forex-scalping.com

Buy stop above and sell stop below a 10 pip range: what it costs when both sides fire

EURUSD has been boxed between 1.0850 and 1.0860 for forty minutes and a data release is due in five. The tempting setup is a bracket: a buy stop at 1.0862 and a sell stop at 1.0848, each with an 8 pip stop and a 12 pip target. Whichever side goes first, you're in.

MT5 doesn't link the two orders. If the buy fills, the sell stop stays live unless you cancel it or an EA does it for you.

Now the ugly sequence. The release lands, price spikes through 1.0862 and you're long. It reverses and takes the stop at 1.0854 for -8 pips. It keeps falling and triggers the sell stop at 1.0848, so you're short. It bounces and stops you at 1.0856 for another -8. That's -16 pips from a single event.

On a $10,000 account risking 1% a leg with an 8 pip stop, the lot size is 1.25, since each pip is $12.50 and $100 divided by 8 pips is 1.25 lots. Two stops cost $200, which is 2% on one news bar. The clean sequence pays +12 pips, or $150, once.

Break-even for the unlinked version: if a share p of brackets run clean and the rest whipsaw for -16, then 12p = 16(1-p), so p is about 57%. If the second order is cancelled when the first fills, a whipsaw costs only -8, and the bar drops to 12p = 8(1-p), which is 40%.

These are my own sums, not journal results, and they ignore the spread and any slippage on the fills.

Fairman, the bracket cost when both sides fire is a good thing to work through before a release. In a spike, the second stop can easily fill with slippage on top, so the total loss of a double trigger can be well beyond the two stop distances on paper. Testing the bracket on past releases with realistic slippage is the only way to know if it works. A dry run on old releases, counting how often both legs trigger, would show the true cost quickly.

Slippage is what really breaks the bracket sums. A stop order fills at the next available price once it's triggered, and in a release spike that can be several pips past the trigger. Say the buy stop at 1.0862 fills at 1.0868. Your protective stop is still 1.0854, so the risk is now 14 pips instead of 8. On 1.25 lots that's $175 when it stops out, against the $100 you planned. The target at 1.0874 is only 6 pips away from the fill. So the reward to risk, planned at 1.5, becomes 0.43 on that trade. A bracket that works at 8 and 12 on a quiet chart can have a different shape on the one bar it was built for.

An expiry on the pending orders fixes the orphan problem. In MT5 you can set an expiration time on each pending order, so if the release is at 13:30 you can have both orders die at 13:35. Otherwise, an unfilled sell stop can sit there for two hours and trigger on a drift you never planned for. Some people instead wait for the first M1 candle to close after the release and enter at market. That gives up perhaps 3 to 5 pips of the move, but you've seen which way it went, and the entry price is at least a real price rather than a slipped stop.

How many events do you need before the bracket's 57% bar means anything? With 20 past releases, a measured hit rate of 12 out of 20 is 60%. The 95% range around a 57% rate on 20 samples is roughly plus or minus 22 points, so 35% to 79% fits the same data. Even 40 releases leaves about 15 points either way. And news events differ: a CPI bar isn't a jobs bar. If you want to test it, mark the first break and whether it held ten minutes, and keep the pair and event type in separate columns. Expect to need more events than you have patience for.

Halving the size of each leg doesn't change the break-even rate at all, since both outcomes scale together. What it changes is the worst case. At 0.5% per leg, the double stop-out is a 1% hit, and the clean run pays +$75, which is 0.75%. If you also have a daily loss limit of 2% or 3%, that difference decides whether one bad bar leaves you any room for the afternoon. The same logic holds for any trade where two orders can both fill: add the two risks together before you decide the size of each, because the market is free to give you both.