Well, here’s where the logic gets a little fuzzy 
First of all, what makes a track “good” shouldn’t be confined to the personal beliefs of the producer (or any producer). As you pointed out, some tracks may take off even though a trained ear can pick out several technical weaknesses (sound, harmony, melody, mix, variation etc). Sometimes, or often, a track is selling because it’s relevant and useful in the buyer’s context. While some may adress hitting the “buyer sweet spot” with pure luck, IMO there are ways of researching how tracks are being used and what matters most to the potential buyer. These could be things like intro length, track duration, ending, specific dynamics like crescendos, breaks, etc, but also the general vibe of the track, how it relates to similar popular music at the time, and of course how the track is titled and marketed, when it’s published, what other tracks are competing for the same search terms at the time etc. These are all things that take considerable time to research, and since every (new) track has its own subset of “environmental variables”, putting down that kind of effort may very well be the difference between a track that picks up those important sales early on, and the track that even though it SOUNDS good never catches an audience.
There is also an important snowball effect on AJ where VH items play a big role. Say a popular VH item picks your track for its preview. If the track is great, chances are your following will increase. If the track is only so-so, that never happens. The same can be said with exposure in high end ads, movies and games.
Quantity DOES play a part, I think we all can see it, but it has mainly to do with the way the AJ search engine gives more exposure to new items. One may think that “the more items in the portfolio, the greater the chance of being discovered”, but that is debatable given the low (and dwindling) ratio of a single author output compared to the total influx of new tracks.
In my case, my best sellers are not necessarily the best “sounding” or “most meticulously produced”, but they all share a high level of relevance, usefulness and effectiveness for the targeted market. These are things I cannot completely address to luck alone - when I look at my bottom sellers I can honestly say they mostly lack that “market sparkle” any top seller features.
So, “statistically”, even if that’s a difficult word to use when dealing with subjective musical qualities, I would argue that 10 VERY good tracks definitely would outsell 100 mediocre ones. It’s just not always true for variables like “time spent” or “quality of mix”. Spending your time right - doing the research and nailing the relevant vibe and structure for the targeted buyer - that’s where the real money is.
I said in another thread as well, although no one seemed keen to investigate, that it would be interesting to take a closer look at those “good tracks that never sell” everybody seems to throw around as an argument in case. My theory is that they don’t really exist. Prove me wrong! You can link to an item that most producers go “wow, that sounds great to my ears” and it has zero sales, but I’m betting I could argue as to exactly why that track never found its way into the shopping cart. Most of the times it’s because at the date of approval there were other tracks in the same category/search scope that were BETTER suited for the buyer’s project. Having 100 of those inferior tracks approved in the same week wouldn’t make a big difference, the buyer would still go for the “champion”, the at-all-times-present big seller in the top of the search result list. However, had the track been of noticeably higher relevance/quality, it would have started to steal some sales from the big seller, eventually securing a place on the first page giving steady monthly exposure.
Even more evident is when you manage to find a relatively unpopulated niche, just by comparing the top 5 search results at any given time, the better item is collecting the more sales. But obviously, this debate is fuelled mainly by those who compete in “corporate”, “epic” or “happy” search terms, where the influx of mediocre, similar sounding items is so enormous that any buyer will find it beyond hopeless to find a gem in the landslide. In those cases, then yes, quantity oftentimes beats quality, and it’s really a game of hit or miss even over long periods of time.
I’m invested in this issue mostly because I feel it would be so much better for all of us if this marketplace was shock full of high quality, relevant items, instead of being constantly diluted with fodder. Envato may have had a stroke of genius during launch years, when they first realised the value of having a large number of authors invested in the site. Every approved author were prone to marketing their items externally and driving traffic to the domain. However, it’s debatable whether this strategy is equally efficient at present time, when AJ is a dominant player in the RF audio world, and positioning the brand and retaining existing customers become more important than driving new potential customers into a wall of copy-paste music.
“Hard work” needs to be defined as to include taking care of your market, not simply churning out raw material that gets in the way of your competitors’ exposure. Search result listings are so broad that just about any track can show up on a popular keyword. The more specific the buyer gets (and can get), the more work needs to be done in catering to special requirements. But fine, if an author decides that the game at hand is all about output and volume, go ahead and snatch those random sales. I’m just quite certain it will never add up to compete with the ones doing their homework. You’ll find that every now and then, a random track hits the sky, but if you look at all the equally crafted tracks that never made it, again “statistically” it’s not the best possible strategy.
We all want to believe what we’re doing is the right thing, and we all have our constraints, be it time, production skills or understanding of the market, but I for one hold it to be true that any investment in relevant knowledge and craftsmanship provides you with valuable assets that serve as multipliers of your subsequent returns.
The evidence is in the pudding