Recent variationist research indicates that grammatical intra-speaker variation (or: optionality) is unproblematic in speech production. Optionality contexts do not coincide with dysfluencies. In this study, we ask a theoretically significant follow-up question about optionality contexts that strongly cue grammatical variant choice: are optionality contexts in which all variants are probabilistically equally likely more problematic than those that strongly cue choice of grammatical variant? After all, unbiased (un-cued or freer) choices are sometimes theorized as being more difficult than biased (cued) ones. We empirically analyzed a subset of the SWITCHBOARD corpus of spoken American English on a turn-by-turn basis. The dataset covers 7,295 conversational turns containing 7,001 optionality contexts (spread over 20 grammatical alternation types), 2,970 filled pauses, and 41,297 unfilled pauses. Contrary to claims in the literature, weak probabilistic cueing does not trigger more production difficulties than strong probabilistic cueing. Unpredictable grammatical choices are not harder than predictable grammatical choices.