Thanks for the advice. Don't worry, I'm not trading this anytime soon. Just something that I'm playing with and researching. Still have a lot to learn. Charles
Here's a chart for the current contract... I wish I could do it for the previous contract but this chart app is limited. It is certain that it will converge to zero by the Dec rollover. The above guys do the following intraday and is quantified at the following. http://www.indexarb.com/. I guess what I was thinking through is too long to be attractive. I'll run the numbers nonetheless. Still Very interesting. Definitely can get swept away with choosing optimimal (ie. max return) pairs.. Also what of spread divergence, when on neutral, reverse position??? Rhetorical Q in some sense... Just putting the notion out there... Thx Charles and Kindest Regards, MAK!
I think if you run the numbers through a black-scholes type of equation you'll find that the cost of carry, plus the execution fees etc make this very tough for the small guys on this time frame. Just a guess, however.... -The New Guy
you cant make money on future vs cash. the MM's make this arb with fast equipment (they always lease the newest for couple months the switch to the next newest thing) and stock loan and stock borrow.
Are calendar spreads on the same issue any better than the sympathetic issue spreads (e.g., ES/NQ)? I've never paid much attention to the non-front month contracts so I'm not sure how closely they track the front month.
I believe it's the same issue. You need BIG accounts and FAST programs to do this, imho. Not to say that there isn't money to be made somehow, somewhere with this type of thinking.... -The New Guy
The difference here is nearly a 9 month spread lock (ie. open and hold through until maturity)... Let the MM and arbs do their intraday thing... Of course money can be made here. If not, the spread would just sit flat through the entire horizon and not converge to zero. Whether or not the profit is worth the capital put up to execute the spread lock is the other side of the coin. If the concern is about fast execution, I would just limit the bid/ask long before the trigger values showed up on the DOM to ensure that arbers would push their orders after mine had executed. This though, is very tricky since converting the DOM into a spread DOM. Not impossible tho.. MAK!
this is not true, the spread converges to zero because of intrest that diminisches over time. be certain that whenever something is starting to get out of line, big money will push it immediately back either because they will arb it or either because the side that is overvalued will be sold or the one undervalued will be bought as part of the hedge of some other trade for example in options. Profit margins on these kind of trades will be minimal.
Somehow, I have managed to confuse and/or articulate exactly what it is that I'm pointing out. Apologies for not laying out the entire details of my train of thought... Kind Regards, MAK!
Charles, I did some testing on the ER2/YM pair (1:1 contract ratio) using intraday tick data (continuous contract) on a +/- 2 SDs from a 60 minute-period, 200 period SMA trigger over the period May 2004 through Sept 2005. It made several profitable trades, but the drawdowns were surprisingly ugly and eventually it took a trade that didn't return to the ratio SMA at entry by the end of the data series (open trade for 12 months). I had the system not take any trades in expiration months to avoid holding a trade at rollover, and to give my SMA indicator time to normalize after rollover. Do you think these two products are simply too dissimilar for this type of strategy? Were you using 1-min or EOD data (or other)? Thank you. Regards,