I believe the system you're working with is the familier 1,2,3 pivot breakout where, in a long situation, point 1 is the lowest point then 2 retaces up, then 3 comes back down, but not as low as point 1, you then go long on the breakout of the pivot (point 2). The problem for me is identifying a valid pivot as there are allot of false pivots. You would then stay long for the duration of the move up as long as the pivots formed on the way up are not violated. I suppose you could say the same for determining high and low touch points for a trendline.
Nice thread!