Constant maturity Futures

Hi Samuel,

I made an attempt some time ago to put together a single price series for
futures contracts.  The problem of course is that to get a data set of
decent size to estimate a time series model I need to go over a few front
month/expiration cycles.  Here are some highlights of my experience in case
you find them to be useful.

I left the project (ie. betting using futures) for several reasons, but as
I recall it is not difficult to do for a single underlying.   My code
became clumsy however when I tried to automate the switch to a new front
month for all the underlyings that I was following since they had different
contract expiration cycles.  This was work in progress when I decided not
to pursue the project.

My experience with obtaining data is as follows.  I no longer had access to
Bloomberg, etc so I found that Quandl had free futures data but without any
adjustments.  They also have a database where they string together contract
prices using different rules.  Access to this database had a free trial
period but was otherwise $50/month.  During the trial period I sent them an
email asking about something that I noticed by eyeball on a price series
that did not seem right.  (I do not remember specifically what it was, but
it had to do with the data values themselves.  If someone is interested I
can dig up the email.)  They replied to thank me for pointing it out and
said something to the effect that they had now fixed it.  However I did not
feel comfortable using that data to bet money.

The other thing that I tried is through my online brokerage.  It turns out
that they keep data for a few months back.  It is not enough data history,
but by waiting until I downloaded data for a few initial months then I had
enough history from that point forward.  This data was free (since I had an
account) and I do trust the quality.  On the other hand, I had to string
the front month contract sequence myself and I had to do so in terms of
trading volume since they did not have open interest data available.

Best of luck,
Jorge

On Thu, Aug 13, 2015 at 3:54 PM, Samuel Wilson <samuelcoltwilson at gmail.com>
wrote: