I don't think people realize what's behind some of the principles at work here. Price either consolidates or trends, so it exists in two phases. Consolidation you could say is an area where multiple traders are fighting it out, some trading 1 lots others trading 2 .. etc. What the zone represents is where traders are giving up to each other as their stop loss is hit. When a larger volume hit or larger trader hits the zone with a large order, the zone breaks into trend phase. If you randomly place trades without consideration for the zones (WHERE) and "WHEN", your going to run into the ruin phase where your taking the losses for all the traders in the zone.
Its highly critical, to look for 'outliers' and martingale the outliers. Outliers aren't just price points, they are time points also. Martingale in essence is a simulation of the market at the micro price structure, the goal is to be towards the last part of the consolidation zone, before placing a trade. Also you want to be towards the last part of time zone also.
This is where statistics comes into play. Take the length of the average consolidation zone in price on a tic chart, for a derivative create a distribution of varying lengths. If you want to minimize risk and maximize probability. Only place the trade towards the tail length of the consolidation zone. Also same thing with 'WHEN', statistically calculate which time intervals price mean reverts or trends. For some derivatives, those stats have been consistent for multi year lengths. Once you collect that data, play the 100th trader about to place the bet. Your looking for price escape out of zone where and when its mostly likely to happen based on multi year tick data. And if you want to be extra cautious with very low trade rate. Wait for the tail in consolidation zone, and time interval, simulated martingale position, for nth number of losses. Than place a live trade with 1 contract.
The fallacies mentioned above present themselves when you apply random actions. But when you bring to bear, the nature of the derivative, and statistics for that derivative, your essentially modeling its inherent behavior.