So lets say I am CPU mining, each share I solve has just as much a chance to find a block as anyone else.. So if I leave my CPU miner running for 6 years when it does finally solve a block its not like it solved a 6 year old block that is no longer valid.. It means it solved the current block and finally (after 6 years) got lucky...
Is that correct?
Yes. However, keep in mind that difficulty is adjusted roughly every 2 weeks, so if it take longer then one difficulty adjustment to solve a block, you'd be better off being in a pool.
By the way, I have read that wiki page.. I read it a few weeks ago, but the lifecycle is still kinda murky... I understand a group of transactions are bundled into a block, then the block has to get hashed into the chain.. And somewhere in there a block is solved and someone (after 120 confirmations) wins 50 BTC.... What I don't understand is the path from bundle of transactions to 50 BTC, and the queing the network uses to apportion work..
You bundle whatever transactions you want along with a 50 + fees BTC transaction to yourself. The transactions change as some are put into blocks by other nodes and new ones are received. They aren't locked until you find a block hash that meets the required difficultly and submit the block to the network.
I am trying to understand so that I can decide if I want to spin up a private pool (of about 10GHs between me and some friends) or stick with my current pool. I am less concerned with variance as I am winning a block but it becoming invalid due to how long we went before solving a block. (after all variance levels out over the long term)
The network doesn't assign you work, you make your own. The network only cares about the result. A block can only be made invalid if someone else finds one around the same time as you, and your node didn't know about that block (the blocks link together in a chain).