Here is a swat at trading a pair in different industries. Note the very important Point #6 below explaining the seriousness of this trade.
AET Short ...... Insurance company with 3 million shares / day traded.
R Long ....... Rental trucking company with 1/2 million shares / day traded.
Two posts (Chart 1 and Chart 2, respectively)
http://www.elitetrader.com/vb/showthread.php?s=&postid=3784777#post3784777
The approach is to accumulate very small lots of shares over time (multiple layers) as position moves unfavorably, and be willing to hold for very many months as needed.
Trade (including mulitiple layers) can close much earlier should a significantly favorable move in position occur.
Points:
1. This pair was selectively chosen for correlation. Caveats: It currently is not at an optimal spread between the Short and Long legs relative to the multiple months (even years) of history I typically demand for pairs. Very long history which I put less weight on for significance, nevertheless includes times of even significantly wider spreads which points to a small possibility of an aggregate, negative return.
2. Whether very long term hold or relatively short term hold, the expectation is for a gain meaningfully net of commissions and shrinkage.
3. An accumulation over time of shares in a long term hold scenario has expectation that in absolute terms, a sizable gain will be had but potentially with a humbled, annualized return (accounting for time in a trade) that an otherwise average pair trader typically enjoys.
4. "Trading of layers" (opening and closing them) over time would be a possibility, but constraints on this trader's time currently will probably not allow this. Historically, though, trading of layers has been overall gratifying.
5. By "being in the trade" this is an acid test to take the experiment seriously and to maintain the discipline over time it legitimately deserves.
6. The measure of success in this trade can point to the viability of--or lack thereof--trading pairs with stocks of disparate sectors and/or industries comprising the Long and Short components of a pair. Only a series of such trades can allow for some statistically meaningful conclusion. The value of such enterprise is to determine if the population of tradable pairs can and should be expanded when carefully selected.... at least to the extent of this trader's style of pair trading, capability for selection, and extent of trading talent.
Two posts (Chart 1 and Chart 2, respectively)
http://www.elitetrader.com/vb/showthread.php?s=&postid=3784777#post3784777