I just read, "Is guessing from context a load of XXXXXX?" and thought I'd respond.
A couple of things...First, a reader does not need to know 100% of the words to understand the meaning of a particular text. So, guessing the meaning from context will depend on the familiarity the reader has with the language (i.e., English proficiency) and the percentage of words the reader does not understand. If it's true that the most 50 common words used in English are functional (as opposed to lexical), and that 45% of all written text are made up of these 50 functional words, then this seems to indicate that lexical words would be a bigger issue when it comes to word recognition. So gaining the meaning from context would only be beneficial if the reader were to understand a lexical word she did not understand before AND that in doing so would enabled her to understand the overall text. Put another way, I see little importance in trying to determine whether a reader was able to get the meaning of a particular word or not from context if not doing so still enabled her to understand the text. "Did the reader understand the overall meaning of the text?" becomes more important that "Did the reader understand 100% of the words in the text?"
Second, I think the blog post (Is guessing from context...) doesn't fairly present the issue if the only examples are at a sentence level. The act of guessing meaning from context is at a discourse level (i.e., beyond the sentence level). There are many types of references (e.g., anophoric, cataphoric, and exophoric references) at a discourse level that help tie ideas together in a given text. Cohesive devices such as reiteration, substitution, ellipsis, conjunctions, etc. provide a means for making a text more cohesive, which incidentally can be a reading strategy that readers can incorporate when "guessing" the meaning from context. But all of these examples of cohesive devices only work at a discourse level and not at the sentence level.
In summary, guessing the meaning from context only becomes an issue when the level of unknown lexical words begin to interfere with the overall meaning of the text. I would argue that most readers (in L1 or L2) fail to understand 100% of most texts, but this does not always interfere with the enjoyment of learning something new. Additionally, guessing the meaning from context occurs at a discourse level and not at the sentence level; if readers begin learning the cohesive devices writers typically use anyway, this will go a long way in taking the guess work out of tackling an unknown lexicon.