Understanding the Power of the Puzzle in Innovation

By Oded Margalit

Oded Margalit, Computer Scientist, IBM Research – Haifa
Oded Margalit, Computer Scientist, IBM Research – Haifa

By day I’m a mathematician working in cyber security at IBM Research-Haifa. But in my “copious” spare time, I’m also the puzzle master for IBM’s Ponder This puzzle, a position I’ve enjoyed since 2009.

The original puzzle master, Don Coppersmith, started the monthly challenge in May 1998 as part of an IBM booklet named Changing the World that challenged inventors to ponder a geometric puzzle when they were stuck on a problem at work.

Solving “Ponder This” puzzles is about being a part of a community of people curious about how to solve problems. Our “solvers” have backgrounds ranging from vice presidents of corporations to PhD students, analysts, post docs, and even my fellow IBM researchers. Some are also high school students and their teachers or parents. Some columnists and research institutes also link to our puzzles (like page 15 of the Mathematical Sciences Research Institute publication.)

We’ve even had a puzzle mentioned as one of the hardest puzzles on a popular puzzle blog. But the point is to test your curiosity about how things work. I personally know and have experienced the carry over of that feeling of accomplishment into my work.

SP-ponderthisFor example, a code checking idea our team recently developed started from playing peg solitaire. The point of the game is to have one peg left in the exact middle – you won’t win if that last peg is too far to one side. So, we tried to show that this game was impossible to win in certain situations, and even applied an IBM model simulation checker to solve the problem.

It didn’t only solve the game, but we discovered that it can be applied as an elegant code and bug checking software. Now, we hope to use it to solve real client problems as a result of this, and are publishing a paper at our Haifa Verification Conference this November (authors: Gadi Alexandrovitch, Alexander Ivri, and our intern Dan Rasin).

Ponder This puzzles don’t require a PhD in math, In fact, April 2013’s puzzle was so easy, our regular solvers thought it was a joke (incidentally, we do release the puzzles on the first of every month). Others, like February 2009 was so difficult, that only one person solved it within a month – and even after extending the deadline for another month and giving several hints only seven more people solved it!

More than 11,000 correct solutions to Ponder This puzzles have been submitted over the years. Maybe we’ll publish a book of them, or even host a conference for our regular solvers someday, but for now, I’m happy with its current format.

I enjoy the challenge of solving problems – puzzles or not – that are a bit difficult. When you look at life from this angle, everything turns into a puzzle.

Want to try a Ponder This puzzle? There’s still time to solve the September’s challenge: Let b = (2^3^4^5) / (e^n), otherwise stated as “two to the power of (three to the power of (4 to the power of 5)) over e to the nth power”, where n is an integer such that 1 < b < e.

Find b, with an accuracy of 10 decimal digits.

I had a nice way to solve it, but, alas, brute force solves it too – so I’m awarding an asterisk to those who find the elegant solution. Can you find it?

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.