Forex august low liquidity: why your bot takes fewer trades

It is mid-July right now. In a few weeks you will notice it. You open your app on a Tuesday morning, see zero open positions, scroll back through your history and count three trades in the last ten days. In May there were twelve. Your first thought is that something has broken.

Nothing has broken. It is August.

The short answer, for anyone who only wants that: forex volume drops roughly 20 to 40 percent over the summer because institutional desks are on holiday, with the last two weeks of August the thinnest, and that means wider spreads and sharper moves on small news. So a good bot takes fewer trades in August, not more.

Why is the market so quiet in August?

Because the people who make the volume are on holiday. Bank traders, hedge fund managers and institutional desks across Europe and the US are largely away in August. The market stays open, but the money that normally flows through it is not there.

This is easy to underestimate when you trade from a laptop. You think in terms of EUR/USD on a chart. The real market is a few hundred trading desks in London, New York, Frankfurt and Singapore absorbing each other's orders. Switch half of them off and the chart still looks the same, but there is far less behind it.

Think of the M25 at three in the morning. Same road, same tarmac. But there is almost nobody on it, so one car slamming on the brakes changes the whole picture. In rush hour, one car braking does nothing at all.

Forex august low liquidity: how much does volume really drop?

Trading volumes routinely fall 20 to 40 percent below peak levels across the summer. August is one of the quietest months of the year in forex. This is not theory, it is the same pattern every year.

Do the arithmetic on your own account. Say your bot normally takes 15 trades a month on setups that tick every criterion. Strip 30 percent out of the market activity those setups grow from, and you are somewhere around 8 to 10. Add that a chunk of what is left is noise the bot deliberately lets go, and you land at 4 to 6. That is your August.

It feels like something is wrong. On your screen, the difference between "no setups" and "bot is down" is exactly zero.

And it does not drain evenly. The summer runs to a rhythm. July starts out normal, with a perfectly decent first half. Around the end of July it begins to sag. The first half of August is noticeably quieter, though things still happen. Then the last two weeks of August arrive and the floor drops out: European desks are near-completely shut and US institutional participation hits its annual low. That is the thinnest stretch of market in the whole year. Fix those two weeks in your head. Early September everything snaps back as if nothing happened, usually within a few days.

The last two weeks of August are the thinnest market of the year. Which is exactly when most people get impatient and start forcing things.

Those two weeks are also where the mistakes get made. Not because the market goes mad, but because the trader does.

A thin market is more dangerous, not safer

Thin liquidity means wider bid-ask spreads, more slippage and less efficient price discovery. Your fills get worse. And that is the calm part.

The real problem is that a thin market exaggerates moves. Relatively minor news or a middling data point can trigger a sharp swing that normal volume would simply absorb. In June a second-tier number dies in the order books. In the last week of August there is nobody on the other side and price runs.

Plenty of people assume quiet also means smaller swings. The opposite is more often true. You get less movement and more violent movement at the same time, which is an awkward combination to trade by hand. The standard risk advice for thin markets is exactly what you would expect: cut your leverage, trade smaller size, and expect worse fills than you are used to.

If you want to see how a single data point can pick a market up and move it, read our piece on NFP and why a bot works around it. In August the same effect applies, just on far smaller news.

My bot barely takes any trades in August, is it broken?

Probably not. A bot that filters on setup quality simply finds fewer of them in a thin market. Fewer trades in August is the expected behaviour.

Turn the question around. What would it mean if your bot took as many trades in the last week of August as it did in May? That it does not take its own criteria seriously. That it mistook a swing on thin volume for a real move. That is the bot to be scared of, not the quiet one.

Before you change anything, check two things. Is your VPS still running, and is the connection to your MT5 still live? If both are yes, and your history shows August trades did happen, just fewer of them, it is working. Silence after a correctly closed last trade is not a fault. A bot that drops out in the middle of an open position is a fault.

What to change in August, and what to leave alone

On the bot: no. On your expectations: yes.

The urge to tinker peaks in August, and that is the danger. You have watched almost nothing happen for two weeks, you get restless, and you start loosening settings so that "something happens again". You drop the threshold for a setup. You switch off the news filter. Now you have a bot taking the most trades of the year in the thinnest market of the year. That is how August costs people money.

If you also trade by hand alongside it, the rules are simple: smaller size, less leverage, and accept that your fills are worse. Our London open setup still works in August, but with less volume behind the open you have to be sharper on your entry. Want to spend the month usefully? Read your risk rules instead of stretching them. There is a full piece on risk management with a trading bot.

Our August looks the same as our March

Nothing special. That is the point.

The bot runs on exactly the same rules. It waits for setups it actually likes, and in August there are fewer of them. We do not loosen the criteria to fill the month out. We would rather take 4 good trades in August than 20 forced ones.

That is easier to hold to here than at most providers. We take 30% commission on your profit and 0 on losses. A quiet August earns us nothing, and neither does a forced trade that loses. No reason to make your bot impatient, because we only win once you win.

And then it is September. The desks come back, volume snaps back to normal, and your bot takes its setups again. Every year. The only people who come out of August with a problem are the ones who decided halfway through that silence meant something had to happen.

Leave it alone. That is the whole lesson.