Friday, February 09, 2007

Finding Opportunity in the Stock Market: A Tool For Tracking Large Trader Participation

In a recent post, I suggested that we can track the presence of large institutional traders by knowing how current volume compares to average volume. Because the majority of volume is attributable to large trades (and large traders), below average volume or above average volume gives us an indication of the relative absence or presence of these large market participants. Since many of these participants are trading directionally, their presence is associated with enhanced volatility and greater likelihood of short-term trending behavior.

But how do we track high and low volume during the day, given that average volume varies greatly through the day? Here is a simple tool I've made available to the prop traders I work with. It's also sits on my desk as I trade for quick reference. It's a synopsis of median trading volume in the ES futures for every 15 minute period in the morning over the past month (N = 22 trading days). The chart above shows the median trading volumes plotted as a cumulative curve, so that you can see what average total volume looks like at each point in the trading day. If today's trading volume in ES, as reported on your datafeed, exceeds the curve, you know we have above average volume. If we're falling below the curve, we have a relative absence of large traders--and a greater likelihood of a narrow, rangebound market, as has been the case most of this week.

You can click on the chart, print it out, and use it for reference during the trading day. I like to update mine weekly.

Here are the median trading volumes--and trading ranges--for each 15 minute period (Eastern Time).

9:30 - 9:45 AM - Median volume is 66,022; median range is 2.75 ES points
9:45 - 10:00 AM - Median volume is 58,107; median range is 2.875 points
10:00 - 10:15 AM - Median volume is 59,408; median range is 3 points
10:15 - 10:30 AM - Median volume is 46,598; median range is 2.25 points
10:30 - 10:45 AM - Median volume is 53,226; median range is 2.75 points
10:45 - 11:00 AM - Median volume is 36,549; median range is 2.125 points
11:00 - 11:15 AM - Median volume is 36,203; median range is 2.375 points
11:15 - 11:30 AM - Median volume is 27,419; median range is 1.75 points
11:30 - 11:45 AM - Median volume is 27,273; median range is 1.75 points
11:45 - 12:00 N - Median volume is 23,638; median range is 1.75 points


Note that both volume and volatility fall considerably during the course of the morning, with a notable dropoff around 10:45 - 11:00 AM Eastern Time. This is why I commonly stop trading (and blogging) around that time. Volume tends to dry up, and opportunity for short-term trading tends to dry up with it.

Also observe that a trading volume of 50,000 contracts at 11:00 AM ET might not attract a trader's notice, because it is not higher than the volume seen earlier in the day. For that time of day, however, 50,000 contracts would represent seriously elevated market participation. You would want to identify what is going on at that time (breakout from a range; news item; etc.) and see if it represents an opportunity for you to join the trend of the large traders. Similarly, if you see below average volume for the first 15 minutes of trading (as has happened most of this week), you'd want to mentally prepare yourself for a slow, narrow-range day.

Finally, knowing these numbers helps you figure out when an economic report at 10:00 AM ET is truly having a significant market impact. If the volume shoots significantly higher than the average volume for that time of day--and is higher than the volume earlier in the day--you know that the economic release is attracting large traders. These are the participants likely to reprice the market. On the other hand, if a release comes out and you don't see an elevation of volume, you have an excellent indication that this will have no meaningful impact on subsequent trading.

(Note that the median 15-minute ranges give you a sense for expectable movement in any 15-minute period, which can be useful information in determining when a move has gone about as far as it's going to in the short run.)

My sense is that these curves can be created for any futures contract and any stock. They provide you with a sense for who is usually in the market--and who is participating now. The periods of enhanced participation will provide you with the periods of greatest market movement, and the largest traders will point the way to which way the market is likely to move.

I will update the data periodically if there is interest. Because most large cap U.S. stocks and equity indices are highly correlated with the S&P 500 Index, a reading on ES volume will provide useful clues as to likely activity in your trading instrument.

13 comments:

yinTrader said...

Hi Brett

Just finished reading your post on tracking large trader participation.

I have printed out the chart and imposed against your ET time with SinT since you are now 14 hours behind us in Spore.

