How Do You Do Overnight Trading? (How To Make Money Overnight)
Last Updated on April 24, 2023
How do you do overnight trading? Is it possible to make money via overnight trading? Can you make money by owning stocks or indices for just one day? Is overnight trading good?
You do overnight trading by buying the close and either selling at the next day’s open or close. We have previously covered how you can buy at the close and sell at the next day’s open in the S&P 500. This article looks at how you can expand the time frame and hold for a few hours more.
The overnight edge
The overnight trading edge in the stock market is a well-known edge and has been covered on this website multiple times. The edge is mainly from the close until the next open (you can find several of these edges among our free trading strategies and in our paid subscription for monthly trading edges).
Why buy the close and sell the next day’s open?
We exit (or sell) on the open because the average gain from the open to the close has been zero during the last 30 years. Below is the chart showing the equity curve if you bought the open and sold on the close (ie, day trading):
But averages are just averages and generalizations. We are not interested in averages, but we are interested in the more infrequent excessive returns that happen when certain conditions occur.
Is it possible to find trading edges that have a better edge by selling at the close (and not on the open) in S&P 500?
Yes, there are quite a few of them. If you develop many of them, it’s possible to mix a nice portfolio of these short-term edges and make a decent risk-adjusted return.
Below we show some equity charts of a few ideas we have, and a few of them will later be published to our subscribing members. All the backtests have been done in the ETF with ticker code SPY since its inception in 1993. You’ll get more or less the same results by using the ES future.
In the examples below we enter at the close, and we sell at the close the day after, thus holding for 24 hours (or over the weekend if it’s a Friday). To better understand how we buy and sell at the close in live trading, we recommend reading this article:
The equity charts below assume no leverage and 100% of equity invested and compounded.
The Turnaround Tuesday trading strategy
We start with the well-known Turnaround Tuesday trading strategy. Despite being well-known for about four decades, it’s still working well. Because this strategy is covered earlier, we’ll only show an example of the Turnaround Tuesday variants:
Holiday effect overnight
In a previous article, we covered how the S&P 500 performs around all the US holidays. Seasonal trading strategies are relatively persistent, and it turns out there are quite a few possibilities for making good short-term trading strategies out of seasonal patterns.
Below is a trading edge that trades around one of the holiday seasonalities:
There are not many trades, 94 trades since 1993, but the average is a solid 0.55% with a 70% win ratio. We believe these are pretty good stats for holding a position for 24 hours, and the strategy is also a bit logical.
End of month effect
The end-of-month effect is also well-known and is somewhat overlapping with the turn of the month trading strategy. We use this effect in our own trading and here is an example of a one-day overnight trade:
There are 233 trades and the average gain is 0.3%, the win ratio is 66%, and the average winner is bigger than the average loser.
Day of week effect – weekday effect
We have mentioned the Turnaround Tuesday Strategy, but there are other weekday effects in trading. Below is a strategy that exploits one of them:
In total there are 305 trades, 68% win ratio, and the average winner is substantially bigger than the average loser.
Mean reversion overnight
Below is an equity chart that shows a classical mean-reversion pattern:
The 373 trades made an average gain of 0.34%, had a win ratio of 60%, and the average winner and loser were pretty close.
Option expiration effect
One peculiar seasonal effect is the options week expiration effect. The equity curve below uses one variable related to the options expiration week:
The average gain is 0.2%, the win ratio is 60%, and the average winner is just slightly bigger than the average loser.
Here is another twist to the options expiration week:
169 trades, the win ratio is a low 57%, but the average winner is 30% larger than the average loser.
A mix of strategies
Below is the equity curve of a mix of overnight strategies – both mean reversion, breakout, and gaps:
237 trades and an average gain of 0.41%.
What if we combine all the overnight strategies?
To make decent returns you need to trade quite frequently when having such a low holding period. Thus, let’s check the performance of all the strategies backtested in one backtest as one portfolio of strategies (with no overlapping trades):
The blue line under the equity curve is the drawdown along the way,
Some facts and numbers:
- CAGR: 14.6% (buy and hold 10.4%)
- Time spent in the market: 19%
- Number of trades: 1373 (47 trades per year)
- Average gain per trade: 0.3%
- Win ratio: 62%
- Average winner: 1.01%
- Average loser: 0.87%
- Max drawdown: 15%
- Profit factor: 2
- Sharpe Ratio: 2.05
Overall, the sum of the strategies performs well on all trading strategy and system performance metrics.
The strategies were all developed in 2017, but for different reasons, we didn’t start trading them ourselves until 2021.
Are the returns realistic?
Yes. Commissions are currently practically zero, and the slippage in live trading is much smaller than what most people believe.
Is the backtests curve fitted? We don’t believe so. All strategies have a maximum of two variables except one, which has three.
Are the strategies liable to survivorship bias? No, all the strategies we made in 2017 are included. Moreover, we didn’t datamine back in 2017.
But always keep in the back of your head that backtests are nothing more than approximations. You can assume that a backtest is always “wrong”, but that doesn’t make it useless. You need to use good judgment when using a model or backtest and make sure it has some kind of logic or structure in it. We believe most of the ones above are.
