Rollover spreads and the false sweep they create

[Shadow Trader]

Around the daily rollover, spreads on many brokers widen dramatically. On a bid based chart, that can produce a spike that looks like a sweep of a level, with a long wick through a previous low or high, followed by an immediate return. In reality, it may just be the price feed reflecting a temporarily wide spread rather than genuine order flow. Traders who see the wick and enter a reversal often get stopped or filled badly. My rule is simple: I do not mark sweeps that occur during the rollover window, and I do not trade the first thirty minutes after it. I also verify with a second price source if the wick looks suspicious. If the same wick does not appear on a different broker or feed, it was likely a spread artifact. This matters for backtesting as well, because historical data may include these spikes and inflate the number of apparent sweeps. Clean your sample by excluding the rollover period, and you may find that your pattern frequency and results change noticeably.

[FXS HongKong Scalper]

Rollover wicks are some of the most misleading candles on a bid chart. From Hong Kong, rollover falls in the early morning, which is also when the Asian session starts to pick up, so it is easy to mistake a spread wick for the first move of the day. Checking a second broker's chart, as you mention, or looking at the spread indicator, quickly shows whether the wick was real.

[Amanhasnoname]

This caught me out for a while in my backtests. I was testing a sweep reversal on GBPUSD using broker data, and the strategy looked unusually good on late evening entries. When I went through them one by one, a large share of the best trades came from rollover wicks that were several pips beyond the level, and on a second data source half of those wicks didn't exist at all.

After that I excluded everything from ten minutes before rollover to thirty minutes after, and the number of trades dropped by roughly fifteen percent. The results got worse, but they were finally honest, and the live trading since then has been much closer to the test.

For live charts, I keep a spread indicator on the screen that shows the current spread in pips in the corner. If a wick appears and the spread was five or six pips at that moment, I don't count it as a sweep. It's a simple check, but before I added it I'd marked several false liquidity grabs on my charts and planned trades around them the next morning.