forex-scalping.com

Bollinger Bands narrow when volatility falls. The "squeeze" describes a period of unusually tight bands. The common belief is that a squeeze precedes a large move.

It does, on average, since volatility clusters and low volatility tends to give way to high volatility. What the squeeze does not reveal is direction.

Traders often guess the direction, enter early and get stopped by a fake move before the true breakout. A more careful approach waits for a candle to close outside the band and enter on confirmation, accepting a worse price in exchange for fewer false starts.

Bandwidth, the distance between the bands relative to the average, can be measured numerically, allowing a rule such as "squeeze when bandwidth is in the lowest 10 percent of the last 100 candles".

Combining the squeeze with a time filter, such as the minutes before a session opens, aligns the setup with a plausible catalyst.

Has the squeeze been tested here with a numeric definition, and what did the numbers show?

I tested a numeric squeeze on AUDUSD five minute charts, using bandwidth in the lowest tenth of the previous hundred candles as the trigger, much like your example. The volatility part held up well. After a squeeze, the next hour's range was clearly bigger than average. Direction was a coin flip, and entering on the first close outside the band gave a lot of false starts, especially in the late Asian session.

Two things improved it for me. The first was the time filter. I only counted squeezes that ended within the twenty minutes before the London open or before a scheduled release, which cut the number of signals by more than half but removed most of the dead ones. The second was waiting for a retest of the band after the breakout candle rather than chasing the close.

Without those filters the system lost slightly after costs. With them it was modestly positive over about two hundred trades, which is not a lot of evidence, so I still trade it small.