I'm sorry to disagree with you, but these days, according to Document Type Declaration tags, HTML5 is the most common version of HTML in use. Also, what browsers are you referring to? Since Chrome has a roughly 80% market share, and all the big players support HTML5, I'd argue that smaller browsers that still aren't capable of parsing HTML5 aren't common enough to count.
HTML4.01 is down to just 2% of web pages, and HTML5 taking 80% now. Granted, my source only took a sample of around 20 million pages, which is just a drop in the ocean, but I feel that is fairly representative.
This post is aimed at people with no technical background, and I've taken some liberties in order to keep it as accessible as possible, it wasn't meant as a technical discussion of the various flavours of HTML i.e. SGML or XML-based.
Sources:
W3Schools - Browser Statistics (https://www.w3schools.com/browsers/default.asp)
HTML Version Statistics (https://try.powermapper.com/stats/htmlversions)
Edit: Oh and yeah you're right about not all web pages being a HTML document, however even the alternatives will still be a HTML document at their core. Again, I'm just trying to keep things simple.
The problem with your assertion is that a lot of pages don't have DTD tags and even if they have, they don't validate as valid HTML4 or HTML5... A lot of Asian websites still require old Internet Explorer due to use of ActiveX controls which haven't worked since Internet Explorer 9 and hasn't worked with Chrome or Firefox at all without extensions.
I accept that there's a small minority of sites that are basically legacy sites, however I still feel my original statement is accurate. Most people reading this post won't come into contact with the type of sites you mention, I don't have a large Asian audience.