Cot Report In Ninjatrader

[rretch]

I'm curious how you are using open interest? I haven't used it myself.

I'll give you the quick version.

If the DX, for example, is rising, and the open interest is rising, this indicates a bullish interest on the DX. Higher open interest numbers indicate more contracts being written between two people. If DX is rising, on the other hand, and open interest is falling, this indicates fewer contract writes and 'could' indicate a bearish DX.

DX price falling and open interest falling = sentiment is bearish in the near term since longs are selling w/o new contracts being written between two 'new' participants (indicating a falling open interest), but it may indicate a turning towards a bullish Dollar once the sells are done and new buyers come in that require writing new contracts.

DX price falling and open interest rising = sentiment may be turning towards a Bearish Dollar

that's the quick and easy of it.....it's much easier just looking at it on a chart.....

[cunparis]

Elliott Wave,

Also, with aCOT Proxy, is open interest considered in the indicator, or does anyone think it could be done? Think or Swim has Open Interest on the DX futures and it's a nice tool to have.

thanks,
-robert

I'm curious how you are using open interest? I haven't used it myself.

[rretch]

Elliott Wave,

Thanks for the indicator. At first glance, it looks like a tool that can be used along with $TICK as a reference....i'll give it a glance on the YM today (since I can also consider COT data).

I wonder if it could help in playing the gap?

Also, with aCOT Proxy, is open interest considered in the indicator, or does anyone think it could be done? Think or Swim has Open Interest on the DX futures and it's a nice tool to have.

thanks,
-robert

[Elliott Wave]

I'm definitely interested in trying out Bolsa's COT indicator. In the meantime I've been working on my own. It's not easy. What I meant earlier when I said I was skeptical.. an example is often the commercials switch sides 3-4 weeks early. So for me the COT gives the direction but not the timing. I'm currently looking at how to introduce a timing aspect which I believe Bolsa has found.

My idea is simply to use the COT as a directional filter and for choosing which markets to trade.

My other great( ) idea was to use the COT to select which currency pair to trade. The futures which the COT is based on are all single currencies, and by finding the currencies the 'big boys' are most long and short of, you can theoretically find the best pair to trade.

After trying to figure this all out manually on pen and paper I decided to wait for a COT indicator in NT and just backtest it...

[cunparis]

Good methodology for evaluating indicators btw. I doubt there is anything to this indicator either, but its something to play around with (maybe others can improve) while we wait to evaluate jrbolsa's indicators.

I'm definitely interested in trying out Bolsa's COT indicator. In the meantime I've been working on my own. It's not easy. What I meant earlier when I said I was skeptical.. an example is often the commercials switch sides 3-4 weeks early. So for me the COT gives the direction but not the timing. I'm currently looking at how to introduce a timing aspect which I believe Bolsa has found.

[Elliott Wave]

Do you lease the JMA or is it a one time fee? From what I understood of their website, which was not very clear, is that it's a lease for ninjatrader and one time fee for tradestation. I've considered purchasing it, so many people seem to like it.

For your oscillator, one way I test mine is to write a simple strategy and if profit factor is > 1 then I am interested, if it's < 1 then I'm not. I use a couple entries but it depends on the oscillator and how it works. Such as:

- enter when it turns up/down from an extreme
- enter when it crosses a moving average
- enter when there is a pullback on its way up/down
- etc.

From eyeballing your oscillator I don't see anything predicting turning points but it'd be good to have some proof. Thanks for sharing.

JMA is a lease. I think the Jurik stuff is actually worth it compared to most other 3rd party stuff, which I don't think has much merit.
You can find an article on the JMA with code here:

http://articles.mql4.com/310

Its on the official MetaQuotes site so I assume its non-infringing. You can probably convert it to NT if you want a real close but not 100% approximation. If you do convert based on that code its probably best to not share it as it could be viewed as infringing.

Good methodology for evaluating indicators btw. I doubt there is anything to this indicator either, but its something to play around with (maybe others can improve) while we wait to evaluate jrbolsa's indicators.

[cunparis]

Here is a different version.

Do you lease the JMA or is it a one time fee? From what I understood of their website, which was not very clear, is that it's a lease for ninjatrader and one time fee for tradestation. I've considered purchasing it, so many people seem to like it.

