Futures Arbitrage

[AgeKay 10]

DugDug, I agree, his attitude might offend some people, but I think I know where it comes from. According to him, he's an old guy, he's been trading over 30 years, he's actually met the inventor of the Market Profile and slept under one roof with him, he's seen it all and I think he is just sick of all the bullshit (as am I though I am still very young) out there. You can call it bitter, but the fact is that he could point out the flaws of every type of information that retail traders base their decisions on using reason and logic. And if you've read some of his posts where he goes into detail, you'll see that he knows what he's talking about. He calls others idiots because he can see all the flaws but these people refuse to acknowledge them. Sure, some people still manage to be profitable, but that is not because of the flawed information they use, but despite of it.

 

Instead of being offended, people should use his criticism to question their beliefs and they will find that their are mostly based on what other people have presented (usually using anecdotes, flawed logic and after-the-fact charts) and not their own thinking. I could give examples but that would just start another flame war. Like you said, it's like religious debate, it's very emotional and you don't get anywhere. After all:

 

"Arguing on the Internet is like competing in the Special Olympics. Even if you win, you're still retarded." Source: "Special Olympics" entry on UrbanDictionary.com

[AgeKay 10]

Well, this "thing" I am talking about is exactly what UrmaBlume calls ''trade intensity". I have a different explanation for why he sees these spikes, but fact is we do mean the same thing and it's definitely done by computers that have a very good reason to enter there.

 

I'd like to take that back (as I can't edit my post anymore). His explanation is actually the same as mine. His explanation in earlier posts were a bit vague which I incorrectly interpreted but I am catching up on his new posts where he does explain it exactly the same way that I see it. Sorry for misrepresenting your words, UrmaBlume.

[Kiwi 16]

DugDug, I agree, his attitude might offend some people, but I think I know where it comes from. According to him, he's an old guy, he's been trading over 30 years, he's actually met the inventor of the Market Profile and slept under one roof with him, he's seen it all and I think he is just sick of all the bullshit (as am I though I am still very young) out there. You can call it bitter, but the fact is that he could point out the flaws of every type of information that retail traders base their decisions on using reason and logic. And if you've read some of his posts where he goes into detail, you'll see that he knows what he's talking about. He calls others idiots because he can see all the flaws but these people refuse to acknowledge them. Sure, some people still manage to be profitable, but that is not because of the flawed information they use, but despite of it.

 

Instead of being offended, people should use his criticism to question their beliefs and they will find that their are mostly based on what other people have presented (usually using anecdotes, flawed logic and after-the-fact charts) and not their own thinking. I could give examples but that would just start another flame war. Like you said, it's like religious debate, it's very emotional and you don't get anywhere. After all:

 

 

 

I find him almost completely offensive. He's an arrogant old sod although he sucks a few in along the way.

 

So what if he's met Mr S. He then proceeded to insult him telling us how much better he was than S. And he kept coming back posting the same advertising pictures from his website.

 

He's not sick of the bullshit ... he is the one promulgating it. Classic hershey style. And I guess there'll be another one pop up in 2011 of 2012 - PT Barnum was right then and they keep popping up whether they're in the good old US of A or Nigeria. Kind of fun to watch though when if don't care about the suckers. Like watching the pick pockets work the crowds in Venice. Locally, he truly seems to have sucked a few people into his particular grail sale. Enjoy

[AgeKay 10]

Well, he definitely sucked me in, but unfortunately he's not offering to sell anything, is he?

 

Instead me and others have actually found the information he has shared so far to be very useful and I thank him. Call me a fan boy or whatever, but I wouldn't even be visiting this forum any more if it wasn't for him because I find all the other information completely useless.

[UrmaBlume 10]

Well, he definitely sucked me in, but unfortunately he's not offering to sell anything, is he?

 

Instead me and others have actually found the information he has shared so far to be very useful and I thank him. Call me a fan boy or whatever, but I wouldn't even be visiting this forum any more if it wasn't for him because I find all the other information completely useless.

 

AgeKay,

 

Thanks for the kind words.

 

There is indeed a dearth of original material on this board.

 

After receiving hundreds of requests to buy different bits of our technical base I find that virtually all of my detractors and stalkers fall into one or both of two categories; 1) people who have tried to buy our gear and were refused and 2) people who have never posted any original work, can't deal with the technical discussion so they get personal and those whose posts are mostly knocking other people's posts.

 

I post here for 2 reasons - to try and get the brightest of the up and comers to think outside the box because that's where success is found and also to try and spot talent.

