Back to topics

How can a good spam filter still flag many genuine messages?

Imagine 10,000 messages, only 100 of them spam. A fictional filter catches 90% of the spam but falsely flags 1% of the 9,900 genuine messages. That produces 90 correct flags and 99 false ones. The small false-alarm rate acts on a much larger group.

4–5 min

Questions you could explore

  • Which denominator answers the question “How trustworthy is a flagged message?”

Start reading

Ready to try your ideas aloud?

Your topic carries over. Record up to 2½ minutes, depending on your plan.

Maths and statistics

When one long wait changes the average

An average wait can sound straightforward until one person waits much longer than everyone else. The mean and median offer two ways to summarize that same group. Their different responses to an unusually large value raise a practical question: what do we want an average to tell us?

5–7 min
Maths and statistics

How is random sampling different from random assignment?

A study can choose people at random, assign them to groups at random, do both, or do neither. Those choices answer different problems: who the findings may describe, and how fairly the study compares conditions.

3–5 min
How can a good spam filter still flag many genuine messages? | Speech Topic | speaking.app