The only drawback is, of course, that you need to buy and sell at the close, which is, for many cumbersome. You need to trade automatically or mechanically, something we have described in detail in our Amibroker course.
If we zoom in and look at the performance in the out of sample period from 2018, the equity chart looks like this:
The Covid-19 mess in 2020 made some windfall profits, and the average gain per trade has been 0.34%.
Now that we have described some potential trading strategies, we end the article with some words of how to construct a portfolio of strategies:
The difficult task of constructing a portfolio of strategies
One of the most important aspects of both long-term investing and short-term trading is to construct a portfolio of assets that perform well together. This means, in practice, that you can’t invest in a portfolio of only oil stocks. This is pretty obvious.
What might not be so obvious is that the same principle applies to short-term trading. You don’t want to trade 10 mean reversion swing strategies in the same instrument, for example, the Nasdaq index, because most of the trades overlap each other.
Even if you have two excellent mean reversion strategies, the two might perform worse than one independently. This is what makes system development a bit time-consuming and hard.
You shouldn’t aim for perfection in trading development. The only holy grail in trading is to trade many different strategies.
We repeat what we have always said:
It’s better to have a portfolio of many sub-optimal trading strategies than one or a few “perfect” strategies.
Correlation of mean reversion in stock indices
Most stocks move in tandem and in sympathy/correlation with each other, and thus trading a mean-reverting strategy on a basket of stocks gives many signals at the same time.
If you have a mean-reverting strategy in SPY or the ES futures, the overlap to most other indices is huge – even for indices in Asia or Europe. We trade the DAX futures, and most trades happen at the same time as with SPY. Thus we need to make changes in the risk settings of when to skip or take a trade to avoid overlaps.
Avoiding drawdowns to compound
If you want to compound the most effectively, it pays off to avoid the worst bear markets. The reason is simple:
Let’s assume you start with a capital of 100, and the market falls 30%. If you own the market, your portfolio is then worth 70. Next year the market goes up 20%, and your portfolio will be worth 84.
But let’s change the numbers a bit. If you managed to avoid much of the downturn in year one and only lost 10%, your portfolio is worth 90 (not 70). If you manage 15% the next year, 5% less the market’s return, you have 103.5 after year two. The lesson is this:
Even if you, in the long term, manage an arithmetic average that is lower than the market’s average, you can still outperform the market. The reason is that you avoid some of the worst drawdowns. If you avoid drawdowns, you can start compounding from a higher plateau.
To better understand the sequence of returns risk we recommend these four articles:
- How likely are you to go broke as retired or FIRE? (Sequence risk, diversification, and withdrawal rate)
- Dollar cost averaging vs. lump sum investing – sequence of return risk
- Mark Spitznagel – Safe Haven Investing
- Why arithmetic and geometric averages differ in trading and investing
The most important factor in quantified trading is having unrelated strategies, i.e. strategies that don’t correlate with each other. That is, of course, no easy task and requires a lot of work.
How to make uncorrelated strategies
There are several factors at play when you develop trading strategies:
Different asset classes and markets
We recommend looking at different asset classes. Oil, gold, and Bitcoin might go completely different than stocks, for example.
Trade different time frames
Vary size and time frames in trading can also be helpful. Day trading and short-term trading could substantially improve results no matter the market’s direction.
To better understand why, please read or anatomy of a bear market. Long has worked very well for 1-3 day strategies in a bear market! The most powerful up days normally happen in a bear market.
Different strategies: Mean-reverting, trend-following, or momentum?
Stocks are typically mean reverting in the short term while trending long term. When one is not working, the other should absorb the losses.
Likewise, seasonal trading strategies are useful for all asset classes.
However, keep in mind that trend-following strategies are challenging to trade because they have mostly a low win ratio. Typically, the rare big winner is what makes all the difference. Very few traders can follow such strategies, something the famous trading coach Brett Steenbarger argued for in his personality test for traders.
Make sure you have both long and short strategies
Most short strategies perform poorly, no matter the asset class. Short is not the opposite of long. If you have a good long strategy, you can be pretty sure it doesn’t work for short if you do the opposite. Short selling is difficult – period.
However, even short strategies that, on their own, are not performing well can still add value to your portfolio. The reason is what we described earlier in the article: risk mitigation and lowering the drawdowns, something described very well by Mark Spitznagel. It gives downside risk reduction and at the same time, might contribute to a better overall CAGR of the portfolio.
How do you do overnight trading?
The beauty of overnight trading is the short holding period. This is not in conflict with buy and hold vs market timing, an article in which we argued that it’s hard to time the market. But a trading edge is not about market timing. Market timing is about being out of the market because you believe it’s overvalued, so these are two separate things.
Why does this matter?
It matters because most of the gains over the past decades come from a small percentage of the trading days. Most of the time, the market has low volatility, and thus, it’s hard to make any money. However, by trading like a barracuda, we believe you can improve the risk-adjusted return a lot.
Barracudas rely on surprise and short bursts of speed to overtake their prey. Overnight trading is much the same.