 

We have just started another trading entity which is a hybrid hedge fund/prop shop and almost all of its traders are members here. They came to us because of the distinctions you noted between some of our trade decision support technologies and those mentioned in other threads. Each of these traders operate from a workstation that drives 8 monitors and none of those screens contain anything from threads like VSA, candles, profiles or Wyckhoff.

 

I have worked in London and always enjoyed it. We used to trade from a desk in Finsbury Circus but that was 25 years ago.

 

Anyway, AgeKay, thanks again for the support and the kind words and if there is ever anything I can do for you, I am usaully up on Skype or I have direct lines to Europe and I would be happy to give you a call.

 

cheers

 

UB

Edited April 24, 2010 by UrmaBlume

[Kiwi 16]

And I repeat. PT Barnum was right.

 

And, as I always recall when I finally withdraw from one of the "catch the sucker" threads and let it continue, people like Thalestrader, Brownsfan, and myself all profit from the suckers. So, why do we keep defending them ... why not just let evolution act?

 

Enjoy

[BlowFish 10]

AK, been away for a long weekend so playing catch up.

 

I guess I am sometimes 'harsh' with UB, that is not because of any lack of appreciated for the ideas he has presented. I just feel that sometimes their value is diminished (actually obscured is probably more accurate) due to hubris (which I try not to let bother me) and hyperbole (which I try to point out as this will confuse). Anyway there is useful information as we discussed by PM when it first was presented. I always pay attention.

 

I find your new interest in trade intensity intresting, particularly your comments about programmatically detecting this behaviour. I know you stepped back from 'charting' dom stuff in favour of watching the raw data. This seems like it is one of those 'patterns' that will always be hard to see raw (of course I struggle to see anything much raw as I mentioned elsewhere). Fortunately I have other tools to determine trade location but am still messing with tools to detect shifts in flow.

 

As an aside did you ever look at marketdeltas footprint charts as a way to visualise this sort of stuff?

[TheDude 10]

how do I delete a post?

[TheDude 10]

Nate,

 

What a great post. Thanks.

 

All of our posts are about what you call point & click trading and that is done from TradeStation.

 

We have never posted anything about the concepts, applications, data and server location that we consider valuable in HF Algorithmic trading and probably, like everybody else, including you, never will.

 

Your view of the markets is of a perspective that is born of both experience and a great understanding of state of the art protocols and technologies and all here can benefit from that.

 

thanks

 

cheers

 

UB

I'm confused. Mind you, I'm not as clever as you clearly are, so please excuse me...

 

....but when you say 'we', who are you referring to? I thought all your secret buddies from GS who are members here, never post. That is what you stated earlier is it not?

 

I liked your first few posts in this thread though. Especially when you describe how you put your PC's in a circle and call it a trading pit! LOL! I did chuckle. Along with the description of your hybrid company 'hedge fund/prop shop'. Your posted chart with stops on a scalping strategy made me smile too. Dear oh dear!

 

You're good value UB. Do keep it up!

 

Thanks,

 

Duderino.

[AgeKay 10]

I find your new interest in trade intensity intresting, particularly your comments about programmatically detecting this behaviour. I know you stepped back from 'charting' dom stuff in favour of watching the raw data. This seems like it is one of those 'patterns' that will always be hard to see raw (of course I struggle to see anything much raw as I mentioned elsewhere). Fortunately I have other tools to determine trade location but am still messing with tools to detect shifts in flow.

 

The initial prototype I programmed wasn't really helpful but I have to admit my implementation wasn't that good. Right now I am just sticking to see it by watching the DOM. It's not as hard as you might think. All trade intensity is is a lot of volume in a very short amount of time. Of course, there is no spike to look for, but if the inside market is bid 300 and all of the sudden you see that 2000 contracts have traded on that bid very quickly and it's still bid 300, you know that some program was just waiting there to get hit on the bid and absorb all market orders. No human being could have refreshed his bid so quickly. It's as if someone had entered an iceberg order there, but knowing that Eurex doesn't support iceberg orders natively, I had to conclude that these are "synthetic iceberg orders" which are just limit orders that keep being refreshed by programs. It really clicked when I saw this post by UrmaBlume:

 

The formation of many local extremes, especially in the futures and options on the equity indexes, is a result of the bid or asked being replenished by auto placement at a rate faster than the market can absorb. Thus action at the top would not reflect this selling and show a top/reversal made on buying instead of the selling that is represented by this auto replenishment of the asked.

