Rejects went up a bit but seem to be sitting steady around 1.3%. was always around .2-.5% before. So, I am happy with the new system. Hope it helps us all out in the long run. Also a quick question. I had never noticed this message in cgminer before, what does "new block detected on network before long poll" mean? Just curious to its exact meaning. Thanks and Keep up the good work Mr. Waffle. 
The new server sends things in two different ways essentially. One of them (the one you mentioned) is a coin switch, the other (new block) message is for a new block on the same chain. Just something small with how we're sending the new data when switching work
Last couple hours are lower, but doesn't look to be outside of normal variance at all... From my admin panel:
Okay. I was looking at the WaffleStats page>Details>Statistics>Last Hour>BTC/Day/1Mh just now, and it was showing 0.0025. Didn't think normal variance went that low.... But I will trust in PW.
Yep, normal variance will definitely drop that low
With our current hashrate (12GHs), the following are [coin] [diff] [avgtime] for some of our larger coins:litecoin, 5735.89666335: 2,083.87
dogecoin, 1395.90888910: 507.14
feathercoin, 231.53089248: 84.12
megacoin, 28.30226493: 10.28
worldcoin, 19.08832739: 6.93
Dogecoin being almost 10min blocks (on average), we'll routinely see 20-30 min gaps. LTC at 35min avg, we'll constantly see whole hours go by with 0 income (if we were mining it full-time). So we'll have some hours where we switch on and off of LTC, and don't find a block, leading to close to 0 income for that hour. On the other hand, we'll have hours where we find 2 LTC blocks, and see a much higher number