Trade Intensity

[DbPhoenix 42]

And last but not least, I am a beginner so what do I know?

 

Beginners often see with a clarity that others do not.

[Eiger 10]

...These traders go to great lengths to disquise their trade - they have automated routines to send a series of small transactions to the market at intervals close to 1000th of a second.

 

Because many of these traders operate from price based models they must get their execution within a very narrow price/time range which means they must execute as much of the volume they want as fast as they can while price is in that range. This results in a huge spike in what we call the intensity of trade ...

 

 

Facinating work. Thanks for posting this. Have you thought about how these traders know where to trigger their high volume trades? They must somehow know where liquidity is pooled, otherwise, (it seems to me) such high volume orders would tend to rocket price and they would be putting price up or down against thier interests. They must also have a sense of the amount of liquidity in their price area and whether or not it could be drained by their high volume orders. If there was too much liquidity price would quickly go against them. It is sort of the obverse of what you have developed. Does your system look for these areas in advance? Would this even be possible?

 

Eiger

[UrmaBlume 10]

Eiger,

 

Thank you for the kind words. We are a small group of private traders, programmers and poker players. We are pround of the work we do.

 

As to your questions "how these traders know where to trigger their high volume trades?" - Most of this high intensity is triggered in 1 of 2 ways - 1) a certain level of premium is present in either the options or the futures and a barrage of basket trades or other premium arbitrage trades are triggered and 2) Buy and sell prices generated by artificially intelligent and genetically optimized trading models are met by the market and the trade is executed in miliseconds.

 

Often it is not required that one seek "pools of liquidity" as, to a certain extent, there are very fast automated routines that both disguise size trade and at the same time stimulate activity and trade flow.

 

Further as to the question of liquidity, besides the automated trading/execution routines, we use certain applications such as the Heads-Up display shown below. This app is further described in separate post but beside measuring the power of surges and buying and selling, it helps to spot places where size is more easily executed.

 

[UrmaBlume 10]

Nite session spikes in trade intensity are very useful in determining commercial intent. During the nite session there is so little retail trade that when the commercials get active it is very obvious, as in the graph below, and often serves as a key reference point during the next day's trade.

 

The spike show below happened about four hours ago, times are PST.

 

[BlowFish 10]

 

Often it is not required that one seek "pools of liquidity" as, to a certain extent, there are very fast automated routines that both disguise size trade and at the same time stimulate activity and trade flow.

 

 

I think this bears some examination as it seems like this is one of your fundamental premises. I would contend that if you want to trade size you require liquidity or you will put the price up on yourself regardless of how you slice and dice. If you squirt 1000 cars at the market 100 every .1 of a second or if you send it all at once it will chew through any limit orders that are resting on the book regardless of whether they show or are icebergs. Further more if there are same side orders (stops) resting alongside them they will trigger competing with the algos. As for slicing and dicing hiding activity, the very fact that most peoples tools don't have the granularity to see the slices or they haven't set them up to doesn't that have the effect of 'unslicing' the salami in their eyes?

 

Incidentally I guess you still need a 'threshold' to be reached (you can't distinguish if its a large sausage coming down the pipe until you get a few slices you dont know how large until you get all the slices) so provided your granularity is such that you can detect this threshold I would contend you don't need wizzbang infrastructure (though that can't hurt). This was one of things that seemed somewhat exaggerated in your original post. No need to exaggerate, your work appears to stand up without. btw Whose data feed are you using and where is your setup located? Geographical location and local loop technology (if any) are the biggest contributions to latency, whole orders of magnitude larger than anything else.

 

I think most would agree that looking at trade intensity and gauging order flow over a short time period can provide an edge (I know you don't use constant time periods ....except on the HUD display it appears you do, as you seem to for normalisation, but that's a whole other discussion) Having said that I do believe more traditional tools used in innovative ways can provide a similar edge. I guess you do as it seems you use Tradestation, if not in production, for prototyping and research.