 

--------------------------------------------------------------

 

As an aside did you ever look at marketdeltas footprint charts as a way to visualise this sort of stuff?

 

Yeah, I trialed them before I even started programming my own charts but I don't recall them having any charts that could visualize this because they don't incorporate the time factor in their footprint charts. After having watched the DOM now though I realized the "flaw" of their buying vs. selling designation what I pointed out earlier in the "Volume Splitter" thread or what UrmaBlume also describes in the same post referenced above:

 

We feel that the bid/asked approach to designating buying or selling volume is flawed in several ways and the synchronization, or not, of these feeds is not part of our calculations.

 

To demonstrate one of those flaws, please consider the situation where a very large buyer places a limit order at the current asked price - the order is larger than the asked size so he is partially filled at the asked and as the order took all the asked size the asked moves up and so does the bid. Now the remainder of this large buyer's order is on the bid and trade that fills this large buyer, on his bid, is recorded as selling volume. Thankfully there are intelligent agents capable of sorting all this.

[BlowFish 10]

One of the more interesting visulisations that I played with was using a time histogram under a constant volume chart. Similar to trade intensity except sampling the time with respect to a certain volume. I posted a couple of charts showing extremes being picked off in the trade intensity thread.

 

The replenished bid/ask is something that can be tricky to see as turns often tend to be as much to do with buyers retreating as sellers stepping up. I remember Jack Hershey talking about 'walls' on the DOM years ago, and I think Rollo Tape (Whyckoff) might even have talked about it back in the 30's (would need to check). I actually wonder if these things are easier to see by focusing on T&S and a glance at the dom rather than simply staring at the dom?

 

As to the behaviour that UB describes (large limit hitting the best bid/ask) this is precisely the sort of event that can cause the L&R algorithm to mis classify trade direction and is discussed in papers that I talked about elsewhere. There is an argument that that buyer is now passive if he is prepared to sit at best bid rather than chase the ask. It is safe to say that they are now providing liquidity rather than taking it. There is no 'perfect' way of classifying volume.

[AgeKay 10]

I remember Jack Hershey talking about 'walls' on the DOM years ago, and I think Rollo Tape (Whyckoff) might even have talked about it back in the 30's (would need to check).

 

I guess you mean large depth with 'walls'. I don't know this Hershey guy, but I've read Rollo Tape, they didn't have something like a DOM back then. All they could do is to call the broker to find out how much depth was sitting there.

 

I actually wonder if these things are easier to see by focusing on T&S and a glance at the dom rather than simply staring at the dom?

 

You wouldn't be able to see it with T&S only, because you wouldn't know that when 2000 contracts traded that someone just simply bid 2000 with one order. The point of trade intensity is that you won't see the size in the DOM until it has actually traded. It will be an average bid since these guys don't want others to know that they just bought that many contracts. I don't even have T&S up on my screen.

 

There is an argument that that buyer is now passive if he is prepared to sit at best bid rather than chase the ask. It is safe to say that they are now providing liquidity rather than taking it. There is no 'perfect' way of classifying volume.

 

I don't think it really matters if someone is passive or aggressive. The end result is that both traders got in or out at the same price. In the case of trade intensity, these computers have a good reason to buy or sell at a certain price.

[zdo 20]

BF,

 

re :" 'walls' on the DOM "

If you have time and energy for it could you please expand and or illustrate this. Many thanks.

 

zdo

[become 10]

One of the more interesting visulisations that I played with was using a time histogram under a constant volume chart. Similar to trade intensity except sampling the time with respect to a certain volume. I posted a couple of charts showing extremes being picked off in the trade intensity thread.

 

The time histogram can be quite close, and certainly a good trader / reader of charts can use it to 'see' pretty much the same activity. The 'trade intensity' indicator is obviously more precise at identifying the activity AgeKay speaks of, though it certainly isn't the 'grail' - what that activity means for future direction still has to be interpreted in context. Perhaps this is different on other instruments; I mostly trade the ES.

 

I've been using the intensity indicator since last Fall, but only on a faster chart - on my slower volume bar charts I use a time histogram. I also use 2 different instances of the indicator, set at different 'speeds'.

 

HTH

[BlowFish 10]

I guess you mean large depth with 'walls'. I don't know this Hershey guy, but I've read Rollo Tape, they didn't have something like a DOM back then. All they could do is to call the broker to find out how much depth was sitting there.

 

No Dom, just an old fashioned ticker tape. It certainly deals with buying pressure selling pressure and even trade intenstity (though not called that!) "

 

If steel has just sold at 50 this figure represent what has happened. It is history. The market price of steel is either 49 1/8 @ 50 or 50 @ 50 1/8. The bid and asked prices combined form the market price.

