Ask HN: How to develop logical skills? | Hacker News Reader

Ask HN: How to develop logical skills?

30 pointsoscardelben21 comments

For what is worth, perhaps very little, I would like to develop my logical skills. I'm referring to the skills needed to complete sequences of numbers, images, etc. The reason I want to develop them is because I really enjoy solving them in the first place. I wonder if someone know some good resources for practicing or he/she has some advice.

Edit: sorry if the question is confusing. What I'm referring to is the ability to solve logical puzzles, like where you have a sequence of images and you need to choose which one comes next. Example from searching on google: http://img43.imageshack.us/f/testdk.jpg/

View on HN

I don't think logic means what you think it does.

Logic is about using deductions to reason about a thing. For example, in sudoku, "If I write a 5 in this cell then the column will contain two 5s. Therefore, this cell can't contain a 5."

Completing sequences of numbers like:

Q: What comes next? 2,3,5,7,11,...

A: 13, it's a list of primes

isn't about logical deduction at all. If anything it's about inference, but I'm probably classify it under pattern matching or something similar. In some sense these problems are meaningless. You could also answer the previous question with

A: -302, the list is the sequence {2,3,5,7,11,-302,5903,pi,....}

and no one can argue with your claim beyond "That's not what I/the author had in mind."

To get better at such things you just need to collect examples. Maybe look at http://oeis.org/Seis.html

I think you may be referring to the "rule following" paradox http://en.wikipedia.org/wiki/Wittgenstein_on_Rules_and_Priva...

http://projecteuler.net/

If you don't understand how to solve a problem, there's always Google to "cheat". The discussions are pretty often very enlightening, and have some wild solutions.

In more abstract means... number theory is handy, and I don't know what else you'd be interested in, the question is fairly vague. Read lots of puzzle books?

I second that. Also you may find this helpful http://projecteureka.org

Project Euler is great as a programming problem when trying to improve your solutions for speed.

Also, as someone who has not touched fundamental maths in 5 years it was also a good refresher; having to think about prime numbers etc again.

A good tip I have found is to use a language that can nicely manipulate longs saving you much overhead, I chose Python because it also has some nice maths libraries.

The best tip I've been given for practice is to actually sit with a piece of paper and write out your thought process whilst solving these little problems.

It's feels a bit laborious at first, but I find it makes for much more effective practice than getting to the answer in a roundabout way, and not really knowing how you get there.

As in hackerblues' comment, sitting and explicitly writing: "If I write a 5 in this cell then the column will contain two 5s. Therefore, this cell can't contain a 5." is more effective than bumbling around and just getting it right, particularly when practicing with simple examples.

Do real math. Not this fake logic puzzle BS.

I second the idea that learning real math will help, but there is a purpose to using "logic puzzles" also - math doesn't help much with learning to recognize various "difficulties" - cognitive biases, like framing issues, and common fallacies, like ad hominems. Besides practicing logic puzzles, you need a good overview book, Robyn Dawes, Rational Choice in an Uncertain World, is the single best volume I have found.

Are we talking about the same kinds of "logic puzzles"?

I'm referring to e.g. stuff in Mensa books.

Get one of those lateral thinking puzzle books (there are probably 100s of websites online as well).

You could also take practice LSATs to get your fix.

I felt I needed to give more than +1 to support this comment. Studying for the LSAT's is an extremely effective way to increase your logical skills. For whatever you think of lawyers and the legal profession, being able to "think like a lawyer" is an incredible asset.

I thought I had a strong ability to recognize logical patterns and avoid making categorical assumptions until I took the LSATs. Thankfully, I am not a lawyer.

Learn to follow Metamath proofs. (http://us.metamath.org/) Metamath allows you to start with the axioms of propositional calculus and go all the way up to set theory and beyond. When you understand Carew Meredith's sole axiom for propositional calculus, you will know a lot about logic.

If you are interested in learning more actual logic, and scientific reasoning, I suggest Bad Science by Ben Goldacre, a book about pseudoscience. The author explains some flaws in some bad science in popular culture, and by learning from him you can learn how to critique bad logic.

I can't recommend H. W. B. Joseph's _An Introduction to Logic_ highly enough. Once you master basic logic, Fred Sommers' Term Functor Logic is exceptionally helpful. http://www.ontology.co/sommersf.htm

http://www.amazon.com/How-Solve-Heuristics-Zbigniew-Michalew...

That book will help you on your journey.

I second this.

This book is the textbook for my 'Heuristic Problem Solving' course @ NYU this semester. I'm normally not one to touch my textbooks, but over Thanksgiving break I decided to crack it open - suddenly I was 200 pages deep into it.

(Check out what we do in the course at http://cs.nyu.edu/courses/fall10/G22.2965-001/index.html )

Study computer science / mathematics? Solve riddles?

Look into abstract algebra. Often those types of problems need to be solved by forgetting any assumptions you make and focusing solely on the information given. Being able to work in an abstract ring or group requires that ability.

I guess it's same question as "How can I learn to play guitar?"

engage in purposeful discourse on something you disagree with or know little about.

Look at LSAT prep materials.

Reply on news.ycombinator.com