Hacker News Reader

Your assumption being that I don't know calculus? Anyway I stand by my original statement, if you're working in data analysis (even as a frequentist) you should be able to see the truth in it; "data scientists" aren't deriving MLEs for wacky distributions and using the mean to parametrise a Gaussian doesn't require calculus. Thanks for the link, I guess?

View on HN

If you read any classic text on statistics like Feller you can't get past the first chapter without hitting calculus. Any statistician will be very, very familiar with the Central Limit Theorem, for example.

Yea of course they will. That doesn't mean they spend their days explicitly enumerating limits of functions, and limits are pre-calc anyway. The point isn't that calculus isn't /integral/ in mathematics, just that it's not _necessarily_ (I did say necessarily) the workhorse of the applied statistician.

No, I'm disagreeing with the statement that you don't need calculus as a statistician. You absolutely do. The expected value is an integral, and is the very first concept in stats / data analysis / whatever you want to call it.

Reply on news.ycombinator.com