 

This market price is like a pair of scales and the volume of stock thrown out by sellers and reached for by purchasers, shows towards which side the preponderance of weight has momentarily shifted."

 

I have to say I mis remembered what it was all about (and probably confusing it somewhat with Humphrey Neils book aroud the same time) but Whyckoff absolutely had a grasp of this stuff and wrote about it 80+ years ago. You might see more in it with your current understanding of the markets .

 

You wouldn't be able to see it with T&S only, because you wouldn't know that when 2000 contracts traded that someone just simply bid 2000 with one order. The point of trade intensity is that you won't see the size in the DOM until it has actually traded. It will be an average bid since these guys don't want others to know that they just bought that many contracts. I don't even have T&S up on my screen.

 

This is partly a reporting issue, CME changed end of last year. Do you report with respect to the limit side or the market side? Sure you can't see the book on the tape but you can much more clearly see 2000 1 lot orders go by with virtually the same time stamp. The thing with the tape it is arguably easier to see stuff as it 'scrolls'. With the dom 3 things are updating bid ask & traded and each over writes itself. Makes it hard (for me at least) to see whats going on. Using time & sales (for it's permanent record) in conjunction with BB BA on the dom helps me a little but I am still far from proficient in assimilating the info in real time. Hence the appeal of 'visual aids'.

 

 

 

I don't think it really matters if someone is passive or aggressive. The end result is that both traders got in or out at the same price. In the case of trade intensity, these computers have a good reason to buy or sell at a certain price.

 

I agree, from the point of view of trade intensity the direction of the order flow (passive/aggresive being the proxy for order flow) matters less than for other 'patterns'. That is not to say that direction is not important at all. Having said that one would expect a shift and a change in intensity at turning points.

[AgeKay 10]

I have to say I mis remembered what it was all about (and probably confusing it somewhat with Humphrey Neils book aroud the same time) but Whyckoff absolutely had a grasp of this stuff and wrote about it 80+ years ago. You might see more in it with your current understanding of the markets .

 

I've read all those old books that fall under the "tape reading" category including Neils' book. I haven't really gotten anything out of these books. One of my charts I am still using though is based on Whykoff's "Volume Figure Chart" in Chapter 8 of "Studies in Tape Reading" (also floating around the net with the ridiculous title "The Day Trader's Bible" )

 

Sure you can't see the book on the tape but you can much more clearly see 2000 1 lot orders go by with virtually the same time stamp. The thing with the tape it is arguably easier to see stuff as it 'scrolls'.

 

If it's bid 2000 and you see 2000 1 lot orders, then I wouldn't classify that as trade intensity because these computers won't show their size and they will be refreshing the bid, not hitting it.

 

With the dom 3 things are updating bid ask & traded and each over writes itself. Makes it hard (for me at least) to see whats going on. Using time & sales (for it's permanent record) in conjunction with BB BA on the dom helps me a little but I am still far from proficient in assimilating the info in real time. Hence the appeal of 'visual aids'.

 

If the market is going down, all you have to look at is the bid and the Volume At Price (VAP) at the bid. In MD Trader it's just another column that you can move around so you can move it next to the bid or ask column. So all you really have to look at is one row made up of 2 cells (bid and VAP or ask and VAP). If the depth of the bid is 300 and the VAP at the bid price is 3000 and 1 second later the bid is still 300 but the VAP is 5000, you know that it has traded 2000 contracts at the bid without it going offer.

[BlowFish 10]

 

 

I'd like to see that documentation. But you always have to know your tools no matter what technology or language you use. All myths aside, managed languages like C# are in fact faster than unmanaged languages like C++ in most real world applications because the runtime knows the system its running on (where it can do system specific optimizations) and it knows the resources available on the system. Especially in trading applications where processing isn't constant but dependent on the amount of data at that time, it can be much faster because the garbage collector can suspend the freeing of memory in favor of processing the data faster during peaks of data and then when it notices that the CPU isn't used that much, it will free the memory, thus not affecting the processing.

 

Common problems like memory leaks are almost impossible in .NET unless you really don't know anything about it.

 

 

Kind of brutally snipped the original post found here http://www.traderslaboratory.com/forums/208/futures-arbitrage-7639-13.html#post94874.

