When you see a battery benchmark on sales collateral or at point of purchase you generally see something called "MobileMark." Everyone in the industry knows this benchmark is wildly optimistic and that the actual battery life you'll get is often less than half what MobileMark suggests. This is because MobileMark measures battery life much like you might measure gas mileage if you started the car, put it in neutral, and coasted down a long hill.

Generally if you want to know how much battery life you will actually get out of a new notebook it is best to take the MobileMark number and cut it in half. In a car this would be like, on a Chevy instead of seeing 16 city and 22 highway you only saw 4.6 miles per gallon -- and you discover this when your car runs out of gas halfway to your destination in a little town we call Love Bubba Texas (where the gas in $140 a gallon, the temperature is 120 degrees, and water is $50 a glass). I'm pretty sure that would be my luck.

So the reality is you buy a notebook, it says 4 hours of battery life. But when you get on a plane figuring you have time for a movie and some work, 85% through the movie you get a battery warning and 95% the battery goes dead. You not only don't get any work done you have to wait until you get to your fricken hotel room to see the end of the damn movie. Not that this ever happens to me, mind you.

You'd figure at least this statistic would make sense because GHz and number of Cores doesn’t actually tell you what you can do either, because different processors, configurations, and components have a bigger impact on overall system speed than any one part does. Even with one part, the Atom, for instance, does less work at a given GHz than a Core 2 does.

So there is really only one metric that looks like you can rely on. And it is wildly inaccurate.

Get the full story here at Datamation.com.