Preferred days in the week for day traders

[Laissez Faire]

I'm day trading, mainly the daily top gainers and top looser (fitting my strategy).

Are there specific preferred days, where price moves more significantly during the session ?

Looks to me like Friday is a great day for day traders, had recently a few stocks which moved 10-60% after open on Friday.

Basically, are there more and less volatile trading sessions (end of the week, beginning)?

I made a thread about this a few years back. FWIW, I haven't noticed anything in my own performance related to weekday, although I do think day of week matters based on what I posted prior in the thread.

Check out what I'm quoting from Mark D Cook in the OP. He found in his own performance that Mondays and Fridays were the worst, so he took Friday off and traded Monday lightly.

https://www.elitetrader.com/et/threads/trade-performance-by-weekday-whats-your-data.342664/

 

[M.W.]

About days of the week I can say with pretty strong confidence that I know because I tested it all. Have you done any work in this space? Then let's share, as there is much more value in that than a accusing the other to not know when you really have no idea how much work someone else has done.

You know that, don't you?

You come across as a know it all who don't know how much you don't know.

Just because you didn't figure it out doesn't mean it's not there. The market's not random.

 

[rb7]

Interesting.

Does that mean you don't trade it on Tuesdays at all?

Correct. I don't trade on Tuesdays.
Although yesterday would have been good, but that's just one day.

Even if I reverse everything, it's still a non-profitable day.

 

[Laissez Faire]

About days of the week I can say with pretty strong confidence that I know because I tested it all. Have you done any work in this space? Then let's share, as there is much more value in that than a accusing the other to not know when you really have no idea how much work someone else has done.

I've done plenty of work in this space as I work off a predictive model based on statistics where weekday is a factor. I accurately predicted new lows today as well based on the same model. Posted yesterday in another thread.

Now, seeing your attitude towards my initial post and asking if I'm on drugs, I will politely decline your request to share my private research (which I already shared quite a bit on in this thread).

 

[M.W.]

Perhaps the business of the underlying you trade is exposed to certain days of the week differently than others? So you claim the same, without showing any evidence on your own, even if I showed you several visualizations that would clearly disprove that certain weekdays generate statistically relevant repeating return profiles than other days? The return profiles for each single weekday is approaching zero the larger the sample size because returns on average for the same days cancel each other out,which shows that there is no specific day that has higher returns than others over a large set of samples. At least for spx500, nasd100, Russel 2000, eurostox50, Nikkei, hangseng, kospi, cac, lse, and multiple others I tested.

Yes, I've analyzed this stuff (days of week) extensively.
And for the strategy/model I am using, Tuesdays are clearly not profitable (on average).

I can understand that for others strategy/model, there are no differences.

 

[M.W.]

I took a jab whether you are on drugs because you asked me twice about something that was very clear from the context. Take it easy, nobody questions your skills or experience. Yet I find it strange that the second I suggest to share more specifics about the methodology how you identified weekday anomalies you pull back. It's not that this pertains to IP or strategy specifics. I simply suggested to share which asset class or specific indices you speak of, over what time horizon or sampling space. Anyway, if that's how you want to hold it then I respect that. I stand by my claim that weekday returns are highly unstable over different time periods and they approach zero over large sample sizes as returns cancel each other out. I have the data to back it up and have done so for most developed market indices.

Another poster on the first page displayed returns for 2018 and they clearly show that they are statistically irrelevant,unless you trade massive size to capture 5-15 basis points.more importantly, you would be exposed to large drawdowns in between if you made the effort to break down the return profile throughout the year.

I've done plenty of work in this space as I work off a predictive model based on statistics where weekday is a factor. I accurately predicted new lows today as well based on the same model. Posted yesterday in another thread.

Now, seeing your attitude towards my initial post and asking if I'm on drugs, I will politely decline your request to share my private research (which I already shared quite a bit on in this thread).