A quick perusal of your chart against the ES ranges show up quite accurately what I have perceived with my own charting tool on 15 mins volume on ESH7.

Hope you will update this from time to time as it is very useful and simple!

Hsieh hsieh ni, gong xi fa cai!

denber said...

This is great. Very thought provoking. This really adds information I don't get from my usual candlestick analysis. I intend to add this to my program and see if it improves its results. I also agree that it should be useful for any issue, not just ES.

It may even explain why I've been having problems making money with ES. As a night person, I generally skip the morning session entirely. Perhaps I should switch to EUR :-)

Simple Trader said...

I had developed a Tradestation indicator that does something similar. It calculates the cummilative average value for each bar over the last x days. Then it compares the current cumullative volume to the average at that time and prints it as a histogram oscillating above/below the zero line as a percentage.

But unless this ratio is substantially less e.g. below -40%, I have found it to be unreliable mainly because volume is spread out unevenly e.g. today the volume had dried up as usual by 11am but at 12 noon heavy volume came in. This is also the case where different days of the week or holiday weeks or days with news events skew the ratio. To improve upon it, I reset the running total to zero at key timeframes e.g. 9:45am, 11am, 1pm, 2pm which seemed to improve it a little better. It also seems to work better in the morning session.

I ended up going back to a basic 2 min chart of the volume, its 5 bar average and a horizontal line at 300 contracts for the ER2. I adjust the horizontal line every weekend by looking at the data from the last 2 weeks visually.

MidKnight said...

Thx for sharing this Brett. I will definitely be checking this out.

Thanks again.

Best regards,
Matt

Brett Steenbarger, Ph.D. said...

Hi Yin,

Thanks for the note; I'll be happy to update. It's also helpful for traders of NQ and ER2 to develop similar charts for their instruments.

Brett

Brett Steenbarger, Ph.D. said...

Hi Denber,

You raise a great point. The AM session is when we tend to have highest volume and movement, so you potentially miss out on opportunity by skipping the mornings. Thanks for the comment--

Brett

Brett Steenbarger, Ph.D. said...

Hi Simple Trader,

Yes, good observation re: Friday's volume coming at noon hour. When volume is much higher than the average volume for that time of day, it is a good indication that large traders are stepping into the market. When that occurs in conjunction with breakout moves, per my recent post, that is when trending moves take off. Thanks for the note--

Brett

Brett Steenbarger, Ph.D. said...

Hi Matt,

Yes, I think it's worth checking out and modifying to adapt to your own trading style. Thanks--

Brett

aiva_trader said...

Hi Dr.Brett I use the same indicator with neoticker, however there is another variation of this for esignal users (for free) by creating different time templates on esignal. For example you can plot a time template that only plots the data on the e-mini for 930 to 945, then plot it on a 15 min chart with volume and a volume ma. This will give the user a look at relative volume for that particular time of day.

Thanks Again for all of your great insight over the past couple of years.

aiva_trader

Brett Steenbarger, Ph.D. said...

Hi Aiva Trader,

Thanks so much for that e-Signal tip. I wasn't aware of that functionality in e-Signal, but will now look into it. I'd also be interested in hearing some time about your experience with Neoticker. I appreciate the note--

Brett

Yaser Anwar, CSC (Trader- Equities & FX | Quoted 7 Times In WSJ, Twice in NYT & FT) said...

Sir-

This is an excellent chart, you're volume posts are phenomenal.

I hope you devote a massive chapter on volume analysis in your next book :D

Just a quick confirmation: for each day, the volume has to breach the curve at each time interval to notice the tracks of the big boys, right?

Thanks!

Brett Steenbarger, Ph.D. said...

Hi Yaser,

Yes, two things to look for is whether the curve for the day is above or below the average volume curve overall and whether volume readings for a particular time of day greatly exceed the average for that time of day. Sometimes you can have an average looking curve, but then greatly exceed the average at a particular time due to the effect of news, a price breakout, etc. Thanks for your note--

Brett

Ing. Alessandro Bertolino said...

Hi, great idea!
Just few questions:

Why 22 trading days? (Not 20 or 30..)
Why 15min the time frame over make volume averages?
And what about a 1 hour time frame?
Bye