trendisyourfriend 

Although i do not use SC i have noticed some differences with my charts and Big Mike's charts. His MP charts differ from mines not by much but they do differ at times by 2-3 ticks on the ES.

That is indeed a problem when the same tool act differently on different platforms. This means that each platform is doing their own calculation, their own view of the tool.
For example, one that messes traders is the cumulative tick volume. SC have a complete different calculation when compared with IRT. This is very problematic. Cumulative bid/ask volume is identical but the tick version isn't? I remember ask SC staff about this and their answer couldn't have be more explicit - our calculations are the correct one, this is the way we feel the calculations should be done.

Now my questions is, which is the correct one? Apparently Gomi version for Ninja is identical to IRT. Can we say IRT and Gomi are wrong and SC is correct?

How can we trust a tool in which there's no consensus in it's calculations?

cusp 

You could if you knew how it should be calculated. A bit of hard work and a little research might prove it one way or another. Show them the truth and they'll change the platform overnight, but only if it is the demonstrable truth.

Which truth, mine or theirs?

I don't use delta based on ticks, only based on volume, so it really don't bother me the differences between platforms, but I know a couple of traders that do use both and they continue to stick with IRT and MD because, according to them, they are the ones that are calculating it correctly.

So here's an example since I still have SC available for the next 2 weeks.






EDIT: Why bother cusp?

I thought we were having here a healthy discussion about SC and it's pros and cons.
We might help people decide when they start to search for a new chart platform.

josh 

Well, when I tried out Sierra yesterday, I was blown away by how far it has come. I mean, with essentially the same setup in NT which takes about 200MB of resident memory, Sierra was humming along at 20MB!! Data loads pretty fast, and while there is a learning curve, in a matter of hours I had my charts almost duplicated, having learned the basics of the Sierra programming interface and already ported over two indicators. I just loved how this thing was working. Run multiple copies for different data providers, do anything you like. This is one flexible program, and honestly I love everything about it... except for.... .... the moment of heartbreak. I have the two running side by side, and for a direct comparison I connect both to IQFeed. Here is a comparison of the tape: T&S comparison: 2012-08-20_1022 - joshtrader's library You will be able to notice that Sierra (on the right) is delayed. It's only maybe 100-250ms. It doesn't sound like a lot, but even before I made the comparison to NT, my heart was a bit heavy as I watched the tape in Sierra. It looked as if it just did not handle quick bursts very well. The bottom line is that it just is not as quick. To be sure, I disconnected both, connected Sierra first, and THEN connected NT (thinking maybe order of connection might be prioritized), but same deal--NT simply is faster.

What you describe here may partially be due to Space-time trade-off.

Space?time tradeoff - Wikipedia, the free encyclopedia

Basically, many tasks have an inverse relationship between CPU time and memory usage, so as CPU usage decreases, memory usage increases, and vice verci. Depending on the needs for the software either can be exploited, so its an engineering decision. There is also the capability to store things on the hard drive rather than system memory, but this typically has slower performance.

With memory being so cheap these days I would usually choose faster CPU performance, but its a personal preference really and depends on the application. Usually you won't run into neither cpu nor memory issues on a mid-range PC purchased today but for some applications you still test the limitations of the hardware. I was a research assistant for awhile and we would routinely use up all our 12 GB of memory on our local PC's running image processing algorithms for example so we had to either use a high-performance computing cluster (and wait for everyone and their dog to finish their previously queued jobs) or "engineer" our algorithm to fit within the memory constraints.

Anyway, just my two cents.

trendisyourfriend 

Although i do not use SC i have noticed some differences with my charts and Big Mike's charts. His MP charts differ from mines not by much but they do differ at times by 2-3 ticks on the ES.

Hello.

Maybe this was the reason?

Quoting 

The method used by SC and IRT/MD are almost identical with the exception that IRT compares 2 price levels above POC to 2 price levels below POC whereas SC compares 1 price level above and below the POC. The IRT site states that method A is similar to that used by esignal, which has been around for some time and I suspect lot of people use this method. Investor/RT - Answers* (LinnSoft.com) IRT also has Method B which is based on the CQG method.

gd lck

Arnie,

The deletion and the "why bother" reflected the lack of content in my post. In other words, why fall into the

"Honey, sorry I can't come to bed yet." ....
"Why dear?" ....
"Someone is wrong on the internet!"

trap. In that case I thought doing some work was more productive.


