Are common typing errors ever considered as a metric or design tool when evaluating new keyboard layouts? Same-finger, alternating-hand, finger travel and so on are all commonly used, but I wonder if studying the frequency and types of errors would be of use? I guess it is a lot more subjective but still of interest. For instance, I read that colemak has a lot of same-hand, while dvorak has a lot of alternating hand. How is this related to a transposition typo (getting two letters in the wrong order).. I am willing to bet that there is a correlation. My guess is that a specific layout influences the kinds of typos produced, since our hands are likely to fail or be weak in certain motions. If this is taken into account, we could get keyboard layouts which are easier to be more accurate on, if this is allowed to influence the letter placement. Common diagraphs should be moved away from positions which are error-prone. Anyone know if this kind of research has already been carried out?