When Your Breakout Bot Should Sit Out: Gap Size, Prior-Day Range, and the Open Conditions That Wreck ORB

Every automated strategy has a market it was built for. Opening Range Breakout was built for a market that opens, sets a range, and then commits to a direction. Give ORB that market and it does exactly what the backtest says. Give it a gap-and-chop open instead, and it will happily buy the high of the day, stop out, and do it again on the reverse.

Traders new to automation usually try to fix that with more indicators. The better fix is a filter: a short list of conditions where the bot simply doesn't take the trade. Sitting out is a position. Here is how to decide when it's the right one.

Why the open is a different market than the rest of the day

The first 30 minutes of the futures session carry the day's heaviest volume and its widest spreads. Overnight inventory gets repriced, orders that queued up while you slept hit at once, and the range that forms is genuinely informative — most of the time.

The exception is the open that has already made its move — one of several places where a strategy has to adapt to the conditions in front of it. When price gaps hard away from the prior settlement, the imbalance that ORB is designed to catch has largely resolved before the bell. The breakout that follows isn't fresh demand; it's the tail end of a move that started at 3 a.m. That's the setup where a breakout bot buys the last tick of enthusiasm.

Gap size: the first filter worth writing down

Measure the gap as a percentage of the prior day's range, not in points. A 20-point gap on ES means one thing after a 90-point range day and something completely different after a 25-point range day.

• Gap under ~25% of prior-day range. Normal. The opening range still has room to form and mean something. Standard ORB conditions.
• Gap 25–50%. Caution zone. The move is real but partially spent. This is where a wider stop or a half-size entry earns its keep more than a full-size breakout does.
• Gap over ~50%. The market has already expressed its opinion. Breakout entries here have a much higher chance of being the exit liquidity for whoever positioned overnight. This is a sit-out day for a pure breakout bot.

Those thresholds aren't magic numbers — they're a starting point you should confirm against your own data. With 4.5 years of backtest history behind the Push Button Trading bots, you can segment results by gap bucket and see where your edge actually lives instead of guessing.

Prior-day range: the compression signal nobody checks

Gap size tells you where you're starting. Prior-day range tells you how much fuel is left.

After an unusually wide range day, the market often spends the next session digesting: overlapping value, false extensions, and a lot of two-sided rotation. Breakouts extend less. After a narrow-range day — the condition our NR7 bot is built around — the odds flip. Compression tends to resolve into expansion, and that expansion is exactly what a breakout strategy wants.

So the pairing matters more than either number alone. A small gap after a narrow-range day is close to the ideal ORB environment. A large gap after an already-wide range day is the one to skip.

The other open conditions that quietly wreck ORB

Scheduled data inside the range window. A CPI or FOMC print landing while the opening range is still forming doesn't create a breakout — it creates a spike and a retrace that looks like one. Blackout windows around known releases belong in your configuration, not in your discretion at 8:29.

Half sessions and the days around holidays. Thin participation makes ranges too easy to break and too easy to reverse. The range forms on volume that doesn't represent anyone's real positioning.

An opening range that's simply too wide. If the first 15 or 30 minutes produce a range larger than your average full-day move, your stop is now enormous and your reward-to-risk has quietly inverted (worth reviewing how your stop and entry order types behave in a fast open). Cap the acceptable range width and let the bot pass.

Immediate failed extension. Price pokes above the range, fails to hold for a full bar, and slides back inside. That's not a breakout day; that's a rotation day announcing itself early.

Sitting out isn't the same as being flat

Here's the part that matters for a real portfolio: when the breakout bot stands down, that doesn't mean your account does nothing. A chop-and-fade open that's hostile to ORB is often the exact condition a mean-reversion strategy like Snapback is built for. Wide-range digestion days favor different logic than compression days do.

That's the whole argument for running a mixed portfolio rather than one strategy at maximum size. Filters that take one bot off the field on the wrong day are far easier to live with when another bot is still working. If you're building that mix, the Bot Portfolio Builder and Analyzer are designed to show you how the pieces behave together — including the days when the right answer is fewer trades, not more.

How to actually implement this

Don't add five filters at once. You'll strangle the strategy and you won't know which rule did it.

1. Pull your ORB results and tag every trade with gap size as a percent of prior-day range.
2. Sort into buckets. Look at win rate and average loss, not just net P&L — a filter that removes your five worst days is worth more on a funded account than one that adds a few winners.
3. Add one filter. Run it forward. Confirm the trade count drop matches what the historical data predicted.
4. Only then consider a second condition.

On a funded evaluation, this discipline compounds. Trailing drawdown doesn't care that a loss came from a "reasonable" setup — it just moves the line. Removing the two or three worst-condition trades a month can be the difference between clearing an evaluation and paying for another reset.

Put the filter to work

Automation isn't about trading more. It's about trading the same rules every single day, including the rule that says not today. Emotion is what makes a discretionary trader take the bad-condition breakout at 8:31; a filter is what stops a bot from doing it 200 times a year.

If you want to see how gap and range filters change results across a portfolio of breakout and mean-reversion bots, start with the 14-day free trial and run the Bot Portfolio Analyzer against your own configuration. No coding required — just the discipline to let the bot sit out when the market isn't offering the setup it was built for.