For your oscillator, one way I test mine is to write a simple strategy and if profit factor is > 1 then I am interested, if it's < 1 then I'm not. I use a couple entries but it depends on the oscillator and how it works. Such as:

- enter when it turns up/down from an extreme
- enter when it crosses a moving average
- enter when there is a pullback on its way up/down
- etc.

From eyeballing your oscillator I don't see anything predicting turning points but it'd be good to have some proof. Thanks for sharing.

[Elliott Wave]

Here is a different version.

-I reversed the plot logic so above center is a rising trend, and below is a falling trend.

-There is a second smoothed plot which can be used with the other line in a crossover system, or as the primary signal to reduce some whipsaws.

-The color scheme has been changed to be a little less ugly.

Hopefully its useful to somebody.


*The version I'm using uses the JMA, but since that's a proprietary indicator, I used the T3 for this version.

[Elliott Wave]

Here is a quick early version of the COT Proxy. I have no idea how it compares to Larry Williams' or Jrbolsa's version, but its a start.

Maybe people can play around with it and see if its useful at all. I have no idea what the best period setting is. 3 was arbitrarily chosen.

[Elliott Wave]

cot proxy emulates cot report for those that don't have cot report (stocks, etc.).
But in the commodities, the cot proxy emulates the behaviour of the cot report very well. I think that i prefer the cot proxy because in a daily basis you get a better information. In the cot report you can't have a "cot report" for one day that differs from another day. In the same week, you have the same data everyday.

regards,
jr.

Here is a formula I found for 'COT proxy'. It looks quite basic and doesn't require any advanced coding like a proper COT indicator does.

(MovingAvg (Open - Close, bars used in average) / MovingAvg (Range, bars used average) *50) + 50

[jrbolsa]

cot report and cot proxy

cot proxy emulates cot report for those that don't have cot report (stocks, etc.).
But in the commodities, the cot proxy emulates the behaviour of the cot report very well. I think that i prefer the cot proxy because in a daily basis you get a better information. In the cot report you can't have a "cot report" for one day that differs from another day. In the same week, you have the same data everyday.

regards,
jr.

[rretch]

I think the COT Proxy is different. From what I've read the indicator 'tries to emulate the COT so that you can use it on securities that dont have COT'. I could be wrong though.

I guess I took 'proxy' as a way to get the data into applications -'a go between language for data read/write'....I've never looked at anything other than CFTC data....but, taking a quick look at what 'COT Proxy' actually is, makes me wonder as well.

I've never really needed a COT Proxy indicator since I trade fx and currency futures (CFTC being sufficient). With equities though, it seems like their respected ETF's and using their non/correlated sector breadth indicators could be compared and used as a type of COT proxy....just a guess.

[Elliott Wave]

I think the COT Proxy is different. From what I've read the indicator 'tries to emulate the COT so that you can use it on securities that dont have COT'. I could be wrong though.

[rretch]

How does Larry Williams make his COT proxy?

I am a bit skeptical of mechanical ways to interpret the COT data. Things change and also change in positions depends on the context of the market. I find the COT quite complicated, good for a general direction but you have to use other TA for confirmation, timing, etc..

I do respect your skepticism, but I'm not sure an application would have to 'interpret' the data. The data is what it is....data. This data is not scientific or historic data, for example, which can be interpreted differently by different background academics. I use an excel sheet to make my own COT indicator, and there is no interpretation while I plug in the numbers....there is nothing to interpret.

Yes, things in the market do change and the COT data reflects this. Though it's possible, I don't think the govt. is skewing the data for any particular reason....though, that 'may' be the only source of potential interpretation mistakes. But, counting contracts (and their open interests) is not of particular difficulty.

I guess the only other area of concern would be in the code, though that can be easily checked with CFTC releases.

-robert

[jrbolsa]

Take a look at this:

http://www.jrbolsa.com/081004.htm

In the graph, "JRCOTIXC_LC" is the indicator that paints the net commercials positions in Live Cattle.
"JRP_JRCOTIXC_LC" is the cot report between 0 and 100 (percentualized)
"JRCOTPROX_LC" is the cot proxy
In my last "silver" comment you can see how we did 42% in "Adolfo Domínguez (ADZ)": http://www.jrbolsa.com/090606adz.png using the cot proxy

If you need, I can show you more examples showing the little difference between the cot report and the cot proxy

regards,
jr

https://www.jrbolsa.com