Why Market Depth is Useless As an Indicator

[UrmaBlume 10]

The ultra short term bursts of uber intense trade that often establish local intra-session extremes is never shown as resting orders it is auto-executed in a series of orders separated by only a few thousandths of a second.

 

The chart below shows a spike in trade intensity but while this high volume of trade was being executed it was never shown at any level as a resting order by market depth. This trade was auto-executed.

 

While commercial traders CAN successfully hide their method and whether the trade was an opening or closing transaction, they can't hide the fact of their transactions.

 

[Soultrader 12]

This is quite fascinating stuff UrmaBlume. Thanks for sharing.

 

What platform are you using to write these indicators? Seems like everyone is out to win the volume analysis race.. trying to dig deeper into analyzing what lies within a volume bar. However, I do like your approach of analyzing the number of transactions with time.... or is this something else? Thanks.

[UrmaBlume 10]

SoulTrader,

 

Thanks for the kind words. Not at all like the responses I got from one of your other, less informed, moderators.

 

You are right about volume analysis - it's all that really matters. After all it is volume that motivates price - Price doesn't motivate price and the passage of time does not motivate price, it is the occurrence of trade (buying and selling) that motivates price.

 

Knowing that it is not the passage of time but rather the occurrence of buying and selling that drives price makes one wonder why anybody would ever base their trade on indicators with price as the prime input (EMA, Bollinger, RSI, Candlesticks, Stochastic, CCI....) and to further compound the error, apply those indicators to time constant data vessels.

 

Three decades ago I was taught that the best indicators of future price don't have price as any part of their calculation.

 

As I have mentioned I run a small PRIVATE trading and technology company and the first thing I teach our new traders is that it is buying and selling that drives price. That none of our intra-session charts are time constant charts and almost none of our indicators have price as ANY part of their calculation.

[UrmaBlume 10]

This is quite fascinating stuff UrmaBlume. Thanks for sharing.

What platform are you using to write these indicators? Seems like everyone is out to win the volume analysis race.. trying to dig deeper into analyzing what lies within a volume bar. However, I do like your approach of analyzing the number of transactions with time.... or is this something else? Thanks.

 

We use Trade Station. I am the oldest contiuous user of products developed by Bill Cruz. I even bought his Temp S&P system almost 30 years ago before there was TradeStation, SuperCharts or Omega Research.

 

TradeStations does have some issues. For us many are centered around the lack of granularity with the TS time stamp. For applications like our Market HUD we use dll's to export the data to applications we have written in C++. For finer granularity in the usage of time we access the windows kernel.

 

A couple of months ago we investigated the possiblity of changing to Ninja and ZenFire data. We were able to speak to management at both Ninja and ZenFire and found the ZenFire data stream to be great but that the way Ninja handled the data stream made the data not suitable for our purposes plus Ninja is very young software and needs a couple of more versions before it will be ready for real power users. The best we got out of the discussion was meeting with Mirus futures. Great attitude, great rates and super day margins. They even support member rates and the CME volume incentive plan that gets total round trips in the ES down to $1.80 round trip. While Ninja is not yet ready for our level of trade decision support processing we may have to get it just for executing trades at Mirus and keep Trade Station for the analysis part.

 

And thanks again for the kind words.

[Soultrader 12]

TRO did share his indicators with the forum. I suppose Urmablume's will remain a secret although the concepts I suppose can be programmed by anyone with coding skills.

 

I find value in Urma's posts, for example I am quite interested in exploring more on tick charts. Lets say by using a 100 tick chart, I can simply program to count how many bars were drawn in lets say a timespan of 1 minute... this would somewhat be similar to what Urmablume is sharing?

 

I was just brainstorming right now of an historgram like indicator... not sure how effective this could be but perhaps someone else can expand on it? Thanks.

[Hlm 10]

To get back somewhat on topic...

 

UrmaBlume, since your indicator/analysis relies greatly on speed I would highly suggest posting a short video showing several (since they only last a second or so) of these spikes thoughout the day. You could even stream your charts live for a day on ustream.com for free. It would also be helpful if you could post an intraday chart (full rth day) showing where these spikes occurred so an individual could get the bigger picture. These would be much more helpful to everyone. Thanks.

[bakrob99 11]

...I was just brainstorming right now of an historgram like indicator... not sure how effective this could be but perhaps someone else can expand on it? Thanks.

 

Hi James,

Have you created an indicator for that? If so can you share the ELD?

 

Thanks

bakrob99

[Soultrader 12]

Hi James,

Have you created an indicator for that? If so can you share the ELD?

 

Thanks

bakrob99

 

Oh no, Im not a programmer and I currently dont use TS anymore. But I suppose it wouldnt be too hard to code up for someone with programming skills?

