Motivation
For desktop, we tested the new paragraph results with a random sample of 300+ articles (see here). Let's do the same for mobile, just to be super sure

Task

• Test with 300+ random articles what the changes to the mobile diff means to the diffing behavior.
• Add a summary about the results to the test page,just like in the desktop version
• Allow to access the test page with a permanent link, so it can be shown to others

Notes

• We can reuse the same script, and the same random articles
• We can also reuse the CSS to better show the changes (T193638)