I think Bruckner is more problematic than many music scholars would care to admit. I don't care for the first few symphonies (including the "zero" and "double zero"), but I do like the scherzo from the first symphony. The problem with the third symphony, is that there are three distinct different versions.
To my ears, the first version of B3 is way too long (rambles a lot, like me), while the final version is too truncated. I like the middle version, which is sometimes referred to as the Oeser Edition. B4 used to be wildly popular, but it's problematic as well. The slow movement to B4 is kind of a dog, and the finale really isn't very interesting either, until you get to the coda. Of course, B4 has that marvellous "hunting horns" scherzo. However, that was the second scherzo that Bruckner composed for his 4th, and it really has very little to do with the rest of the symphony, thematically speaking. In some ways, I think that the first version is a truer picture of what Bruckner really had in mind with this symphony. Eugen Jochum said that the 4th should be subtitled, "the mystical", and not "the romantic".
I think that Bruckner's 5th is a marvellous symphony, with a very fine Adagio and an outstanding finale. However, I can sure see how most listeners could lose patience with it. There's a lot of mechanical, J.S. Bach-like, grinding out of short thematic material (cells, really), which appear to be going no place. However, they do, infact, arrive upon huge musical vistas. It has, by far, the greatest ending of any of his symphonies - going back to the start of the big brass chorales. I also think that the ending of Mahler's first sounds a bit like the ending to B5.
For the first three movements, the sixth is easily Bruckner's most progressive symphony - both harmonically and rhythmically. However, the sixth is somewhat let down by its finale. Jochum stated that he often times put cuts in the finale; in places where he thought that Bruckner's improvising skills on the organ, got the best of him.
For me, the seventh may be the least flawed of all his completed symphonies. The first two movements are melodically rich, while the scherzo is like a kaleidoscopic rehash of Wagner's "Ride Of The Valkaries". Again, the finale is a bit of a let down - in spots - but has a terrific coda to end it. I really like the 7th.
Most Bruckner buffs will tell you that the 8th is his greatest, but I find it a bit problematic as well. Again, there are two distinctly different versions of it. Most people will say that the first version is too flawed, musically speaking. But I think it's actually a far truer representation of where Bruckner's head was at, when he began composing this piece. It's far more raw, fiery, unfiltered, youthful, and optimistic sounding. It seems to relate much more to Bruckner's more youthful composing. The revised version is a filtered and highly polished product. But it's also a bit more elephantine, staid, marmoreal, and downright old sounding. It was as though Bruckner had been forced to grow up, after Hanslick (or whoever it was) had rejected his first version. But I think that Bruckner, to some degree, overshot his mark. I like it, but I find it hard to love. I prefer the first version, especially in the Arte Nova recording conducted by Dennis Russell Davies.
For me, the 9th - without the conjectural finale - is one of the greatest orchestral compositions ever! In fact, I prefer it to the Mahler 9th (and conversely, I feel that M10 had/has the potential to be an even greater work). Die-hard Bruckner scholars will try to sell you on the necessity of including a finale. I think that's a fine idea, in theory. In reality - to my ears, anyway - what pages do exist, sound more like the beginning of a 10th symphony than a finale to the 9th. Those pages are harmonically far more wild; as though Bruckner suddenly realized that he had to get more in step with the times (or try to anticipate them, a bit). They're interesting, but I just don't find the argument convincing. Anyway, one can decide for themselves by getting any of the three recordings that include a conjectural finale (Tintner/Naxos, Talmi/Chandos, and Inbal/Teldec), or picking up Harnoncourt's excellent Vienna Phil. recording, where he chooses to do a lecture/demonstation on the finale - something that I feel is much more helpful.
By the way, I think you guys made excellent observations! I'd like to read more thoughts on Bruckner. Why does he seem to be slipping out of sight, to some degree?