According to what i've read between Scrypt and Skein, 1mh (scrypt) is equivalent to 300mh (skein).
We have a median of 100mh on p2pool scrypt and 3gh on the skein p2pool. (just taking p2pool stats)
So 100mh scrypt is equivalent to 30gh skein if i'm right, so we have 10x time more power on Scrypt p2pool but we find much more less blocks on the Scrypt p2pool than Skein p2pool.
Correct me if i'm wrong but if i'm right that mean Skein is much more easier with 10x time less power... so their might be a problem with difficulty retargetting on Skein

Taking the last 10 hours worth of data available from
http://insight.dgb.cryptopoolmining.com I've come up with the following stats:
Groestl........185 blocks found
Qubit...........182 blocks found
Scrypt.........113 blocks found
SHA-256.....196 blocks found
Skein..........190 blocks found
based on blocks found, maybe the problem is on the difficulty retargetting on scrypt, all other 4 algos are about ~190
the difficulty should adjust such that a block is found every 30 Second Blocks (2.5 Min per Algo)
2 * 60mins * 10hrs = 1200 blocks
1200 / 5 = 240 blocks per algo in ten hours, but scrypt is way down only 113 blocks