 

I was going to start a thread on development environments and managed vs unmanaged environments would have been one of the key discussions. Still a bit premature as I have not decided whether to 'roll my own' yet. I have however been giving it some serious thought the last few days. Yesterday I had a few hours to kill and went into Waterstones (a UK book store) and spent the time reading Wrox's Professional C# 4.0 .Net 4. On the whole I was impressed.

 

I have to say I am tempted to go managed. There is just something niggling in my mind that if I ever do want to scale up performance .NET will hit a wall way before an unmanaged approach. The thing is if that wall is a million miles away from what I need who cares? For example if I want to calculate the delta for every one of the constituent S&P 500 stocks (I don't but might want to one day) I would make an educated guess that .NET might start to creak at the seams. As I have said in this thread (rather off topic really...sorry) I am an old dinosaur, writing a simple framework to manage stuff (or pinching elements from the public domain) is not that big a deal.

 

Anyway I re read AK's post linked above and he makes a compelling argument for managed above. One of the things I liked the sound of was a new chart type in a few hours but wonder if that is because of a good library of controls or because of .NET? I'm pretty sure you could have done that in Borlands old stuff too .

 

Anyway I rambled off a bit, really I just wanted to offer this link that has quite a few hints and tips for writing faster managed code. It does make you think that managing your own stuff is actually in some ways easier than trying to optimise managed code. Of course if you aren't hitting any barriers then the whole argument is moot.

 

Writing faster managed code

[MidKnight 12]

What gets you with the unmanaged stuff is when the GC kicks in, you usually get a noticeable stall in performance. If your code is continually allocating memory and also marking references to null just as often, you are gonna get the GC kicking in quite often and it wouldn't be unusual for it to kick in when you don't want it to Damn Murphy!

 

With kind regards,

MK

[AgeKay 10]

One of the things I liked the sound of was a new chart type in a few hours but wonder if that is because of a good library of controls or because of .NET? I'm pretty sure you could have done that in Borlands old stuff too .

 

Don't want to take this off-topic, just want to comment on this one. I think it's both. .NET is two things: The run-time (languages) and the framework (library). The run-time is updated every few years and its mainstream languages (C# & VB.NET) tend to incorporate the best ideas and functionality from new innovations and old languages. Especially in the latest releases, many functional and dynamic language features have been integrated into the mainstream .NET languages which make them even more productive. I think just the C# 3.0 release made me 2x more productive compared to C# 2.0. In addition, it supports pretty much every popular language ever used which can all share the same code base and access the framework. The framework is massive and integrates pretty much every popular technology. I've been working with it from the beginning and I know maybe 5% of all classes (maybe only 1%, who knows). You can develop web sites (ASP.NET, Silverlight), databases (LINQ to SQL, Entity), rich client applications (WPF, WinForms), networked applications (WCF), mobile device applications (.NET Compact Framework, Silverlight), games (XNA) and cloud applications (Azure) and much more using the same development environment, code base and framework. The leverage you get from that is incomparable. For example, my charts are using WPF (Windows Presentation Foundation) which is a library to develop UIs. It maintains all state (retained mode) and it's using DirectX (what games use) to render the graphics so everything is hardware accelerated (graphics card does all the work instead of the CPU). So out of the box you get the easiest to use and fastest graphics API. This makes me 100x more productive than the next guy using some native graphics API (GDI or DirectX). Using my own little charting abstraction that is based on WPF, it takes me about 15min to create a completely new chart type from scratch if I knew in advance which color scheme I want to use and it's performance is faster than anything you've seen (think how much is going on in games nowadays compared to those few stupid lines and rectangles we are drawing to render a chart). Even when I had my 5 charts open, my CPU was always pretty much idle (using 1% maybe) even though there could would be tens of thousands of objects on my screen.

[BlowFish 10]

Thanks AK. Yeah I should have started a new thread but what the hey.Going to have a play with C# seems like you can use unmanaged code for 'back end' stuff if the need arises. Seems like some synchronicity with VS 2010 just hitting the streets. It'll probably come to naught so much to do so little time. Now if you can only connect to a data provider api in 3 hours too hehe .......

[johnnyh 10]

I make my living spreading different markets. I spread the curve, ranging from Eurodollars all the way through 30yrs. Of course there are different hedge ratios and relationships with every spread, but in short, yes I spread different futures markets. The cme now even has implied spread markets for the treasuries ie NOB 10yr vs 30yr. I would certainly be open to speaking more about this topic.

Good luck!

[BlueHorseshoe 10]

Rhetoric usually employs three audience appeals, ethos, pathos and logos.

 

Someone's been reading Northrop Frye . . .

 

BlueHorseshoe