Well nothing concrete to contribute but cf computers 10 d4 is actually Junior's first thought! 

It does then seem to relegate it to second place but seems to float around -1 or so. Which probably says rather more about the way its programmed than anything else

Watson does give this a very brief mention in mastering the chess openings volume 3 but only to say that the position after 12 Bg5 is 'fantastically complicated'.

I happened to come across an old John Watson review on TWiC http://www.chess.co.uk/twic/jwatsonbkrev34.html where he discusses a gambit idea of his, where white is giving up a considerable amount of material to build up an attack. So much material in fact, that it probably still confuses the engines today, more than ten years later. 

But still, +ten years is a very long time in engine evolution, so it would be interesting to see if we the chesspubers could finally refute or validate Watson's 10.d4!? 

I checked my database, but couldn't find a single game with 10.d4, despite that it was also mentioned in Kosten's The Dynamic English - maybe because: no one dares enter it, it's already refuted, or maybe it's been made redundant by some better option earlier? 

Anyway, it could be some harmless summer fun to see what can be found. Here's a pgn of JW's TWiC lines and the few similar tries.