There's actually two questions to be answered. "Why do we play A meta?" and "Why do we play THIS meta?"
At present, League of Legends suffers from (or is blessed by, depending on your perspective) a very stagnant / stable metagame. When LoL players say "meta" what they're referring to is the general strategy and team comp typically employed. This strategy has persisted despite significant changes in the jungle and items from Season 2 - Season 3. So what IS the current meta?
ADC (auto-attacking DPS champ) and support - generally bottom lane on the map below
Bruiser - generally top lane on the map below
AP caster (bursty magic damage champion) - generally mid lane on the map below
There's some caveats to this and we'll get to those in a minute but these 5 comprise 90% of the games played competitively both before and after season 3 started. What I wanted to get to is the reasons why this meta gets played and why it's so stable. The first question above, "Why do we play A meta?" is relatively easy. Solo Queue is generally a collection of individual and duo queueing partners. They don't have the time to do a coordinated, complicated strategy. They need a template to work off of so that the team's expectations are set. So while Talon may be a great counterpick to LeBlanc mid, if you pick him you throw off the physical / magic damage balance of a standard team comp and those ripple effects can't always be handled in the 3-4 minutes of ban / pick time. As a result, you tend to see a little more innovation from true team play (Fnatic winning Dreamhack with a triple bruiser comp or M5 in IPL4 with a double jungle, for example) but in solo queue, innovation is greeted by calls of "Troll Pick. Screw it, I'm feeding!"
The broader question of why THIS particular meta is a more nuanced, but arguably more important, question. The answer lies in an understanding of some basic facts about the map.
- Baron is closer to top lane and Dragon is closer to bottom lane.
- Dragon spawns at 6 minutes. Baron spawns at 15, and for all practical purposes, generally isn't killable until 20+.
- Mid lane is the "shortest" lane both in nexus - nexus distance and outer turret - outer turret distance; the latter is probably more important.
- XP is shared based on proximity (with a 10% bonus to each player involved) but gold is based solely on final blow.
- Auto-attacking characters tend to scale better off of gold while casters ("ability oriented characters") tend to scale better off of level.
- A balanced team of magic and physical damage minimizes build efficiency for the other team; i.e. armor and magic resist are least efficient, gold wise, when you're dealing exactly 50/50 split magic vs. physical.
- Any strategy that can be neutered just from bans and a single counter-pick can't be considered stable.
So let's look at what these mean in aggregate
- Control of bottom lane (i.e. Dragon) is more important early game than top lane, so playing two here is probably best.
- A support is best teamed with an ADC because they can feed all their gold to their laning companion while leeching xp.
- Jungling provides a captive source of gold and xp that needs to be "mined" for the team, so we need someone (or multiple people) to be responsible for clearing the jungle. The jungle also gives the ability to gank and move around unseen. For these reasons, a dedicated jungler is generally preferable to the laners simply picking up the jungle neutrals as they have time. However...
- The jungle lacks sufficient gold/xp to support two junglers.
- AP casters tend to be squishy so putting them in the middle where they have a shorter distance to safety makes sense.
- Top lane is most exposed due to have only one person and being isolated from the rest of the team.
- Because top lane is so exposed and mid lane generally contains champions like Katarina and Diana (pre-nerf) that snowball hard, the junglers will frequently focus on ganks here. It helps, as well, that 2v1 ganks are easier to pull off than 3v2 ganks.
So the natural team composition that flows from this is:
- An ADC and support take the bottom lane.
- The AP caster mid, where he's least exposed due to the proximity to the tower.
- Bruiser at the top. Most bruisers tend to be AD focused but there are plenty of exceptions.
- Jungler (which are generally hybrid damage dealers but, again, lots of exceptions) jungles.
That doesn't quite get to the goal of 50/50 magic / physical, but it's pretty close, and the support will typically be contributing some magic damage as well. The actual ratio will vary wildly depending on the choice for top and jungler. If you play Malphite and Amumu, you're going to have a LOT of magic damage. If you play Lee Sin and Hecarim, it'll be far more physical.
There are three developing trends in this otherwise stable meta.
The first is the trend in lane swapping. Many purple teams (more on why this is largely a purple team phenomenon in a minute) are opting to send their ADC/support duo top and create a 2v1 on top and bottom. Riot has admitted (and recent tournaments have validated) that the blue side has a "slight" statistical advantage. This lane swap evens out those odds somewhat. Generally there are a few reasons why teams switch. The first is that they're avoiding a perceived bad matchup (either top or bottom) or they have a hyper-carry like Vayne or Kog' Maw who would benefit from an easy laning phase. The other factor, and I think this is the reason that you don't see the blue team initiating these lane swaps, is the location of red buff. Red buff is bottom lane adjacent for blue side but top lane adjacent for purple side. As a result, the decision for purple to lane switch is a lot simpler. What AD carry wouldn't want to have a 2v1 lane and have the red buff in easy access? The answer is none!
Remember too, the goal is to catch someone off guard and quickly take down the top tower. If you find your team swapping, you're effectively giving up control of the Dragon so it's important that they aggressively counter jungle and ward to maintain vision and disrupt the inevitable Dragon attempts that will be made.
The second growing trend is away from traditional "sustain supports" to more aggressive supports. A "sustain champion" (Soraka, Taric, etc) is one who specializes in keeping their laning companion active and alive. This is the closest thing to a "healer class" that we have in LoL. The more aggressive supports (Blitzcrank, Alistar, etc) work to pressure the lane and focus on the assumption that a dead enemy ADC is better than a heal on your ADC. The current trend is to pick supports that can grow into carry or pseudo-carry roles. At present, Zyra is very, very popular at high ELO and competitive play. The reason is that she can be played very aggressively (landing her root is often all it takes to kill an enemy) and she also grows into a second AP carry late game once she can get some gold. As a result, her power level is highest at the two phases of the game that are most important (first ten minutes and last ten minutes).
The final "small" trend is to have an AD caster mid and an AP bruiser top. This is a fairly small tweak but it's definitely more common to see a Talon mid and Malphite top than it used to be. The problem with this is that there aren't that many AD Casters in the game right now. Talon is clearly the top tier of this small type with Pantheon and Zed being the other members of the family (Rengar and Kha'Zix could, in theory be one but they need bushes to be effective and there aren't any close enough to work in mid lane). Zed has a crazy high skill cap that even the pros struggle with (go watch Chaox's play of Zed from this past weekend) and Pantheon is generally spurned (though I'm not sure why). Three champions is insufficient to have a stable meta, however. The list of AP bruisers is similarly short (Malphite, Vlad and Irelia are the ones that come to mind). So banning Talon and Malphite (something frequently done lately, from what I've seen) can really hinder this plan.
What changes has season 3 brought? Not many to be honest. The potentially big meta shifting events (nerf of oracles and sightstone which means a lot more wards, jungle changes, re-itemization) haven't yet pushed any big changes. There was a period of time where Black Cleaver ran amok and you saw teams of 4-5 AD Bruisers running around but that's been "fixed" with the cleaver nerf.
There hasn't yet been a tournament played with Season 3 post-cleaver patch. There was a solo-mid series tournament where the final round was played season 3 but with cleaver still in its OP state (pre Nami / Vi as well). As a result, it's probably not worth exploring too thoroughly for clues as to the meta. We'll have to see what happens in the next big tournament to really get a good feel. In the meantime, get some friends and explore. Today's "Troll Pick" can be tomorrow's meta-breaking discovery.