 

If I can code this up in CQG I will share the code. Thanks.

[Head2k 10]

Lets say by using a 100 tick chart, I can simply program to count how many bars were drawn in lets say a timespan of 1 minute... this would somewhat be similar to what Urmablume is sharing?

 

I was just brainstorming right now of an historgram like indicator... not sure how effective this could be but perhaps someone else can expand on it? Thanks.

Or it is possible to plot simple time histogram to 100 tick chart, for example. I played with tick and CVB charts with time histograms when I was looking for a way to understand price movement. Hovewer, later I found that my brain easier processes or understands simple time interval charts with volume histogram. Anyway, you still get the same information, that is activity over time, or time over activity.

What is different in UrmaBlume's approach is imho just a speed of evaluation and hence a speed of his reaction, too. He doesn't wait for 100 tick, 100 CVB or 5 sec bar to finish. That's why the need for a low latency precise data feed. But the principle is still the same: A reversal on high activity.

[honvly 10]

A couple of months ago we investigated the possiblity of changing to Ninja and ZenFire data. We were able to speak to management at both Ninja and ZenFire and found the ZenFire data stream to be great but that the way Ninja handled the data stream made the data not suitable for our purposes

 

Thanks for the great information. Can you go into details about how Ninja handles this data stream making it unsuitable?

[taotree 10]

Thanks for the great information. Can you go into details about how Ninja handles this data stream making it unsuitable?

 

I don't know if this is why but I think I saw UrmaBlume or someone else mention in another thread that Ninjatrader changes the timestamp that comes in from Zenfire so you no longer get the microsecond granularity. That might be what he's referring to.

 

Not sure why it matters. Zenfire has an API you can interact with directly so any shortcomings of Ninjatrader are irrelevant. Doing the level of analysis that UrmaBlume is doing, I would just go direct with the API. Then you don't have to be stuck in some charting vendor's paradigm of how they think you should be analyzing the market. Paint anything you want (I prototyped a fun zoom panel you can zoom in and out of the market ad infinitum) however you want. You want higher timeframe data together with other data? Trivial in any way, dimension, etc. In fact, you could even define a non-linear time axis. To paraphrase Scotty on Star Trek: "Candlesticks? How quaint".

[UrmaBlume 10]

To paraphrase Scotty on Star Trek: "Candlesticks? How quaint".

 

Perfect. Very Well Said. LOL.

[mickeybh 10]

SoulTrader,

 

Thanks for the kind words. Not at all like the responses I got from one of your other, less informed, moderators.

 

You are right about volume analysis - it's all that really matters. After all it is volume that motivates price - Price doesn't motivate price and the passage of time does not motivate price, it is the occurrence of trade (buying and selling) that motivates price.

 

Knowing that it is not the passage of time but rather the occurrence of buying and selling that drives price makes one wonder why anybody would ever base their trade on indicators with price as the prime input (EMA, Bollinger, RSI, Candlesticks, Stochastic, CCI....) and to further compound the error, apply those indicators to time constant data vessels.

 

Three decades ago I was taught that the best indicators of future price don't have price as any part of their calculation.

 

As I have mentioned I run a small PRIVATE trading and technology company and the first thing I teach our new traders is that it is buying and selling that drives price. That none of our intra-session charts are time constant charts and almost none of our indicators have price as ANY part of their calculation.

 

My comment to you is simply that buying and selling occur over time, so time is a factor. When the market is testing a top it may take 2 minutes to break through or 20.

 

Also the resting orders are extremely important as the market order hitting the bid is a buyer buying on the bid. sell 1 buy 1. When bids in a 5 tick range suck in 10 12 15 thousand contracts, where is the aggression. Buying or selling.

 

If you trade every ocurrance of volume spikes and change in Delta your bank account will go down.

[numbertea 10]

The ultra short term bursts of uber intense trade that often establish local intra-session extremes is never shown as resting orders it is auto-executed in a series of orders separated by only a few thousandths of a second.

 

The chart below shows a spike in trade intensity but while this high volume of trade was being executed it was never shown at any level as a resting order by market depth. This trade was auto-executed.

 

While commercial traders CAN successfully hide their method and whether the trade was an opening or closing transaction, they can't hide the fact of their transactions.

 

 

 

I use both bid/ask data as well as volume, trade events, and price movement. All these things impact future price. I trade the ES. I am very interested to hear about how you can possibly predict the market without taking price into consideration. Because of the tick size I had to specifically build my system for the ES. If I want to use my system on other instruments I would have to rebuild it with the other instrument price considerations. Any information about how to consider market sympathy without price data would be greatly appreciated.

 

Cheers