If the ‘seeing’ (typically 3 arc sec as you say) is thought of as comprising a uniform disc, and this covers the slit in a symmetric way, then I think the geometry gives (after a little work…) an upper bound on light loss.
I reckon that my f10, 200mm SCT (Celestron 8SE) with angular slit extent of 2.37 arc sec. incurs only 11% loss (I think Robin’s setup is a bit more at 32%). Having tried to get my 0.63 focal reducer to fit with a nominal recommended 110mm distance from reducer lens to sensor and failed (calibration module prevents this) I am therefore inclined to try without.
cheers
Kevin