I was under the impression that these commercials hedge their physicals with futures. They even get more generous leverage than traders. Geez, if they're not hedging their physical operations, then they are as much speculators as any trader.
Well, Glencore is a speculative commercial like Louis Dreyfus, et al. They are trading around a physical inventory book. They carry a great deal of physical but they are not producers or users per se. And IMHO that is the conundrum - it makes dynamic hedging very difficult. A commercial that does not run a spec book would have a much easier time with hedging.
Speculating on the physical ... that requires a very high level of confidence in your price forecast. These commercial specs must be experts in their field and have balls of steel.
I received an email from an ET Member, who is not a client, and he wanted to relate to me his principal observation at a recent job interview with a Chicago proprietary futures trading firm. In his words, and I quote: "While I was there I had an opportunity to interact with a number of the traders, lots of them were trading spreads (Yield Curve, Crack spread, TED Spread). At that point, my knowledge of spreads was pretty limited (and still needs lots of work to be honest), but one thing was clear: Professional Chicago Prop traders were trading spreads, and they made it clear that it is a much better way to trade." That was his observation, YMMV. That was his strong impression. Spread trading is not the be all and end all, nor is it an ATM machine. Point being, that he as a scalper, was genuinely shocked by the methodologies employed by a large number ( but not all ) of those traders.
During client meetings this week, we came across a few examples where the proprietary study takes the client out of the trade before our stop-loss is hit ( we set the stop-loss and profit targets at trade entry time ).
And as far as charting goes, eSignal has 7,009 exchange supported spreads listed in their library to date. Please chart the exchange spread in lieu of a synthetic combination whenever possible - much better data.
Two thoughts: 1. If your charting package supports exchange supported spreads, all of your fave technical studies will work just fine with them. 2. Hint - the prompt months are largely spec volume and that driver makes them delta directional with the front month; much more even with a calendar ( pair) spread. You really can't model roll behavior to the degree of likelihood you get farther out in the curve. Long story short: play where the volume is largely commercial / institutional.