a rich tree of mathematical ideas—probability rules, binomial distribution, expectation, variance, laws of large numbers, Bayesian updating, run theory, hypothesis testing, simulation, and psychological pitfalls. Each coin toss is a miniature laboratory for statistical thinking. As an IB student, you might extend this work by designing your own experiment with a possibly biased coin, collecting data, performing a chi-squared goodness-of-fit test, or exploring the arc-sine law for the proportion of time the number of Heads exceeds Tails. The coin, deceptively simple, is a key that unlocks the door to probability and statistics.
通过这次探究，我们看到一次硬币抛掷如何衍生出丰富的数学思想——概率规则、二项分布、期望、方差、大数定律、贝叶斯更新、连续理论、假设检验、模拟和心理误区。每一次掷硬币都是统计思维的微型实验室。作为 IB 学生，你可以拓展这项工作：用一枚可能有偏的硬币设计你自己的实验，收集数据，进行卡方拟合优度检验，或者探究正面次数超过反面次数的时间比例所遵循的反正弦定律。看似简单的硬币，其实是打开概率与统计学大门的一把钥匙。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图