Alright guys.
So I've been watching the impact of the bumped (hard-locked) difficulty of 512. From my side, we're looking rediculously better. All around load on servers is down by about 70% due to the adjustment. This is from a few factors, first being we don't need to track every miner, and all of their previous shares (and times) for calculating and readjusting difficulty, second is the reduced amount of shares we have to log into the db (and then process into rounds, etc), third is the reduce in bandwidth required, which allows us more upstream bandwidth for outgoing block propagation.
All of these things combined makes for a _much_ better server-side. From a miner standpoint, I can understand why it is a slight negative (smaller miners don't see shares nearly as often, may think something is broken). But unless there is a major outcry about this, I think we need to stick with it.
The next thing is, we've boosted connectivity (a lot to do with bandwidth changes from difficulty), so we should be seeing fewer network-orphaned blocks, which is extremely good. A bit of that was showing up earlier on the stats pages (from about 0.0080 --> 0.0086).
On top of that, I've just enabled ReddCoin, KarmaCoin, and PenguinCoin. These are the first of their kind, in that we have to manually trade them (they're not on Cryptsy, where we auto-trade), and aren't listed on any exchanges that support API based trading. What that means is two-fold: 1) Unexchanged balances will be significantly higher than normal (they wont move from unexchanged --> converted very often). 2) Unexchanged balance approximation will be off significantly more than it was previously, since we're exposed to market swings between the time they're earned, and the time they're converted (probably 2-3 times daily). But even with the added risk, I think they should be more profitable on average.
Lastly, I've disabled a few more coins for the time being (under 15 difficulty), as when we switch to them, the block propagation time compared to block-find time is way too high, and results in us fighting ourselves for blocks (orphans). As a simple example, say it takes about a full second for any block we find to make it around the world (propagate fully). With coins under 15 difficulty, we have an average block time of about 7 seconds. Meaning that 1/7th of the time, we might have already found a block, and not seen it yet (found at EU, hasn't gotten to USWest). And we end up orphaning our own blocks about 1/7th of the time.
The long-term solution to this was discussed previously, which is to switch portions of our hashrate to lower difficulty blocks, in an attempt to keep our block-find time at a constant value (ex: never switch more power than would give us 20s block times). But this is a reasonably large change to the profitability switcher, and the easier short-term solution is to just disable those coins until we can make those changes (a few days most likely). So in the mean time we added additional coins (see above).
tldr; added new coins, disabled some coins, don't trust unexchanged balances
For those folks who keep asking poolwaffle questions that were answered two pages back. BTW, thanks for all of your work PW. I was probably the 10th person to join this pool when you first posted on Reddit, and have enjoyed watching the many improvements you have made since then. Also a big thanks to Wil Schrader for creating the Wafflestats page. Great job!