Here's a little piece of wisdom that I figured out on my own some time ago. I call it the Two Episode Rule and goes something like this: To decide whether you want to watch a series, watch the first two episodes. If you liked them, chances are, you'll like the rest, too. A corollary to this rule states that multi-parters only count as one episode.
I actually discovered this rule on accident, thanks to a sitcom that most of you have probably never heard of, Comeback. A few years ago, when this series first aired, my family and I watched about the first ten seconds of the first episode and then switched to another channel in disgust. But then we decided to give it another chance, and actually watched the full second episode. I have remained a fan of this series ever since. Now, our skipping the first part was actually a very fortunate thing, as later, when I watched it, I found out that it was pretty lackluster, compared to the rest of the series. I think that had we not skipped it that first time, we would've not tuned in the next week and thus missed out a lot of the later Comeback goodness.
Over the following years, I have put a lot of thought into this phenomenon and eventually arrived at the Rule in its current form. Here's the reasoning behind it: The pilot episode will invariably be different in tone from the rest of the series. Not only is its goal to introduce the viewer to the universe, but also to provide an exciting start to the overall story to rope the viewer in. The pilot is also often a two-parter, which is the reason behind including the corollary. Keeping that in mind, it is easy to see that the second episode is actually the first "normal" episode the viewer will see, which makes it a better indicator of the things yet to come.
One weird thing that I've noticed since I formulated the Rule is that the second episodes are often among the best ones - take for example Red Dwarf's "Future Echoes". It's as if TV makers are aware of this phenomenon and live by the rule "put all the exciting stuff in the pilot and all the good stuff in the second episode". I'm not sure whether that would prove the Rule or invalidate it, but it'd certainly lend credence to the logic behind it.
Now I know what some of you are thinking: "But T-Jack, wouldn't it be easier to just ask a fan of the show to recommend an episode and base your decision on that?" And I know that because that's what a friend of mine said when I explained the Rule to him. Well, you could technically do that, but that method has a few drawbacks. First, a fan of the show is almost certain to point you towards his favorite episode, i.e. the best one, which is not exactly representative of the whole series. Second, you'd have to be wary of spoilers. And third, this method is not easily applicable if you can't contact a fan, maybe because the show is really obscure or just starting out. For that matter, this is what the Rule tries to emulate - a viewer's reaction to a new show. You are being introduced to the series exactly the way the creator intended.
I should note one more thing, though: Don't forget that this is only a rule of thumb. In every show, episode quality can fluctuate wildly even within one season. It is usually safe to assume that after five good episodes the quality won't go down the toilet, but it can happen. The Rule can also give false negatives - some shows are known for getting better along the way, Growing the Beard as it is called in tropese. So I guess don't rely only on the Rule to tell you if a show is good. Remember, it is only supposed to be the first indicator of many.