Drafting queries is an artform, not a science IMO. What QLH can help you do is draft a query likely to appeal to the widest audience of agents - that means following some constructs that conform to many agents' stated likes and dislikes.
But queries that break these constructs, and bad queries even, land agents all the time. The key is, the more non-conforming or "bad" your query is, the luckier and more talented you need to be.
If your book is genius and polished and perfect, or very high concept, you will have an easier time attracting the attention of an agent, even with a weak or flawed query.
But the less high concept your book, the less exceptional your concept or writing, the harder your query has to work.
So, people will tell you never or always all over QLH - you need to decide:
- is that person's opinion trustworthy, and
- does that "rule" work for you.
When multiple people tell you the same thing, and you have read some of their other critiques, and agreed with those critiques, and it all is backed up by your own research, then you should think hard before ignoring that advice. But when you get conflicting advice from seemingly equally reputable sources, and the specific agent hasn't opined on the subject, then you need to go with your gut instinct and what is best for your book.
And remember, a query's sole purpose is to get enough of the agent's attention to cause the agent to look at the enclosed/pasted in pages, or to ask for some pages. So...forget perfection. Try to draft an effective query, and then tailor it to match the expectations and likes of the specific agent.
So, for example, if Agent A asks that you include a market comparison, do it for Agent A, but you don't have to do so for all the other agents. And if Agent B posts what she thinks is a fantastic query, you might want to look at her comments and why and use it as a model for yours. And if she loves rhetorical questions, use them for her, but know that many other agents hate them.
In QLH, there are some very experienced critiquers, and some completely inept critiquers. You'll have to figure out which comments work for you, and which seem less helpful. And you'll ultimately need to draft a query that sells your book.
Mine was far less voicey than I often push other people to be in their query. I made the calculated decision to go with a less voicey, but more clear and straightforward query, and hope the pages would hook an agent. That risk worked for me. Might not work for someone else.
Ultimately, all the query has to do is get the agent's attention enough to cause them to read the pages - that's it. So, even flawed queries work sometimes (maybe even a lot of the time) - but it still seems, to me, to be best to try to draft the most effective query you can.
And I'll second IceCreamEmpress - go read a lot of agent blogs for how they read queries - in addition to the ones she listed, I'd suggest checking out Queryshark's blog (agent Janet Reid) and there are also a lot of other agents blogging and discussing queries online. So, check out the agents you are interested in; see if they have blogs on which they have critiqued queries, or if they have critiqued queries in any online contests, etc.
good luck.
~suki