Oops, no, it was my mistake. In some of the tests I had put the actual and expected solutions the wrong way around - thanks for pointing this out!

The random test cases in Javascript expect results such as "MMMMMMMMCDLXIII" despite the description claiming that "there can't be more than 3 identical symbols in a row."

Nice kata. Have you seen this guide for writing your own kata? It's really useful, particularly for tips on writing tests. I think it's really handy being able to see which inputs caused a test to fail, and you could make some random tests too.

Your solution doesn't work if there are more than 2 of a given number in the original array. As explained in the kata's description each sub-array should contain all of the duplicates of a particular number. Which is why group([15,10,15,14,13,19,18,13,13,10]) should result in [[15, 15], [10, 10], [14], [13, 13, 13], [19], [18]], not [[15,15], [10,10], [14], [13,13], [19], [18], [13]]. I hope this is clear, let me know if not!

Woops, done it now sorry dude

Thanks for your feedback. I'm afraid I don't understand what you mean by "if the number of takes greater than 2 (including them in one array)", could you explain?

Lovely, thanks :)

Thanks :) Just one more question - when I run it, at the end there's a passed test that just says "Warning". I can't work out what this is or where it's coming from - could you explain?

Thanks for your translation :) Is there a reason the example test cases don't provide information about what they're testing on? Seems so handy to have