A common rule: enter when a new bar opens, if the previous bar met some condition. In backtests on bar data, the entry price is the open of the new bar. Simple and clean.
In live trading, the entry is sent when the platform detects the new bar, which happens on the first tick. By then, the price has already moved from the previous close. The platform also needs a few milliseconds to process, and the order needs time to reach the broker.
Many other traders and scripts are doing the same thing at the same moment. The top of the minute is crowded. Spreads can tick wider, and fills often come at slightly worse prices than the open that the backtest assumed.
The aggregate effect is a small but consistent slippage that affects every trade. On a strategy with a 1 pip average gain, a 0.3 slip is a 30 percent loss of edge.
The remedy is to model the delay in the backtest: assume the fill occurs a few hundred milliseconds after the open, at the then-current price. If the strategy survives that, it has a margin.
Is the first tick after a bar opening a crowded moment on the pairs traded here?
Entry on the first tick of a new bar
-
FXS London Scalper
- Posts: 976
- Joined: Sun Oct 18, 2026 12:25 pm
Re: Entry on the first tick of a new bar
The first tick of a new bar can be far from the bar's open in fast markets. Accounting for that in backtests, by adding some slippage on bar-open entries, makes them more realistic.
FXS Official Team | London Session Desk
-
Amanhasnoname
- Posts: 2060
- Joined: Fri Oct 30, 2026 9:45 am
Re: Entry on the first tick of a new bar
On EURUSD and GBPUSD it definitely is, at least at the top of the hour and the half hour. I ran a small test last spring where I logged every fill from a bar open entry on the five minute chart next to the open price the tester would have used, and the average difference came out a little worse than I expected, with the worst cases clustered right on the hour when the London and New York desks were both active.
What surprised me was that the spread was the bigger part of it, not the price drift. For the first second or so after the new bar the spread on my account would often widen by a fraction of a pip and then settle back, so a market order sent instantly paid that extra every single time.
The fix that worked for me was simply waiting. I now let the expert advisor hold off for two or three seconds after the bar opens and then check that the spread is back to normal before sending. I miss a few moves, but the average fill got noticeably closer to what the backtest showed.
What surprised me was that the spread was the bigger part of it, not the price drift. For the first second or so after the new bar the spread on my account would often widen by a fraction of a pip and then settle back, so a market order sent instantly paid that extra every single time.
The fix that worked for me was simply waiting. I now let the expert advisor hold off for two or three seconds after the bar opens and then check that the spread is back to normal before sending. I miss a few moves, but the average fill got noticeably closer to what the backtest showed.