Let's suppose you could use the fundamental inputs to develop a price model for NG. Your model is developed over historical data, where all the fundamental values are know. Sounds like something that can be done with sufficient data and good regression package or programming skills. You would still need accurate forecasts of these fundamental variables to calculate a future price. Where will you get accurate forecasts for weather, storage, and distribution disruptions? That's where these fundamental models become challenging.
This is not the way to look at it. For one, fundamentals do not change often. They are fairly easy to forecast and there is not a lot of discrepancy even for publicly available forecasts. The hard part is dealing with the dislocations. Let me provide an example in the stock world since it might be easier to follow. Say I built a cash flow model for NVDA. I have the projected cash flows for the next 3 years out, I have assigned the proper growth rate to those cash flows, weighted them by a certain probability factor and properly discounted them for given risks. Whatever this number is, its not going to change much. Sure, every qtr I'll tweak the numbers based on forward guidance but for the most part, it's static. Yet the stock is all over the map. Is my cash flow model going to predict the pullback it had from 120 to 95? Probably not, it only lasted a week or two. Would my model have predicted the velocity of the move from 95 to 160? Almost certainly not. So there is where the real rub is. The fundamentals for oil or nat gas are not changing much and they are pretty static. The much harder part is given a certain set of fundamentals, how do we go about modeling the variance around these fundamentals? This is where people get paid. Nothing is easy in life and I certainly won't pretend it is, but people are doing it. I think my models are fairly robust and they properly account for both fundamentals and sentiment. Of course I trade spreads not the outright so the spreads are much cleaner in terms of capturing the fundamentals.
As a spread trader, would you share your thoughts on modeling the spread itself versus modeling the individual legs of the spread separately?
I'm not sure what you mean by modelling the individual legs. The spread in both oil and natural gas is not a combination of two months but an actual economic value, it's the cost of storage. It's something real. So when you model a spread you are modeling the economic benefit of storing oil or gas vs selling it into the spot market. Or put another way, should I sell it today or at some point in the future. There are two storage functions. One is the real cost of storage. If I call up Henry Hub for natty or Cushing and ask for a quote on 6 months of storage that is a real price. The spread in the futures market is a synthetic price. It is some function of the real price. The two don't have to equal each other and rarely do, but the value of the spread does govern the economic decision every asset holder must make. The individual legs are just that, legs. I'm not sure there is any value there. In the commodity world, most people focus on relative value. It's easier to understand, interpret and model. In the advent of big data, so many people are modeling this stuff that there is much less meat on the bone then there was 20 years ago. There is so much data out there and it's so easy to get. This is true of all markets, not just energy. One of the nice things about energy vs say bonds or stocks is that mother nature is an independent contractor. She doesn't work for anyone and doesn't take crap from nobody. If she says there is going to a storm, then by the powers that be, a storm it is!
@Maverick74 The only edge in oil & gas that is scalable is trading the physical with your own inventory around the world and profiting off the spread. Deliverables are a good thing if you're the one buying it for a fraction of the home market price. Your variance on those products by out right trading will kill 99% of traders over the long run(months, not years.) Ring ring.... you've got a margin call to make. I've tried speaking with you before about a DeltaOne strategy, and I know how this game works. Quid pro quo. If you've got recommendations or if you've got the time to help, just shoot me a message. Thanks
That is not true. Optimizing physical assets is one way for sure. But energy markets are still highly inefficient. Edges can be found in any market through data and statistical analysis. As I said before, I do think this is getting harder as data is becoming easier and easier to get. I apologize that I can't seem to recall speaking to you about this before or even what deltaone even means. You can PM me if you want. I work with two partners and keep proprietary stuff close to the vest for obvious reasons. But if you are looking for general feedback I'll be happy to oblige. Fyi...Mark's son Daniel has a natty fund, just trading paper, he is doing well. I also know a lot of guys in the electricity space trading FTRs, all virtuals, no physical assets doing well. I live in Houston and there are a quite a few energy funds down here with no assets that are doing just fine. Again, it's a lot of hard work and I certainly would agree with you that most retail guys on ET would get killed trading energy on their own.
I just scanned my inbox, unless you used a different alias we have never chatted before privately so perhaps you spoke to someone else. Anyway, feel free to PM me.
Just sent a PM. Referring to the physical assets, I believe many companies already profit off the spread by utilizing weak currencies/US gov protection/and economies of scale. It's not a cheap mans game certainly. Currently, I believe the profit potential is in Venezuela for physicals but I do know about John Arnold and Enron. I've been raised around it. But, finding an edge that disappears after the JP Morgan's catch on isn't something to be remembered.
Brent is by far the hardest. I think Natural Gas is the easiest, followed by heating oil, gasoline and then WTI. None of these are easy, just on a relative basis.