Averaging, is generally frowned upon. It is used to suggest that initial entry wasnt as optimal as it could have been, hence the need to average. (although averaging into a winning position is accepted)
the losing position will blow the account if the number of positions are too great. so maybe limit the number of positions, or ensure they are wide apart.
this causes dissonance.
the market starts to move away from your personal position, and rather than accept that the sup/res range has broken, you end up adding to an existing losing position.
if you have 4 positions, and you're wrong, and you add a new averaging position, you now have 5 positions, and the market is going against you (faster).
the essence here is to know the point where you think you have failed, and to exit the losses.
unless, of course, if you accept the position is a losing one, 4 positions for example, you decide to offset it with a 5 positions in the new direction. here, you're just "fixing" your 4 position loss, and have a nett exposure of 1 position in the new direction.
the grid needs a point where you accept the grid has failed. thats the principle that tends not to be considered, and causes exponential failures.
(you can see I have thought about this!)