Stuck closed yes, open, not so often.

You would think they would chose a design that was not subject to a high pressure closing problem.

If they need to purchase different, more appropriate, valves then there could be a lead-time while those valves are made. Although I have no insight in Boeing’s thinking, they may want to test the multi-engine system separately before performing another full-assembly pad abort test.

In my experience, the darnedest things fail. Technologies that are well established and assemblies that are checked out fail at the most inconvenient times. Sometimes it is because they were used in ways not intended, as Kirk mentioned, as happened with valves on the Mars Observer satellite, and as with a Russian Fregat upper stage.
http://behindtheblack.com/behind-the-black/points-of-information/independent-arianespace-investigation-cites-design-error-as-cause-of-russian-launch-failure/

Sometimes it is because assemblies were not built quite right, which is uncommon per assembly, but the failure of any one of thousands of assemblies can be disastrous. We have seen this recently when Russian rockets were built with parts made from materials that were not as strong as the drawing called for.

A couple of decades ago, I bought a brand new car, but the tubing for the cockpit heating was not properly connected, and my first drive in it turned into a colder-than-expected experience. One difference between a problem on Earth and one in space is that in space the crew might freeze to death. Another difference is that we can still fix the one on Earth. We necessarily require and demand far higher quality in our spacecraft than in our earthbound craft.

It seems that in this case Boeing tested their system with a single engine and that this was their first test of the four-engine system, expecting the one-engine success to carry over. Quite often, individual parts are de-rated (e.g. a 1/2 amp resistor is used where up to 1/4 amp could be expected) in order to assure the part is not over stressed. It seems that the Boeing designers did not account for a need for additional valve performance when used in the multiple engine system.

Generally, we don’t hear much about the troubles that most development programs run into, because these (e.g. cars, computer chips, etc.) are done by private companies that try to keep their plans and progress away from the competitors’ eyes. Blue Origin once worked in secret, but their test flights had to have some amount of public exposure, such as when they asked for FAA approval to fly them. Eventually they embraced the publicity, webcasting live test launches.

Kirk’s Ars Technica link has a good quote or tagline: “These development programs are hard. Especially for human spacecraft.“

A commenter on yesterday’s Ars Technica article said Boeing claims that “valve failure was caused by the four-engine configuration of the service module and would not have manifested in single-engine tests” and went on to speculate that “My educated guess is that something about the way the four engines are connected to fuel tanks exposed the fuel valves on all four engines to flow/pressure conditions which were outside of the parameters under which the engine was previously tested, so they all failed in the same way”. (The commenter sounded informed, but did not cite a reference for their Boeing claim.)