EBIT/Tev (one of the best if not the best predictors of returns in general) had it in the cheapest decile of >10B US equities when Buffett was buying. Add to that the fact that they are the low cost producer, a history of strong capital allocation, a history of strong returns on capital, and a rock solid balance sheet. Bonus points for oil as a bit of an inflation hedge if CPI picks up. That is a recipe for success.

Pretty consistent with Wesley Gray/Greenblatt research.

What is EBIT/Tev?

Earnings Before Interest and Tax/Total Enterprise Value. You could go further up the income statement to EBITDA yield, but I prefer EBIT when running loose screens (uses depreciation as a proxy for capital deterioration, which as we know isn't true). Papers that I have seen seem to show it is roughly a wash on performance between the 2 stats.

Enterprise level metrics show superior correlation to results as they normalize for leverage. P/E and for that matter book/market screens give an "unfair advantage" to levered up firms. Finally book screens don't seem to work well on mid-large cap stocks in practice.

Return on Capital seems to add some juice, as Greenblatt claims in the Little Book. The Magic formula ranks stocks on a 50/50 factor of ROC and EBIT/EV... but, you should know that Tobias Carlisle contends that the cheapest value decile stocks, ranked by ROC, should outperform the seemingly random 50/50 factor that Greenblatt has chosen.

At some point of course, over fitting may be an issue, but I suspect Tobias is closer to the truth, hence my comment on XOM.

Perhaps we should move this discussion to another thread. I do think it would be valuable to discuss apples/apples metrics though. I have noticed quite a bit of confusing ratios that cross compare firm, value, and equity level metrics. Each metric can of course be useful, but only as long as you are comparing apples to apples, if you follow what I mean.

EDIT: Oops...I realize you may just be looking for the XOM EBIT/ev yield?