correlation: for example, ETFs on silver & gold, or ETFs on the Dow 30
& S&P 500.
My idea was to trade ETFs, so the "roll" of futures (around
expiration) doesn't affect pricing so much.
So, say metals ETFs: silver ($slv, iShares ETF) & gold ($gld, Spider ETF)
Index ETFs: Dow ($dia, Spider) & S&P 500 ($spy, Spider)
There are, of course, many more examples.
So, the idea is to get the average spread between the prices of the 2
chosen securities on some specified time period: maybe every minute
for a week. Historical data would suffice for any of that statistical
stuff.
Then, once an average spread is determined, you'd use this equation to
make a decision to enter a trade (shorting the higher priced security,
and going long the lower priced security):
*SD = standard deviation
( Current Spread - Average Spread ) / 1SD > 2
So, if the spread is wider than 2 standard deviations, you wager the
spread will tighten. Conversely, if the spread gets < -2, you switch
the long & short: wagering the spread will widen.
==============================
That's the basic idea. One also has to then program in that if one
trade is already on, not to open another trade that isn't a closing
trade, or whatever other rules one wants. For example, if volatility
goes crazy, don't enter any trades, or if the market drops like 3%,
don't enter any trades, etc, etc - to avoid unpredictable market moves
to reduce potential losses.