Thanks DJ. I'd like to add some additional insight on this particular type of failure if you don't mind?
The following pictures are of intake valve stems from two different failed engines with cosworth cylinder heads... and guess from which cylinder they're from? Yep, cylinder #2!
Uploaded with ImageShack.us
Notice, first of all, the valve stem is still straight after the failure- which means the failure wasn't a result of piston contact, or foreign debris in the cylinder. If that were the case the valve stems would have bent. Instead what we have here is a clean break, like the valve head said, "later man, I'm outta here", and took a vacation.
Secondly, look at the beach marks on the broken surface of the stem. See how the half-moon lines on the face begin on one side of the stem and get bigger? This is a sign of fatigue, and in this case, plastic deformation. It happens when a slightly bent valve tries to straighten itself out over many miles of operation. And at the weakest point of the stem (in this case where the stem is necked down, just before the radius into valve head begins) a microscopic crack appears then gets bigger and bigger over time. Eventually, POP goes the valve head.
Note this will always happen to the intake valves and not the exhaust due to the thinner (weaker) stem, and the larger head diameter- they bend easier.
Also, if an engine failure such as this happens at a low rpm and the driver shuts the engine off quickly enough (like the examples above), it is likely you will see only one, not both intake valves in cylinder #2 that will look like this because both valve heads will never decide to pop off at the same time, and it is always that first popped valve head that takes out the other intake valve (and everything else for that matter) before it has a chance to do it on it's own. However, a said failure at higher rpms will just obliterate everything, including the chance you'll find a relatively clean valve stem break to look at such as the ones you see above.
When it comes to installing a timing belt, I imagine the common logic of some engine builders/mechanics is to think the cam sprockets should only be rotated in the direction of engine operation, and that you always begin with the intake cam gear first... because after all, the mechanical cycle of engine operation is dictated by the intake valves opening before the exhaust valves do, right? So when it comes to lining up your marks to install the belt, it would make sense to phase the cam gears in the same manner, right?? WRONG WRONG WRONG!! Especially if you are running +1mm over-sized valves, be warned that you WILL bend the intake valves when they come in contact with the exhaust valves, in cylinder #2, on an EJ engine if you adapt this logic.
This is why it is very important to refer to the factory service manual when phasing the cam gears before you install the timing belt. If you don't, you run the risk catastrophic engine failure, and not to mention a very expensive repair.