 

A specific on the trade intensity. Is this essentially looking at order velocity? It seems to me if you are sampling data by volume (or price change i.e. range bars for that matter) that looking at changes in the velocity of that volume arriving would give a similar indicator. Smooth it a bit and I think that might be similar to what you have, or am I way of base here? Its something on my back boiler for further research.

 

The work you are doing certainly looks innovative and interesting but I can't help feeling that you are 'bigging it up' somewhat:) Having said that I can certainly understand why you are proud of your baby, it looks impressive. I hope you take this in the spirit of provoking discussion as requested

[UrmaBlume 10]

Blowfish,

 

Thanks for the kind words and the rational discussion.

 

In these posts I am trying to walk the fine line between stimulating the conversation, providing clues to new concepts and methods and not p*ssing off my partners by giving away the farm.

 

In reply to your first paragraph:

 

1) In the futures market the market depth means almost nothing. Size is almost never posted - It waits

 

2) Those who either want to disquise their trade or create liquidity do so by placing, according to a certain algorithm, orders on both sides of the market. In the old days they used to say "sell a little to buy a lot." Sometimes a quick burst of selling will trigger other selling and maybe even some stops which adds tremendously to buy liquidity.

 

Without regard to whatever trading concept, market or indicator, there is a very small, in number, set of size traders that, not necessarily in concert, set local bottoms and tops. Without some knowledge of where they are active and what they are doing puts other, smaller, intra-session traders at a huge disadvantage and no MACD, RSI, CCI, BollingerBands, etc, buschwa will overcome that disadvantage.

 

I am not saying that knowledge of this activity is the only way to trade successfully or that one can not profit without it. I am saying that most know nothing of when or where this activity occurs and that they would have a better chance at success if they did.

[BlowFish 10]

I wondered if point 2 was one of the conclusions that you arrived at. In the old days a large participant might hire half a dozen brokers to work a large order for them. To buy or sell there lines they would need to work both sides of the market, I guess there is no reason algorithms could not do that in micro time frames. I guess if you are talking about arbitrage trades based on things like fair value premium and such that is a somewhat different animal.

 

As an aside have your group looked at 'internals'. There used to be a time that you would often see tiki extremes (tiki is like the tick for the dow industrials) when programs triggered. Seemed clear that in those instances the things where just buying futures and selling baskets without to much subtlety.

 

I also wonder whether in darkened rooms people write algorithms to try and capitalise on other algorithms. I don't know if you work on the programming side or whether that is handled by your associates. When god was a boy and I was a programmer there was something called 'core wars' that a few of the nerds got into. Essentially battling algorithms http://en.wikipedia.org/wiki/Core_War. The (rather surreal) thought of battling algos reminded me of it.

[UrmaBlume 10]

I also wonder whether in darkened rooms people write algorithms to try and capitalise on other algorithms. I don't know if you work on the programming side or whether that is handled by your associates. When god was a boy and I was a programmer there was something called 'core wars' that a few of the nerds got into. Essentially battling algorithms http://en.wikipedia.org/wiki/Core_War. The (rather surreal) thought of battling algos reminded me of it.

 

Blowfish,

 

What an interesting analogy. At least part of what we do is exactly what you describe. Some of our technologies look to find high intensity, auto-executed, market activity and other bits and pieces work to execute trades in a timely enough fashion to take advantage of the information.

 

cheers

[bakrob99 11]

I'm pretty sure most of any latency I get under normal conditions is simply down to geographical location and the speed of light

 

I don't wish to quibble (but am going to anyway) but I think your 1000th of a second is probably off by an order of magnitude or two. Light would only travel about 3Km in that time for example

 

I think you have missed by an order of magnitude here. Light travels at a little over 186,000 Miles per seconds which works out to about 186 Miles in 1/1000 of a second or approx 300 Kms in a milli second (1/1000).

 

And a hard disk's head doesn't have to travel far... besides your O/S writes in a memory cache which is considerably faster again.

