The need for some form of explicit grammar instruction is recognised in most current approaches to second language teaching. Since the usefulness of explicit instruction is at least to some extent dependent on the difficulty of the rules that are taught, an important question for teachers is whether their judgements of rule difficulty are reliable. This study investigates the question of whether there is a significant relationship between the teachers' assessment of rule difficulty and the learners' ability to produce controlled output based on explicit rules. In the study, 25 Polish teachers of English were asked to evaluate the difficulty of 12 pedagogical rules of English grammar. Subsequently, 50 Polish learners of English were asked to produce example sentences based on the same rules. To test the null hypothesis of no significant relationship, a Spearman rank-order correlation coefficient was calculated (r[subscript s] = -0.9). The results were found to be significant (p less than 0.01 for a two-tailed test). They are discussed in terms of what they mean for the process of selecting rules for explicit instruction. (Contains 6 tables.)