lol at your attitude in every post, 9000 posts of an armchair trading warrior... the chart is to indicate the relationship, you hero
I didn't realize I upset your delicate sensibilities. Mea culpa for asking a question. What you sittin' in? Who would've thought copper futures would be correlated to a copper producer. Ok, so it's to indicate the... relationship. Pure champion stuff. And for the record.... I may be an asshole, but unlike some I can build a spread chart.
No haircut would be given to trade Comex against RTP. And it's a loss to attempt the spread on such a frequency. Equity swaps on low-frequency.
Bone - An insightful number of posts. When it comes to pulling the trigger (entry or exit), what are the inputs to your proprietary method - without giving too much away? I assume it's not a TA pattern/set up. My guess would be the speed of divergence between the legs on some statistical basis. However, if I'm not mistaken, that would show on the chart perhaps as a 'sudden move', which in turn would take us back to trading the equivalent of a TA pattern. I'm familiar with seasonal spread trading where the current market is compared to previous years performance. Are similar measures possible for intraday trades?
I have custom technical studies I designed expressly for spread trading. My clients use my studies along with eSignal, CQG, or Bloomberg charting packages - depending upon what they are comfortable with. We use statistical methods like correlation and cointegration studies to design and build synthetic spread combinations. Our hedge ratios use volatility and currency adjustments to take as much delta directionality as practical out of the trade. The volatility and daily trading range of a particular spread dictates it's holding timeframes. Big difference between a Eurodollar Condor and an RBOB Crack Spread. Some of my clients use the intermarket correlations to arbitrage between products utilizing 'lead-lag' strategies and skip the spread stuff altogether. They are spread traders who never get around to closing the legs, essentially. That is what I was trying to illustrate with the ES Correlators chart, and the Copper vs. Rio Tinto chart. There is alot of multi-dimensionality and flexibility to "spread" trading. Arbitrage, lead-lag, fully hedged multiple leg positions designed for fairly long holding periods, market making, etc. etc.
I do not put a great deal of stock into seasonal strategies as pushed by Moore Research Center and Joe Ross and Jake Bernstein. I think that when you test them for yourself, or trade them live, that the hype is just hype. I thought that the Wall Street Journal article on Jake Bernstein's seasonal strategies was very poignant. Absolutely. Look at the very definition of Statistical Arbitrage being a very prominent example. Having said that, I think that the cheap capitalization requirements and the "better behavior" and general trendiness of most spreads as a general rule lends themselves to longer holding periods.
So these studies give the enter/exit trigger? Surely, just because 2 legs are highly correlated or show a degree of cointegration (I had to wiki that one ) doesn't in itself give an entry condition. Or, thinking about it,.... Given cointegration is stochastic, could it be that once correlation is found between legs, the cointegration is used as the timing tool? My awareness of cointegration is minimal I admit.
Can you elaborate a bit on what a cointegration study is - are you using software to estimate your cointegrating vectors?
Hi Everyone Here is a chart of today's price action on the Crude Calender spread Aug/Dec. I entered this as a custom spread in my esignal. Now check out the price action. I note the bid/offer spread on the time and sales is mainly 1c or 2c. Bone - would it be possible to trade this price action manually legging in and out. Would slippage make this viable? London