 

My only point is: your math is off...but I am not commenting on URMA BLUME's ability to determine if the trades are travelling at the "intensity" she claims. YOu may still be correct about that.

[JBWTrader 10]

UrmaBlume,

 

What I find very interesting is to ask what sets these (at times Disguised) volume "spikes" off in the first place. I think what you have captured is evidence that possible various uncorrelated/correlated strategies are firing off during the session.

No doubt this type of evidence/signal serves as a possible reliable reason to enter a trade. (although I dont know if you saw in a previous post of mine where I asked if such a signal has favourable win/loss ratio etc etc ie is there an edge to this piggy backing on such a signal...?)

 

Without giving the total game away have you gone into why such events occur in the first place ?

 

for eg one thing I have learned is that the markets are very, very precise on where they turn and why.

 

 

All the Best

 

John

[BlowFish 10]

I think you have missed by an order of magnitude here. Light travels at a little over 186,000 Miles per seconds which works out to about 186 Miles in 1/1000 of a second or approx 300 Kms in a milli second (1/1000).

 

And a hard disk's head doesn't have to travel far... besides your O/S writes in a memory cache which is considerably faster again.

 

My only point is: your math is off...but I am not commenting on URMA BLUME's ability to determine if the trades are travelling at the "intensity" she claims. YOu may still be correct about that.

 

You are right I read it as 10^−4 having said that the general point is still valid despite my :doh:, apologies, the eyes aren't as good as they used to be especially with strings of 0's. Actually maybe I didn't make the point well either The comparisons I made where from here http://en.wikipedia.org/wiki/Millisecond. I simply commented on the seek time not how often a disc needs to seek.If data is cached somewhere it does not need to seek, besides there are solid state discs for mission critical applications.

 

The fact remains that latency introduced by processing and application are going to be fairly insignificant compared to the latency introduced by geographical location and the local loop (where it is not fibre optic or you are not hosted in a data centre). btw that was why I asked about data feed and where the application was hosted.

[bakrob99 11]

UrmaBlume,

 

What I find very interesting is to ask what sets these (at times Disguised) volume "spikes" off in the first place. I think what you have captured is evidence that possible various uncorrelated/correlated strategies are firing off during the session. ...

John

 

Couldn't they be the result of PROGRAM TRADING ... Compare the events with the ESINX feed.

[JBWTrader 10]

Yes bakrob99 they are the result of programme trading , in fact I would take it further and say that all turning points are dictated by programme trading based on the same set of rules played over and over again.

Best

John

[UrmaBlume 10]

UrmaBlume,

 

What I find very interesting is to ask what sets these (at times Disguised) volume "spikes" off in the first place. I think what you have captured is evidence that possible various uncorrelated/correlated strategies are firing off during the session.

No doubt this type of evidence/signal serves as a possible reliable reason to enter a trade. (although I dont know if you saw in a previous post of mine where I asked if such a signal has favourable win/loss ratio etc etc ie is there an edge to this piggy backing on such a signal...?)

 

Without giving the total game away have you gone into why such events occur in the first place ?

 

for eg one thing I have learned is that the markets are very, very precise on where they turn and why.

 

 

All the Best

 

John

 

 

We find the signal to be profitable and that it comes in plenty of time to correlate with other indicators and then execute the trade.

 

The markets turn for only one reason - a change in the balance/imbalance in the order flow.

 

This indicator is designed to find short bursts of intense (usually automated) trade while certain other indicators indicate buying or selling intensity.

[deanz 10]

Here are UrmaBlume's first 3 posted charts shown here on NinjaTrader with Zenfire feed in 5 Tick charts and 1 tick...... the volume spikes are nothing out of the ordinary, looks like Ninja missed those trades, or maybe it just has a unique way of displaying them

 

UrmaBlume throw us over that fiber cable would ya.... yeah that one just sitting over there....

 

Dean.

 

[Charts are in EST Time and blue vertical dotted lines are placed every minute.]

[BlowFish 10]

Deanz could it be a times scale issue? Tradestation can label with exchange time or local time. Perhaps UB could confirm?

