When I was in college, I accidentally took a course on Statistics. Fast forward ~ 20 years, the learnings of that one single course helped me become a better SRE.

Some sample questions I shared with some engineers over the past few months, and many of them got it wrong.

Throughput on Monday was 10tps, Tuesday was 30tps, and Wednesday was 50tps. What was the throughout for those three days?

My office is 10 kms away. I drive up at 40kmph, and down at 60kmph. What's my average speed?

Out of 15 requests, 50% were served within 50ms. Another 50 served within 100ms. What's the P90?

There are 90 requests in a minute. From 10ms till 100ms. What's the P0 and P100?

Current epoch timestamp is 160034560. How many seconds from epoch is it in IST?

Pod A' throughput is 10% lesser than PodB. Which is 11% more than PodC. Are Pod A and Pod C' throughput the same?

Again, these are not interview questions that I ask. These come up in middle of another conversation and most engineers don't feel the need or gravity to be able to answer these correctly.

And this situation is alarming for some who apply as Distributed Systems experts.

If you have a weekend free, start with these simple-and-small books.

Fortunately, my job requires me to keep reading these "mini-books" to make sure my ability to deal with numbers doesn't erode.

However, if you don't find yourself having to deal with such problems:
- Practice with your kid
- Practice at the local grocery store, next time you go shopping
- Practice with the taxes mentioned on the invoice

You will be surprised, how many daily errors you will catch ;-)