The earlier observations were too imprecise and error-filled

Do you have a quantitative statement of the error bars on normal (ie, not Brahe) astronomical data? I have heard it claimed that Ptolemy's data is systematically biased in favor of his model, which is only meaningful if the error bars are small enough that such data could distinguish models. (though there's also the issue of whether anyone understood error bars)