Predicting Crude using Intermarket Analysis

Quote from Tyren:

It's more than XLE/$XOI to look at.
In my opinion one should look at the other commodities and compare them to crude oil for divergence. HG, GC, SI, HG, HU, HO.
Now has been the the driving-season over there, and HU is important.
And then you have the futures, the prices in 2009, 2010, 2011 for crude oil.

Good luck.

Tyren
Norway(we found oil in late 1960's, never looked back since... but production are going down now and we are empty in 20 years time or so)

I decided to discuss the results of using silver and gold to predict crude. Because the data goes back further we were able to go back to 1984. We found that neither of these are predictive. Silver produced only 27K as the best set of parameters when optimized from 5-40 for both sets of parameters. Also for both metals, only 8 combinations were profitable. The best set of parameters , which was silver which was 5,5 made 27K since 1984 and lost -$350.00 since 2000.
 
Quote from Murray Ruggiero:

I have decided to optimize our original data series against not only Crude but also , Heating oil and Unleaded Gas. We found that Heating oil gave us the best bang for the buck. Using the same parameters with EXV as with Crude 40,25. It produced $103,761 and had a drawdown of $11,369.00.

In addition the equity curves correlation of these very similar markets is not as high as you would think. If we look at Crude and Heating oil it is high but not extremely high. It is .59 on a monthly basis. Heating oil and Unleaded Gas have a higher correlation at .77 on a monthly basis.

I use Crude oil to predict Heating oil as someone requested. I ran it from 1984 and the best sets of parameters made between 65K and 75K ,with about 35K since 2000. The drawdowns were about double what was produced using the oil sector groups. So you can see that not even Crude can forecast heating oil as well as these stock baskets.
 
In case you want to see the optimization report here it is for Heating oil using crude.
 

Attachments

Quote from Murray Ruggiero:

I decided to discuss the results of using silver and gold to predict crude. Because the data goes back further we were able to go back to 1984. We found that neither of these are predictive. Silver produced only 27K as the best set of parameters when optimized from 5-40 for both sets of parameters. Also for both metals, only 8 combinations were profitable. The best set of parameters , which was silver which was 5,5 made 27K since 1984 and lost -$350.00 since 2000.

Silver, copper and partly gold trend much more than crude oil and S&P500. So one can probably not use it on 1-2 month swings. I just look at them at 2-8 days pullbacks, divergence over that periode when I have a pattern(pullback).

So if it is 1-2 months trade you have in your test you can test crude oil vs. S&P500.
S&P500 often leads with 1-2 months.
 
Quote from Tyren:


So if it is 1-2 months trade you have in your test you can test crude oil vs. S&P500.
S&P500 often leads with 1-2 months.

I ran the same type of intermarket model as I used for my oil stocks using the SP500 Futures. I tried both a positive and negative correlation and found the best set of parameters used a negative correlation with 10 day average for the Crude oil moving average and a five day moving average for the SP500 made 73K since 1984 but with a 37K drawdown. This is not as impressive as the other relationships I have shown. Using longer moving average periods the results are even worse.

The empirical evidence for the SP500 is very strong over the past three years but before this three year period this relationship lost money.
During the period 1/1/2004 to 5/22/2006 this SP500 to predict crude using these parameters made about 83K but in the 20 years before it lost about 10K. This relationship also did well in 1991, the year of the first gulf war. It seems like this relationship is based on middle east war effects between the stock market as a measure of how the war is going and crude prices which go up if the war is going bad and down if it is going good. If we remove the war effect this relationship lost 24K in 20 years, this is because the oil stock in the SP500 have a positive correlation and there effect can be seen during non war times.
 
Volume today is perhaps 10 times that in 1998 for crude oil. It perhaps therefore makes smaller waves with ca. 1.5 months lead.

