Originally Posted by
squark
If it's true that you're as likely to mistype as your quasi as you are your mirror, and not as likely to mistype as your contrary, you could use that to help determine mirrors. IOW, if deciding between EIE and IEI, you could look at the plausibility of misytping as IEE vs EII. If it's the former, then EIE is more likely, and if it's the latter, then IEI is more likely.
It makes sense, but there might be cases where a contrary typing does seem more likely, such as strong and obvious rationality/irrationality. But, in those cases, you're not going to be looking at your mirror as a choice to begin with.