Poker is a gambling game, and as such those who play it are subject to the whims of lady luck. This is not to suggest that poker is a game based purely on luck. On the contrary. Playing poker well (especially limit hold’em) requires a significant amount of skill. It also takes a significant amount of skill to decipher who is a good player and who isn’t, because our typical methods of assessing quality of performance do not hold true in poker.
In just about every endeavor that we embark upon, the litmus test for the quality of that endeavor is the results. How fast did you run the marathon? What did you get on the test? What did you shoot for 18 holes? All of these tasks produce tangible results that provide a reasonable metric for judging the quality of the effort. You shot 74? You must be a very good golfer. You shot 128? You must be a very bad golfer. Both assumptions made here about the quality of the players are reasonable and accurate; if you shoot 74 you have to be good and cannot be bad, and if you shoot 128 you have to be bad and cannot be good.
Poker results also can provide a good indication of the quality of the effort, but the scale that is required to obtain meaningful results is one we are not used to dealing with. A round of golf takes a few hours. A marathon is also measured in hours, as are most tests. But poker takes thousands and thousands and thousands of hours to produce meaningful results.
The poker tools website ev++ has developed a poker variance simulator that can be used to demonstrate just how much luck influences the results of poker.
Assume we have 1,000 poker players who play $10/$20 6-handed limit hold’em online. They play 4 tables at a time and play 600,000 hands in a year. (This equates to about 1,500 hours of “work”, and is a schedule that an actual professional player can easily adhere to.) All the players are of the exact same ability and are expected to win 0.5BB/100 (the same as 1bb/100), which equates to $60,000/year. (Basically they are all clones of the same player.) Plugging those numbers into the simulator along with a reasonable standard deviation, we get the results below in the form of a graph. Multiply the vertical axis by 10 and you get dollars. Multiply the horizontal axis by 100 and you get number of hands.

To make the graph easier to read, it only shows how the “hottest” (green) player did over the year and how the “coldest” (blue) player did. The results of the other 998 players fall somewhere between the green and blue lines. The dashed line indicates how the players would have fared if there was no luck at all involved.
All of the players are projected to make $60,000 based on their skill level, but the player who lady luck smiled upon made in the neighborhood of $145,000 and the poor fellow that she decided to smite ended up LOSING about $38,000. Both are “winning” players who make decisions that will ultimately show a profit, but one got lucky and the other one didn’t. This was over the course of an entire year, and still there was over a $180,000 spread between the biggest winner and the biggest loser. And this is a sample of only 1,000 players. The poker boom of 2004, where every frat boy and ESPN watcher gave online poker a whirl is dead and gone, yet at this moment there are still over 160,000 people playing at PokerStars.
It’s easy to see where a layperson could go with this. “You played all year and LOST almost $40,000? You need to quit. You must be terrible. That guy over there made over $140,000. Maybe you should pay him to teach you how to play.” (And thus was born one of the biggest rackets around; the poker coaching industry.)
It gets even more dicey if we replace the 1,000 players who are projected to win $60,000/year with 1,000 who are all projected to lose $30,000/year.

In this case the player that lady luck smiled upon won about $57,000 over the year. His results look like what the results of the first 1,000 players were supposed to look like, but he is not nearly the player that they are. But his results say that he is. (This is actually very unlucky, because often this player will think he is a good player and start playing higher stakes and even quit his job to become a professional poker player. Needless to say, things do not end well for him.)
If the results of an entire year of play cannot paint a clear picture of one’s ability, how do you know who’s any good and who isn’t? Ultimately it comes down to the decisions that players make while playing hands, and the rationale that they use for those decisions. There are plenty of players out there with fine results, but when you ask them to explain the reasons for the decisions they make they either cannot do it or the rationale they give is easily shown to be faulty. Many of them know that they cannot give a reasonable explanation for what they did and laugh it off by saying something like, “I dunno I was just smashing buttons”. After all, they’ve got great results over 600,000 hands so they must be doing something right. Right?