Have any of you discovered that some days of the week are better to trade than others and can vary depending on the instrument? If so, please reply to this thread.

In a 12 month backtest of a few strategies, for example, Bonds were more profitable over the past year on Tuesday and if only traded on Tuesday, would have been almost as profitable than any other combination of days. Statistics such as Profit Factor and Win % rise dramatically, albeit with fewer trades, obviously. The benefits include less commissions, time-in-market risk and trader stress.

I find this fascinating and wonder what causes this to occur. I have found this phenomenon in each and every instrument I test and the strategy rules are consistent across all tested markets.

I plan to start a journal with a portfolio of instruments trading only the optimal day of the week to compare live results with the tests.

Tundi 

Have any of you discovered that some days of the week are better to trade than others and can vary depending on the instrument? If so, please reply to this thread. In a 12 month backtest of a few strategies, for example, Bonds were more profitable over the past year on Tuesday and if only traded on Tuesday, would have been almost as profitable than any other combination of days. Statistics such as Profit Factor and Win % rise dramatically, albeit with fewer trades, obviously. The benefits include less commissions, time-in-market risk and trader stress. I find this fascinating and wonder what causes this to occur. I have found this phenomenon in each and every instrument I test and the strategy rules are consistent across all tested markets. I plan to start a journal with a portfolio of instruments trading only the optimal day of the week to compare live results with the tests.

For bonds the treasury auctions have an impact. I think that the results could be even more interesting, if you compare the weeks that have scheduled auctions to those that have none.

Tundi 

Have any of you discovered that some days of the week are better to trade than others and can vary depending on the instrument? If so, please reply to this thread. In a 12 month backtest of a few strategies, for example, Bonds were more profitable over the past year on Tuesday and if only traded on Tuesday, would have been almost as profitable than any other combination of days. Statistics such as Profit Factor and Win % rise dramatically, albeit with fewer trades, obviously. The benefits include less commissions, time-in-market risk and trader stress. I find this fascinating and wonder what causes this to occur. I have found this phenomenon in each and every instrument I test and the strategy rules are consistent across all tested markets. I plan to start a journal with a portfolio of instruments trading only the optimal day of the week to compare live results with the tests.

you would be surprised by how much both w/l ratio, win rate can be improved if you add a simple trend filter. Please make sure you test various parameters to avoid curve fitting. The optimal parameters are always a trap.

Having tested this idea in-depth, there's certainly a day-of-week edge to be found, just as there is for hour-of-day (or time-period-of-day), but keep three very important things in mind:

1. Make sure your data pool is large enough to produce trustworthy data. If you're looking at a period in which just 50-100 trades have been taken, thats approximately 15 or so trades per day-of-week, which is nowhere near enough data to provide you with results upon which you can draw conclusions.

2. One way to help with the issue mentioned above is to test day-of-week results across an entire sector, or grouping of instruments. . . for example, in my futures trading, I like to test all currencies together, as a cumulative whole, and seek commonalities. This provides a larger data pool, and because there is some correlation across instruments, you'll occasionally see agreement in the day-of-week or hour-of-day results, across the board. . . this is often significant, and should be paid attention to.

3. I strongly suggest ignoring day-of-week profit breakdowns that look like a chaotic roller coaster, such as monday performing extremely well, tuesday very poorly, weds extremely profitably, thurs poorly, etc. The only time I find myself truly intrigued and excited by day-of-week breakdowns is when I see a clear linear trend. . for example peaking at the end of the week, and trailing off to a minimum at the beginning of the week (or vice versa), or a peak on a certain day, lets say wednesday, with the neighboring days being the 2nd best performers (tues/thurs, in this case), and falling off from there (mon/fri slightly weaker, sunday perhaps the weakest of all).

Hope this is helpful

Things like "day of week" probably aren't strong enough to develop any sort of system that most people would want to trade. Honestly, most things people call edge are simply too weak to make them worthwhile. But, when you start to think about something like this you need to think about WHY something may be the case. Why should the markets trade differently on different days of week?

Here are some reasons that certain days may perform differently. I'm thinking primarily about S&P 500 but other instruments also will follow.

1. Traders unwilling to hold over the weekend. This would show up in Friday/Monday biases.
2. Reports being released or tending to be released on certain days.
3. Any sort of structural component dealing with leverage. For example, day traders have leverage. In the ES/S&P 500 is not uncommon for the market to late day shake out those who strictly can't hold over the closing gap. If there were some any sort of structural or cyclical leverage factor involved it might show up in day of week biases.

As for the S&P 500, I used to always see a bearish Thursday bias. I don't know why. But, one possibility might be that anything bearish that came out or was expected to come out resulted in traders reducing exposure before the weekend. Why Thursday and not Friday? Something like that could happen because traders are anticipating other traders actions. If they think other traders might close on Friday then they might try to close on Thursday to get the jump.

Now for a bias like this, think about what might happen if the market becomes very bullish. It might not be too surprising for it to reverse. For example, if traders are very bullish they may want to get long before the next report or news cycle comes out or they might be willing to take the risk of holding over the weekend, i.e. to capture the profits because the risk is very low.

Keep in mind that statistically some days of the week must do better than others. For example, if your strategy took a large loss on a particular day of the week then that might skew the results. It would not automatically imply that such a loss would be more likely going forward though.

It shouldn't come as a surprise but perhaps worth mentioning that 1 or 2 days leading up to holidays are typically illiquid and not worth trading, at least for me. Christmas, New Years, Thanksgiving, bank and religious holidays, Labor day, or basically any long weekend that a family would be away I find markets don't always behave normally, but on the day after everything is back to normal with no hangover.

Thanks. I noticed this past Friday seemed to be a bit off, not realizing that Monday is a banking holiday. Next time, I'll pay attention.

This is interesting as I've noticed something similar myself.

1. I live in New Zealand so markets open 11am on Monday but the rest of the world isn't awake yet. What I've noticed is the markets more often that not behave very strangely. Everything from fundamentals and technical's combined will hint one direction but the instrument will creep the other direction, sometimes fly back and forth when the larger continents wake up then will finally revert to the original direction.

I've stopped trading on Mondays as I find it results in losing trades by having my stops wiped out. I find it's better to wait for the rest of the world to wake up, do their thing, then come at it Tuesday morning and observe what's happened.

2. On the flip side, I've noticed there's quite often a sell-off at the end of the week due to a large majority of traders closing their positions before the weekend. It quite often opens up cheaper prices for me to buy at if I'm going long-term!

Mondays and Fridays are great days to lose money. Scalpers would probably say the opposite


Sent from my iPhone using NexusFi

Not sure exactly how to get my point across, but isn't this like missing the bigger picture?


What I mean by that is

#1 Wouldn't it be more likely that certain days with certain instruments potentially on average act differently or trade a certain way? If so it would be better to look into how to trade that, instead of just avoiding that day because your P + L isn't as good.


#2 Also, wouldn't this mean the indicator, strategy or theory you're using to trade with has a weakness. Because if it's based on actual math or solid theory, than you would hope it would pick up on what's happening so that you could avoid the trade or learn how to read it, in order to trade that particular setup / day better.


Hopefully I understood correctly what everyone was saying and that my reply makes sense.