But what does that tell you? Let's say I ran a pool with a 5% comission. So I get 2.5 BTC from each block. What stops me from having my pool software simulate a few 5970 clients, awarding me their score in the pool rankings. That of course would be done with seperate addresses so it wouldn't show up on block explorer. I can't think of a way to protect against something like that.
It's enough just to compare your expected payout to real one. It should be the same on average, every participant can do this.
Second, these pools are mostly being run by one person and lack things like fault tolerance. If Joe pool operator makes a code change that breaks things, the pool goes down. He isn't likely to have a development environment, test deployments etc. So wwe have 2-3% comissions then there is the downtime which reduces efficiency even further.
So in that respect, I am thinking solo provides the greatest efficiency. There is no overhead loss due to comissions or pool outages.
Do you have any statistics on pool outages or proofs of fault tolerance lack ? How would you know if pool operators have test deployments or not ?
Please, don't make such assumptions.
Solo mining efficiency is higher only if you have hashing speed high enough to find blocks faster than difficulty adjustment steps.