 

[rb7]

Perhaps the business of the underlying you trade is exposed to certain days of the week differently than others? So you claim the same, without showing any evidence on your own, even if I showed you several visualizations that would clearly disprove that certain weekdays generate statistically relevant repeating return profiles than other days? The return profiles for each single weekday is approaching zero the larger the sample size because returns on average for the same days cancel each other out,which shows that there is no specific day that has higher returns than others over a large set of samples. At least for spx500, nasd100, Russel 2000, eurostox50, Nikkei, hangseng, kospi, cac, lse, and multiple others I tested.

I'm not arguing with you. It's not about believing it or not.

You've done your tests, using a particular approach and found no evidence. So for you, there isn't a day worst or better profit wise. I respect and believe your results and conclusion.

I've done my tests, using a different approach (obviously!) and found evidence. And I'm using this results for my trading.

We are both right based on our analysis.

FWIW, I'm day trading the ES.

 

[M.W.]

But you would agree if both analyses focused on the same asset over the same sampling space that one would be correct, the other not?

Perhaps we speak of different things entirely, I referred to the basic return profile of weekdays. After re-reading your posts you seem to be talking about your strategy performance profile broken down by weekdays. Very different things.

I'm not arguing with you. It's not about believing it or not.

You've done your tests, using a particular approach and found no evidence. So for you, there isn't a day worst or better profit wise. I respect and believe your results and conclusion.

I've done my tests, using a different approach (obviously!) and found evidence. And I'm using this results for my trading.

We are both right based on our analysis.

FWIW, I'm day trading the ES.

 

[rb7]

But you would agree if both analyses focused on the same asset over the same sampling space that one would be correct, the other not?

Perhaps we speak of different things entirely, I referred to the basic return profile of weekdays. After re-reading your posts you seem to be talking about your strategy performance profile broken down by weekdays. Very different things.

You are right. We are not talking about the same thing at all.

I don't really care about the basic return profile of weekdays which is too academic for my needs.

As I previously mentioned, my results are based on the strategy/model I'm using. And a different one would give different results.

My strategy/model is based on specific market conditions and those conditions are not present on specific weekday.

 

[Laissez Faire]

I took a jab whether you are on drugs because you asked me twice about something that was very clear from the context. Take it easy, nobody questions your skills or experience. Yet I find it strange that the second I suggest to share more specifics about the methodology how you identified weekday anomalies you pull back. It's not that this pertains to IP or strategy specifics. I simply suggested to share which asset class or specific indices you speak of, over what time horizon or sampling space. Anyway, if that's how you want to hold it then I respect that. I stand by my claim that weekday returns are highly unstable over different time periods and they approach zero over large sample sizes as returns cancel each other out. I have the data to back it up and have done so for most developed market indices.

Another poster on the first page displayed returns for 2018 and they clearly show that they are statistically irrelevant,unless you trade massive size to capture 5-15 basis points.more importantly, you would be exposed to large drawdowns in between if you made the effort to break down the return profile throughout the year.

Well, I wouldn't say taking a jab at someone is the way to proceed if you want to exchange opinions and research, is it now?

Fortunately, I'm not easily offended as so many others on this site who would have you blocked by now. LOL.

I did some very detailed research across weekdays on daily return and ranges and other stuff I look at and while I found some differences across weekdays, I don't think it's something you can exploit on that basis alone. Day trading or trading in general is about accuracy and average results are useless for that.

ES from 2005 to 2023:

As for my own approach, I generally keep that close to the chest, although I've shared a few bits and pieces here and there including what I posted in this thread. In summary, I think day of week matters, but not as a simple answer that's directly applicable which is what the OP is looking for. I'm still learning and I certainly don't have all the answers.

This is now a High Low week. If the statistics I posted earlier come true this week as well, we should be expecting more downside come tomorrow or Friday. It's just a roadmap to help guide me through the week. Some of the weekly biases are already exploited by now, so it's hard to tell. But it warrants caution against the long side for sure.