Take a probability triple, $(\Omega,\mathscr{F},\mathbb{P})$, representing a single coin toss.

Then \begin{align} & \Omega = \{H,T\}. \tag{Prop. 1} \\ \end{align}

Now, \begin{align} \text{A } \sigma \text{-algebra must be closed under countable union}, \tag{Prop. 2} \end{align}

so \begin{align} \text{if we have } \{H\},\{T\} \in \mathscr{F} \text{ then we must also have } \{H,T\} \in \mathscr{F}, \tag{Prop. 3} \end{align}

but \begin{align} \{H,T\} \text{ is not a possible event resulting from a single coin toss.} \tag{Prop. 4} \end{align}

What's gone wrong?

  • 9
    $\begingroup$ You are mixing up events (collections of outcomes} and outcomes. $\{H,T\}=\Omega$ is an event since it is the set containing two outcomes, but it is not, by itself, an outcome. $H$ and $T$ are outcomes, $\{H\}$ and $\{T\}$ are the corresponding events consisting of single outcomes. $\endgroup$ Jun 3, 2020 at 21:30

1 Answer 1


The event $\{H,T\}$ is that the result of the flip is either $H$ or $T$; this has probability $1$

The event $\emptyset = \{\,\}$ is that the result of the flip is neither $H$ nor $T$; this has probability $0$

So there is no problem; $\mathscr{F}= \{\emptyset,\{H\},\{T\},\{H,T\}\}$ as you might expect

  • 8
    $\begingroup$ So proposition 4 is wrong. I guess that $\{H,T\}$ got interpreted as 'tails after heads' instead of 'tails or heads'. In that case it resembles an equivocation fallacy. $\endgroup$ Jun 3, 2020 at 14:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.