You realize that a persons writing style can, and is, used in solving crimes (if relevant to the crime) to reduce the number of suspects.
A persons writing style is akin to a signature and is generally unique from person to person.
I'm aware that it's used in criminal investigations but I'm not writing expert and sure, you could run a statistical analysis on the writing to determine if the styles are different but the analysis you're referring to is actual handwriting where the writing is unique.
Let's face it though, there's going to be no proof more solid than a digitally signed message. It would be irrelevant what his writing style is.
To prove/disprove that it's not the same person based on a small sample of text and a single space between 'blockchain' is as watertight as the current claim. If you've run the analysis on the handwriting samples then sure but otherwise, this is all muddying of the waters.
If he can produce a signed message then your analysis is wrong. If he can't produce the signed message then your analysis is of no consequence because he hasn't proven anything.
Ultimately, a signed message is all we're looking for.
(This isn't an attack on you, just that people are going to debate back and forward endlessly about this issue when he could easily just prove himself in one action.)
You can use plain text analysis too, which I did some research on for a project some time ago. Natural language processing techniques, sentiment and gramatical analysis amongst others are sufficient these days to achieve 85%+ identity certainty on text alone. You do need quite a lot of it from both parties, but a few hundred Satoshi posts and some blogs postings from Mr.Wright would probably be sufficient.
Think about the complexity of the English language, and the myriad of ways to write just one sentence. With just a few hundred sentences, there are billions of combinations of words, punctuation, etc. Too much for a human, easy for a computer.
I'm not saying that my 'blockchain formatting' should be taken as absolute proof, but when you apply that discrepancy to the many others that seem to be apparent, then I personally am strongly biased towards a "NO!"
Of course, you're correct that the only way to be sure it
is Satoshi, is with a digital signature (preferably multiples!), which I think we wont ever see.
Easy to dupe some journalists with technical ramblings and a "fake" signature so that they post an article declaring your sovereignty.