Are these the rule or the exception? Are more cases won across industries or are more cases lost? You've just named a few cases that have won without showing that it is the norm. I understand the sometimes that might happen, but you didn't really answer my question at all.
Here is the answer you were looking for...
After doing a little more reading in to the subject, it seems that the win/lose of trademark infringement across industries is entirely dependent on which state in the U.S. you decide to initiate your legal battle and how big your pockets are. In fact, this scenario sounds oddly familiar...
Likelihood of confusion is not necessarily measured by actual consumer confusion, though normally one of the elements, but by a series of criteria Courts have established. A prime example is the test announced by the Ninth Circuit Court of Appeals in AMF, Inc v Sleekcraft Boats, 599 F.2d 341 (C.A.9) 1979. The Court there announced eight specific elements to measure likelihood of confusion:
Strength of the mark
Proximity of the goods
Similarity of the marks
Evidence of actual confusion
Marketing channels used
Type of goods and the degree of care likely to be exercised by the purchaser
Defendant's intent in selecting the mark
Likelihood of expansion of the product lines
http://en.wikipedia.org/wiki/Trademark_infringement
Proving dissimilarity of industries plays in to most of these points. Although, trademark dilution laws in the U.S. were changed in 2006 to be much less specific.
The new law revises the FTDA so that the plaintiff only needs to show the defendant's mark is likely to cause dilution. However, the revision also reduced the Universe of marks falling under its protection, requiring that marks be nationally well known to qualify for protection from dilution.
http://en.wikipedia.org/wiki/Trademark_dilution
In the case of "Polaroid Corp v. Polaroid Electronics Corp", the defendant won the case, even though the marks were clearly similar. The plaintiff lost because the products and services weren't similar. This case was from 1961, it's old and laws have changed since then, but I would debate that you could use this in court as evidence of prior cases.
On the other hand, as you've already said, I can find examples of cases that have gone the completely opposite direction (to the point of ridiculous).
I'm no lawyer, clearly, but it seems to me that the defendant has a fairly strong case, even taking in to account trademark dilution laws have changed since 2006. It's really going to come down to the size of their pockets I think.
I'm going to hand this one to Google IMO.
Oh yeah, and I quoted Wikipedia. It's not the preferred resource to quote, but it serves for my arguments.