In my experience, ratings are an inaccurate measurement. If you want to learn how to cook Pad Thai, are your going to trust the 5-star recipe with 30 responses, or the 3.9-star recipe with 3,100 responses??

An active comment section tends to be a step above reviews, because at least you can see if other people find the recipe too spicy, or boring, or an exciting base for other ingredients. And no, 2 comments that say "I loved this recipe!" and "Thanks!" isn't sufficient. This is usually how you can tell a recipe on an aggregate of authors like FoodNetwork, AllRecipes, or NYTimes is legit.

In my opinion, one step above an active comments section is following individual recipe writers that you jive with. An individual writer will usually have a measurement of success that you can agree with (health-conscious, budget-friendly, unique flavors, wide appeal), so you can understand their motivation better. Also, they have some credibility on the line. When you find writers that you agree with, you may even find their "filler" text to have some substance (another reason a "poseur" would feel obliged to include vapid copy before their recipe).

I guess that's what ratings are for. Who knows if it'll taste good to your taste buds, but if many others like it that could be a good indication.

Honestly, I tend to rely on the "brand" of where I'm getting the recipe. Ratings are only useful if you trust the site to have good ratings (either an active audience with similar taste or editors filtering and publishing with similar taste). I haven't seen it much, but I also worry about food safety for random recipes.