Averages are weighted only by the statistics of the point. Longer observations will develop more realistic statistics as the individual positions (epochs) drift over time and give a better indication of the precision of the shot. With a short observation (for example, Quick Topo is only 2 epochs long), there aren't enough epochs to give a realistic error estimate, so it is likely that it will give have a higher weight in the average even though it's more likely in reality to have a better precision than a longer observation. It's best to only average points that have similar duration of observation, particularly the number of epochs in the solution.