1990-99 it was tech, tech, tech in Usa. S&P500 did not follow macro-phases very well. I made charts using Nikkei and showing the 5 months lead to crude oil. And the TYX/USD 5 months or so lead.
Charts :
http://www.aksjeinfo.com/ubb/Forum1/HTML/000157.html
 
Mr. Ruggiero,
can you please reply to this old illiquid question in this thread?
http://www.elitetrader.com/vb/showthread.php?s=&threadid=72664&perpage=6&pagenumber=6

>>I guess my main question is: don't you need a good reason/angle in comparing 2 markets for predictive purposes, other than just slapping 2 charts over each other and saying sometimes they move together, sometimes not? Why not concentrate on markets where there is a direct logical relationship -- say, taiwanese dollar vs korean won, crude vs gasoline, Cheescake Factory vs Wendy's etc? Is it your point that there is no more edge to be found in these "obviously" correlated markets, and that now we must stretch our horizons in finding more esoteric relationships?>>

Thank you,
Bernard
 
Quote from illiquid:

Well that was my point -- comparing dollars to crude is much like comparing 2 stocks in completely different sectors. In any case, when it comes to currency differences, all markets are affected by simple exchange rate changes, eg, IBM staying at $70 will fluctuate in terms of euros. But that doesn't mean there will be a predictive edge in comparing the movement of the euro vs IBM, only that you will know the price of IBM in euros will go down if the euro goes higher and IBM stays the same in terms of dollars. There is as much edge here as knowing that we're in a bull market and that most stocks will generally move higher under such conditions.

I guess my main question is: don't you need a good reason/angle in comparing 2 markets for predictive purposes, other than just slapping 2 charts over each other and saying sometimes they move together, sometimes not? Why not concentrate on markets where there is a direct logical relationship -- say, taiwanese dollar vs korean won, crude vs gasoline, Cheescake Factory vs Wendy's etc? Is it your point that there is no more edge to be found in these "obviously" correlated markets, and that now we must stretch our horizons in finding more esoteric relationships?

You are missing the point completely. I am not picking esoteric relationships. I am picking relationships which are logical and strong. Companies which drill for oil will outperform when analyst believe oil prices will increase. The related stocks groups leading the related commodity is something that John Murphy pointed out almost 15 years ago and it has worked well all this time.

This is why energy based stock groups to predict , crude, Heating oil and Natural Gas makes sense. It not esoteric , but requires understanding the stock groups you are looking at.

Another classic relationship is the Fidelity Select Chemicals Fund to TBonds. This is because US chemical companies are treated like commodities and these stocks rallying is a precursor to inflation.
 
Quote from Murray Ruggiero:

I am really surprise we don't have more interest in predicting Crude...

Hi Murray,

Appreciate the work your doing.

Keep at it because I personally know a dozen or so ET members that have a strong interest and profitable via intermarket analysis of Crude including myself.

However, although I don't currently trade crude (I do a few times per year), I too have been doing such via intermarket analysis since the 80's.

Yet, your approach is more mechanical while my approach is rule based while not mechanical via a different angle.

Keep at it because there's many different approaches to intermarket analysis and its good to see a different angle being discussed.

By the way, I use intermarket analysis of Oil along with the intermarket analysis of other key markets to help with my profitable day trading of the Russell 2000 Emini ER2.

I'm interested in your angle because I'm constantly asked about mechanical sources of such and can now point traders I meet towards your direction.

Intermarket analysis is one of those things that's a tough pill to swallow for most retail traders here at ET due to the fact that price action only trading is not as popular as indicator based trading.

Once again, we are here and listening although not posting often.

Mark
 
Murray, I just wanted to mention that I still follow your FSENX/CL intermarket system, my version of it anyway. FSENX peaked on 6/27/08, a couple weeks before CL. The system gave a sell signal on 7/25/08. CL subsequently collapsed as we all know, but the system got me out for a profit. I didn't want to get out; I thought oil would go higher. I guess that's why they say you should always follow your system.
 
Back
Top