 

Zenfire is arguably one of the most complete and timely freeds available to the retail customer.

[UrmaBlume 10]

Here are UrmaBlume's first 3 posted charts shown here on NinjaTrader with Zenfire feed in 5 Tick charts and 1 tick...... the volume spikes are nothing out of the ordinary, looks like Ninja missed those trades, or maybe it just has a unique way of displaying them

 

UrmaBlume throw us over that fiber cable would ya.... yeah that one just sitting over there....

 

Dean.

 

[Charts are in EST Time and blue vertical dotted lines are placed every minute.]

 

I think you are missing the difference between a spike and volume and a spike in intensity.

 

You should know that Trade Station does not have the ability to measure time with enough granularity to do this work. We call certain dlls that do this work outside tradestation and then import and graph the results.

[Shamal 11]

UrmaBlume,

Is it possible to illustrate some realtime trading examples based on all this, taking yesterdays ES market or day before.

[bakrob99 11]

In most markets and certainly in the equity index futures markets there is a small set of very big traders whose entries into the market often have an immediate and very tradeable impact on price.

 

Thanks for these charts. They are interesting.

 

Could you also post a chart where the TRADE INTENSITY failed to show the correct direction.

 

Thanks

[monad 10]

kind of simple in my trading approach, all this sound pretty sophisticated stuff to me, a few solid concrete examples on an intraday chart would not be out of place.

[UrmaBlume 10]

kind of simple in my trading approach, all this sound pretty sophisticated stuff to me, a few solid concrete examples on an intraday chart would not be out of place.

 

Every thread I have started and most of the posts I have made include examples on intraday charts.

[Hlm 10]

Every thread I have started and most of the posts I have made include examples on intraday charts.

My guess is that the comment was geared towards providing something that's not a month old and one little snap shot out of the day. For example, maybe you could post a few times it happened today (Friday) or next Monday. Since one cannot just open a chart and load the indicator, it would be more helpful if a decent sized and unbias population was supplied for better discussion.

[UrmaBlume 10]

My guess is that the comment was geared towards providing something that's not a month old and one little snap shot out of the day. For example, maybe you could post a few times it happened today (Friday) or next Monday. Since one cannot just open a chart and load the indicator, it would be more helpful if a decent sized and unbias population was supplied for better discussion.

 

Below are 5 screenshots taken at different times during Today's trade. The times on the charts are PST and there were about 50 such signals on Friday.

 

 

 

 

 

[darthtrader3.0beta 10]

Zenfire is arguably one of the most complete and timely freeds available to the retail customer.

 

There is really a more obvious question that comes up with all this, I'm surprised you didn't notice it.

So I guess the thread starter has jacked tradestation to take a high frequency data stream other than tradestation? Something I asked about early on that was ignored. Considering we all know that TS is not a TICK data stream...let alone a high frequency data stream...

Which obviously brings up the question of why such a group of "poker players, programmers and traders" would even bother with tradestation to start with...I'm sure the fact that alot of newbs use tradestation which all this stuff sounds like some grail had no consideration to such a group...even though there are such better choices at the retail level, let alone rolling your own.

But then you have to start thinking about the structure of such a group...

If this was just to "brag" I would probly post a link to autumn gold..but then again I guess this group just all does their own thing with no entity...and no interest in OPM....

They dig down to this level, understand the stat arb guys come in at certian levels of futures premium but don't seem to bother with the cash market...

The real ironic thing though is I'm pretty sure this is deja vu and these idiots have posted this stuff before. Everyone called bullshit, so they came back with more "technicals"...

[deanz 10]

I think you are missing the difference between a spike and volume and a spike in intensity.

 

 

Contracts traded are contracts traded..... the total over the same time period have to be the same no matter how you view it..... don't they ?

 

I like to give people the benefit of the doubt but I think I'm with darthtrader, especially after visiting their website..... I think I smell a rat.

 

Dean.

Edited January 24, 2009 by deanz