However it's nice to see that @slickiam (nice name) had done the extra work. I'll quote the post he referenced:

joeroma @ Sierra Chart Support Board

This issue keeps coming up but to understand it I think you have to appreciate the history of MP and realize that MP is based on statistical theory. I addressed this issue last year and I will repeat it here. The method used by SC and IRT/MD are almost identical with the exception that IRT compares 2 price levels above POC to 2 price levels below POC whereas SC compares 1 price level above and below the POC. The IRT site states that method A is similar to that used by esignal, which has been around for some time and I suspect lot of people use this method. Investor/RT - Answers* (LinnSoft.com) IRT also has Method B which is based on the CQG method. A few years back I asked Don Jones of Cisco-Futures why he compares only 1 price level above and below the POC whereas Dalton's book (Don's son is one of the co-authors) said to compare 2 prices above and below the POC. He said that, when MP was introduced in the 80s for the bonds, there was a problem when they tried to apply it to the grains because of low volume at some price levels. Consequently, they decided to use 2 price levels. However, he stated that because there is plenty of volume with the es, ym, etc., the most appropriate way is to use only one price level above and below the POC. There seems to be different methods floating around and people can choose whatever they want. Personally I'll stick to what Don taught me since he was involved with MP from early on with Pete Steidlmayer and it is more consistent with MP theory and my understanding of statistical theory. This is the method that SC uses. However, I think that in the long run it probably will not make much of a difference which method you select since, as we all know, trading is not an exact science. When many of us select an area where we plan to enter a trade that area may encompass the VAH/VAL of different methods. Indeed, I have noticed that more often than not these different methods seem to produce values that differ only by a few ticks. The only exception seems to be on bigger range days when the differences tend to be significant. This is why I also look at vol-based VA and tend to place more weight on vol-based value areas. Hope this helps.

On that basis I don't think the Sierra guys will change it


Personally I don't use market or volume profiles. I've looked at them but find multi-timeframe supply and demand zones give me a better trading tool. I do have a fair bit of customization and some automation and some trading direct from Sierra. Like others I also use more than one copy. For me Sierra Chart has been a very satisfactory product.

Silvester17 

this is supposed to be a market profile (not volume profile) ,es, 7/27/2012, only rth, 3 min chart

That "indicator" was rounded to a 5 tick setting... This is why it is appearing to be higher than price. In reality, price was higher than the VAH. The referenced lines were simply well... reference points. MP isn't just about the Value and POC areas... There's nothing wrong with Sierra's indicators.

I find this rather comical...

josh 

As if by magic, or perhaps divine inspiration, I found a setting in Sierra under "General Settings" that allows the user to change the chart update interval. I changed this from 500ms to 1ms, and look at her go now!!!! T&S, NT on left, Sierra on right: 2012-08-20_1130 - joshtrader's library Sierra DOM: 2012-08-20_1133 - joshtrader's library Chart: 2012-08-20_1156 - joshtrader's library So it appears that Sierra is actually faster than NT, not the other way around as I originally thought. NT also has an update interval but the minimum is 100ms, FWIW. Feeling quite the doofus right now, I ask myself: do I have the courage to try Sierra for real? The answer is YES! I will try it this week on sim, and see how it goes. Still lots of learning to do, but as it stands, Ninja has to show me something pretty spectacular (note that I do own a ninja lifetime license and wish I could have that option for sierra).

Sierra is an excellent platform and far superior than most available. I left NT quite a while back and could not be happier with Sierra. I was evaluating between Market Delta and Sierra and went with Sierra for a variety of reasons.

Private Banker 

Sierra is an excellent platform and far superior than most available. I left NT quite a while back and could not be happier with Sierra. I was evaluating between Market Delta and Sierra and went with Sierra for a variety of reasons.

PB, do you use the sierra DOM? Could you take a screen shot, and anyone else who uses it? I am experimenting and have this currently, but am looking to eliminate that big space on the bottom left and fill it with something more useful.

Private Banker 

That "indicator" was rounded to a 5 tick setting... This is why it is appearing to be higher than price. In reality, price was higher than the VAH. The referenced lines were simply well... reference points. MP isn't just about the Value and POC areas... There's nothing wrong with Sierra's indicators. I find this rather comical...

well, like I said, I saw something strange and only wanted to know if anybody made some comparison with other platforms. I never said it's just about value or poc, nor did I say that sierra's indicators are wrong.

but thanks anyway. aes isch hoechschti isaebahn fuer mi.

Ninja Trader's Customer Service is spectacular, the moment you own or lease the license, even in the evening or on the weekend. I often had to wait days to hear back from SC.

ETA: NT needs to catch up in the Market Profile area, and maybe they will in the next release.

The fact that so many brilliant coders use NT and generously share their gifts means it's a party every day.