I will use for the sake of example League of Legends, because it is the game I am most familiar with, but I believe related effects occur in other games as well that are prominent (Starcraft II, Street Fighter IV, Marvel vs. Capcom 3, etc.).

Anyone with the internet connection required to play League of Legends can also open up Twitch.tv or Own3d.tv and find streams of high level players playing the same game. The fact of a highly visible metagame does very interesting things to the community and the metagame. In my opinion, most of the results are negative for both players at most levels except the highest (WCG entrants, 1700+ ELO, etc.)

Those running the streams make a living from the advertisements, which pay per viewer, helping support their plane tickets to Korea/Europe, paying for their food, etc. Some of the most popular streamers I believe are even self-sufficient or making significant amounts of money from the practice. They are known to do 24-hour streams occasionally, I believe to help make even more money. At 3 cents an ad view, 6000 viewers (hit by the most popular streams pretty regularly) means 180 dollars, times three ads in between every match, means the player is making 540 dollars per 45 minute match, which is a generous assumption for match time with the average probably being closer to 30 or 35 minutes. That is certainly enough money to live.

Economics aside, the top tier players also benefit from the open streaming arrangement because it lets them assess their opponents play. Since most of the top teams have someone streaming, it is easy to scope out what sorts of item builds, strategies, team comps, etc. people are using the most and get an idea of their play style. While some people may reserve their best/secret builds or strategies for tournament play, it seems like most of that is talk (ie the secret weapon pick that never happened at IEM Kiev). Keeping at a high ELO, which is how most of the streamers got famous in the first place, requires them to be on their game, if not their absolutely best every match.

The problems occur when the rest of the community, those 6000 viewers and their friends, decide to ape the strategies, runes, masteries, and item builds of the top 1%. While many of the top tier LoL strategies are successful or can be successful at every level of play (Ryze, for example, is a champion with no skillshots whatsoever who is as easy to pubstomp with at 1800 as at 300 ELO), there is an unfortunate assymetry in the metagame. To consider the metagame as a singular, community-encompassing thing is precisely the mistake. The people playing at 500 ELO and those at 1800 use entirely different strategies (or lack of them), making different champs, item builds, and team comps viable. This is precisely the reason that people "smurf" (making new accounts in order to play with bad players), because they can get away with wacky, fun, ridiculous strategies while completely dominating those worse than them. Most of the problems, I imagine, occur in the 1000-1600 ELO range, where players pay attention to their build, team comp, and overall strategy. These are the players that watched Own3d.tv and seeing toplane Shyvana utterly dominate a tournament (again IEM Kiev), decide that they too should play Shyvana, except they do not have the encyclopedic knowledge of jungle timers, insane map awareness, or pro team in vent backing them up that make the strategy effective. This can lead to a flavor of the month effect that actually makes everyones playing experience worse. When the flavor of the month is an overpowered champion that is easy to play, things are not so bad. When flavor of the month is the skillshot-oriented AD Kennen, people who queue with tryhards are in for suffering. Playstyle variety is also effective. When people on your team bitch at you for playing anything other than champions that are high on the widely viewed tier lists put out by Reginald or whomever, your playing experience suffers and you decide, next time, to go with the "safe" pick. The experimentation that the top tier players use to actually come up with innovative builds like AD Kennen is frowned upon if you are at lower ELO.

The fundamental problem seems to be that peer pressure from those playing-to-win, by no means a bad thing on its own, reinforced by an always evolving but continuously monitored 24/7 televised metagame, makes the whole community try to play in ways they may not enjoy. I don't really have a solution other than suggesting that people take what they see on stream with a grain of salt rather than as gospel truth, or compare the top-tier metagame with what they see as effective at their current ELO before deciding what to choose, but I think it's a fascinating community action problem.