To continue (in the interest of being even more long-winded),
Let’s address the idea of a “teaching grammar”. This is a grammar book, or grammar portion of a more general book, which is supposed to proactively help you acquire grammatical understanding through explanations and exercises.
If you are taking a class, and using a book with dialogues and readings, the textbook may select a sentence from the dialogue like “If I’d gone to the party, I wouldn’t have had a good time” and explain what it means. It may or may not use grammtical terminology, such as “In the ‘if’ part use would + the past particple, and in the other part, use “would have” + the past particple.” Is the term “past participle” confusing? Maybe. But it’s helpful to have a term to describe the verb form in English that is neither present tense form nor past tense form (although in many cases the past participle is the same as the past tense form).
The book may get more technical and try to talk about “first conditional”, and so on. IMO, the less terminology, the better. The main objection I have to many grammar-based materials is that they are written in such a confusing way. In my own study of Japanese, I think I’ve been lucky as most of the materials I’ve used have been like “this” means “that”. " X wo mita koto ga arimasu ka" means “Have you seen X.” Nice and simple. Very helpful. If the book also wants to tell me that “no” and “koto” are nominalizers in japanese, that’s fine. I know what a nominalizer is, so it’s helpful. Or I can just remember it as a pattern.
Then, (to continue with our class and our textbook) you may be asked to make some sentences such as “If I’d lived overseas as a child, I…” and use these as the basis of a conversation where you discuss what would or would not have happened if you had or had not done something in the past. Is this sort of thing completely useless? I don’t think I’d say that. I don’t think the evidence exists to say that this sort of thing is of no benefit. However, there also isn’t conclusive evidence showing this sort of thing helps.
Will you consciously remember how to construct this sort of sentence? Probably not. The next you see or hear that kind of sentence, will you be more likely to understand it? Maybe. Will trying to actively use that type of sentence in a meaningful conversation help you to understand it the next time you encounter it? Maybe. I don’t think we yet know the answer to those questions.
There are plenty of studies showing that traditional grammar-based instruction is not more effective than lots of comprehensible input - even when it comes to taking grammar tests. Students exposed to lots of comprehensible input acquire an intuitive understanding of syntax/etc. and are able to perform as well on grammar tests as students who are primarily instructed in grammar and a translation-based approach to memorizing vocabulary. Given that this seems to be the case, there’s absolutely no reason to have a class structured primarily around “grammar” study, as many unfortuantely still are.
Other books are instructional grammar books where sentence patterns, or morphological changes, are introduced and then students are expected to complete exercises to see if they can correctly construct these sentences and so on. Most people don’t like doing this. Is it helpful? I don’t know. When I began Japanese, I had some books that did this for verbs and adjectives. I did these exercises intensively and memorized the changes. atsui, atsukatta, atsukute, atsukereba, etc. oyogu, oyoida, oyoide, oyoganai, etc. I found that it helped me because when I saw a past tense verb, for example, I was able to figure out quickly what the present tense form probably was, and I could then check it in the dictionary. When I learned a new verb, I was also able to immediately use it correctly in the past tense. My friend who didn’t do this memorizing was unable to do this for a very long time. I didn’t spend years doing this - I just did it intensively over a few days. I personally think it turned out to be a good investment of time, although I agree that I probably would have gradually acquired this knowledge over time if I hadn’t “studied” it in the beginning.
However, can we claim that explanations of syntactical or morphological points have absolutely no benefit? This seems crazy to me. It’s like saying that consulting a dictionary is of no benefit. Of course, there are different opinons on how much we should consult a dictionary, whether we should consult a bilingual or unilingual dictionary, and so on.
Steve, for example, uses his online dictionary a lot - a lot more than most advocates of comprehensible input would recommend. However, he uses an online dictionary which he can check quickly and he doesn’t try to memorize the English equivalent - so although he uses his dictionary a lot, he uses it as a tool to make difficult content comprehensible, and he uses it quickly so he can still expose himself to lots of content.
Others may like consulting a reference grammar more than others - either in their native language or in the target language. I think the jury is out on a lot of this.
So, other than saying that grammar-focused instruction is not more effective than input-based instruction in developing grammatical accuracy, and is less effective in virtually every other way, I don’t think the research is there to say that grammar-based resources such as reference grammars, and possibly even instructional grammar books, provide no benefit to the learner, or have no role to play in language learning.
As Steve says, let’s just do what we like and what we feel is beneficial to us. As long as we remmeber that the majority of our time should involve exposing ourselves to meaningful, comprehensible input in the target language, we’ll probably be fine. Focusing on grammar “rules” at the expensive of input will not get you there. Memorizing “rules” you don’t understand will not help. Memorizing a bilingual dictionary is not a good idea. Trying to directly translate everything is not a good idea. Exposing yourself to lots of incomprehensible input will be of limited benefit.
Grammtical resources, if they do anything, help to make input more comprehensible, just as dictionaries do. Steve doesn’t like consulting grammatical resources very much, but uses his dictionary extensively. I don’t like looking lots of words up in the dictionary, but I certainly use my dictionary. I also consult my reference grammar if I’m curious to understand something. Everybody’s different - although Steve is a much more successful language learner than me 
Yiddish has no simple past and only a past-participle. An interesting feature of a language which I love so much.