intraday volume analysis image

[caementarius]

I made the following chart by taking the average % of volume in each minute for the S&P500 from Jan -> 10/3 2008. One thing that surprised me was how pronounced the 10AM, 3:40, and 3:50 minutes were. I mean, I had always heard program trading kicked in during certain periods or whatnot but I thought things would be more smooth.

Now, how would one trade on such patterns? Not sure - but in the spirit of sharing what I'm looking at in the hopes of feedback, herewith the chart:

 

[nazzdack]

1) The only certainty is that the first and last hours of the day are busiest. Mid-day is slowest.
2) The 10am "peak" may also have more to do with economic reports that are released at that time compared to other times of the day.
3) The 3:40pm and 3:50pm peaks can also have to do with daytraders evening-up before the close.
4) Take it with a grain of salt.

 

[silenttrader]

Few days back wall street journal published an article saying that late day surge in volume and prices are due to new ETFs, whose fund managers buy or sell depending upon new orders. others have blamed late day price movements to hedge fund redemptions.

 

[caementarius]

Quote from nazzdack:


2) The 10am "peak" may also have more to do with economic reports that are released at that time compared to other times of the day.

Ah, right. Makes sense. There are tiny peaks on each hour, too.

By the way, the dark line is what volume would be if each minute had an even percentage (100%/390min).

Thanks.

 

[Lucrum]

I've played around with average "time of day" volume. My only comment is if you're going to try to use it you may want to look a longer time frame.
Probably 5 mins at a minimum and even better 10 - 15mins.

 

[nazzdack]

Quote from caementarius:

There are tiny peaks on each hour, too.

That may be from traders who use 30-minute and 60-minute bar charts along with the Market Profile guys at the beginning of each hour. Take it with a grain of salt.

 

[nonlinear5]

Quote from caementarius:


Now, how would one trade on such patterns? Not sure - but in the spirit of sharing what I'm looking at in the hopes of feedback, herewith the chart:

I think this is very typical U-shaped trading volume distribution with the peaks at the open and the close, and the bottom during the lunch hour.

 

[Spydertrader]

Quote from nonlinear5:

I think this is very typical U-shaped trading volume distribution with the peaks at the open and the close, and the bottom during the lunch hour.

It is.

- Spydertrader

 

[caementarius]

Quote from Spydertrader:

It is.

- Spydertrader

Yes, exactly, what's kind of interesting to me (not implying that it's very significant) is the regularity of the aforementioned pulses. 10AM makes sense as pointed out. Not sure about 3:40 and 3:50 -- seems like if I were doing large volume for a fund I'd pick 3:37 or 3:53 just to avoid the compressed time period.

Also interesting how the afternoon ramps up while the morning volume drops precipitously. The afternoon appears to have much more of the day's volume.

Market theory folks would say trading happens in response to the release of information (10AM makes sense this way). Yet the pent up information from the last evening or weekend gives relatively meager volume in the AM. Instead, there's a hurry-up-before-the-close pattern that reaches a fever pitch right before the close.

I guess it's all common knowledge - I'm just trying to get a good idea of what happens on an average day.

 

[jack hershey]

For trading equities, you will find that doing 100 cycles a year is well supported by this chart and rules. It is necessary to use high beta stocks to optimize making money.

Elswhere you can google or do searches on how manage increasing holding by an order of magnitude or two orders of magnitude within the duration of the non stationary platform for maintaining the look up tables associated with attachment.

The same applies for the commodities.

You will find that you can segment volume ranges for various gating purposes. As you may expect determining the ranges is best done with respect to price volatility and price overlap.

There are also seasonal adjustment considerations.

Finally, you have found the influence of two types of RTH news announcements. As you examine this further you will acquire an interest in how blow outs work during "cascading" after news.

Were you to do an analysis of the components of market premium drift, then you would arrive in the ballpark of how to use the more sophisticated gating signals for trading ahead of market bots and having your trades pushed by them.

There is a general question in the space about who controls the market. Once you get into that ball park, then you get to see how counterintuitive the markets really are. I hope you turn the corner before you go too far down the road.

If you like backtesting, give yourself a real scare by testing the attachment with a 3 beta or better universe.