Three years ago, The Wall Street Journal ran an article entitled "Apache Sees Profit in Discards." There, Apache's (NYSE: APA) habit of acquiring and breathing new life into tired assets like BP's (NYSE: BP) Forties field was contrasted to that of peers like EOG Resources (NYSE: EOG) and Anadarko Petroleum (NYSE: APC), which tend to seek an edge by carving out leading positions in hot new plays like the Bakken and the deepwater Gulf of Mexico.

Not a lot has changed in the ensuing years, save for decisive shifts by Devon Energy (NYSE: DVN), Encana (NYSE: ECA), and to a lesser extent Talisman Energy in favor of unconventional onshore North American resource plays like the Haynesville and the Marcellus. Apache is certainly involved in unconventional activities, such as in the Horn River Basin and the Granite Wash, but the company hasn't gone all-in like several of its peers. Apache is no trend-chaser, as was reaffirmed by a classic contrarian move announced Monday.

We've been watching Devon shed its offshore and international assets for a while now. BP took $7 billion worth of assets -- including all of Devon's remaining deepwater Gulf of Mexico interests -- off the company's hands just last month. Left behind was a package of shallow- water U.S. properties that couldn't be much more unfashionable. Apache is happily snapping that up for the sum of $1.05 billion.

So, is this a good deal for the detritus developer? Apache is paying a little more than $25 per barrel of proved reserves and around $55,000 per flowing barrel. The latter figure is quite cheap for an asset that's roughly half oil and liquids by both reserves and production volume, as seen by comparison with other recent deals. Apache's acquisition immediately adds to cash flow, whereas SandRidge Energy's (NYSE: SD) latest Permian purchase isn't modeled to add to cash flow until 2011. The company also sees a lot of upside potential across the properties, including 79 recompletion candidates and 26 exploratory prospects identified.

It's great to see Apache zig whiles the rest of its peers zag. This has been a winning formula in the past, and I see no reason for it to stop working now.