Steve Fenton

The cost of juggling

On the 2nd of September 2017, in Middletown USA, Patrick Ponce achieved the world record for solving a Rubik’s Cube by twisting that damn thing into the correct position in 4.69 seconds. He got a certificate for it from Guinness World Records Limited. Rafael Solano commented on this saying:

This guy is a great inspiration and I hope to be somewhat just as fast one day.

I’m sure we all feel the same.

Now, I can’t be sure about this (unless you’re reading Patrick, in which case please get in touch to correct my numbers) – but I would estimate that Patrick could solve three Rubik’s Cubes, one after another, in about 15 seconds. Perhaps 20 seconds if he needed to check his watch and grab some popcorn.


“But how long would it take Patrick to solve three cubes while juggling them?” I hear you ask. As far as I know, he hasn’t done this yet; so we’ll need to take a look at 12-year-old Que Jianyu. He juggled three maddening 3D combination puzzles while solving them, and completed all three in just five minutes and six seconds. CNET Magazine suggested that this achievement also has the rubber stamp of the Guinness World Record organisation, although I couldn’t validate this by finding a certificate (this all breaking news at this stage, so I’ll add a link when it gets published).

But wait… it takes 00:05:06.000 to solve three annoying puzzle lumps while juggling, but (probably) only 00:00:20.000 to solve them if we don’t juggle them. Seriously, it takes 15 times longer to do something if you add juggling to the process. And remember, we’re counting a popcorn break in our 20 second time!

I’d love to know how long it takes Que Jianyu to solve all three without juggling. Message me!

Less juggling

“But not all tasks are as complicated as a Rubik’s Cube!” I hear you exclaim. And I anticipated just such a statement. Or more honestly, while I was researching Que Jianyu I found a certificate for Teo Kai Xiang from Singapore, who solved one irritating cuboid matrix while juggling it along with two juggling balls. This somewhat simpler version of the task was completed in 22.25 seconds.

We need a table!

Rubik’s Cubes and Juggling
Amazing Feat Amount of Juggling Time
Solving 1 Rubik’s Cube None 00:00:04.690
Solving 3 Rubik’s Cubes None (possibly popcorn) 00:00:20:000
Solving 1 Rubik’s Cube Cube + Two Balls 00:00:22.250
Solving 3 Rubik’s Cubes 3 Cubes 00:05:06.000

Or maybe a graph!

Rubik's Cube Times

We didn’t really need a table or a graph. What we needed was to simply acknowledge that you can solve three infuriating puzzle blobs sequentially faster than you can solve a single one when you’re juggling.

Not just speed

Let’s not forget that there is more to life than flat out speed. If you review the video footage, you’ll see that the juggled version doesn’t result in a cube that is neatly aligned. The juggled version is messy. It took longer… and it’s not such a great outcome as it would have been if it was done one-at-a-time.

Que Jianyu

Let me be fair here, I have solved one Rubik’s Cube… ever. If it was being timed, the stopwatch battery would have died before I solved it. But I had to solve it because my five-year-old daughter was devastated that it was muddled up. If I could routinely solve those tiresome six-sided devils, I would still be impressed that someone could do it while juggling. I’m not saying that I’m not impressed; I’m saying that it would be done better if it was done without all the juggling.

So what?

Maybe you are juggling stuff.

In software development we call it work in process and it can often be as complex as solving an irksome puzzle block – sometimes even harder because the goal isn’t always as clear as “make each face a single colour”. This means we need to pay more attention to juggling, because it makes no sense economically and because quality suffers.

I say this having done it. It isn’t merely a dream; it can be done. I was working on a team that gradually stripped away juggling. We started by limiting the number of concurrent projects. When this was 1, we limited the number of concurrent features. When this was 1, we limited the number of concurrent acceptance criteria until it too was just 1 (idea credit: Neil Killick).

We nailed six months with zero defects and with no confusion between the clinically-trained product owner and the clinically-awesome (but otherwise not clinically-trained) technical team. Our feature cycle times went from weeks, to days, to literally less than 3 hours per feature – at the same time as our quality hitting the jackpot.

It can be done; but it starts by reducing the juggling.

Written by Steve Fenton on