I'm excited to see this. It reminds me of a paper by Cameron Browne (http://ccg.doc.gold.ac.uk/wp-content/uploads/2016/03/browne_...), which does similarly, but using less advanced technology. While the approach was janky in some ways, the game that the system invented, Yavalath, has received some praise for being an interesting and compelling take on the n-in-a-row genre.

I've been dreaming of doing similar with an alpha-type system, and so I look forward to reading this paper.

Edit: After skimming it, it doesn't look like they're going the distance like Browne did. Browne had the computer evolve games that the AI would rate ideally, while the DeepMind paper only has the computer consider a handful of pre-selected variants. I don't anticipate it would have been much extra work to evolve games, and I would have found it much more interesting.

Interesting paper concept. I'll read it.

I also notice Kramnik is the last-listed author in a group of four, the rest of whom work at DeepMind. Is this a new marketing "thing" in published corporate research?

You may already be aware of this, but for those who don't: depending on the discipline and other factors, the last author in the list may be known as the "anchor". Often, frankly, the preceding authors do the majority of the work, while the anchor functions as a supervisor.

I'm ignoring "vanity" authorship and other horrors.

Did you read the paper? It is quite long, mainly because of Kramnik's manual analysis of some selected games. It appears his role was to explore game playing factors, which are hard to catch with statistics.