The concept of a "Lazy Masternode" was pointed out by taushet recently. I also alluded to the same concept a few weeks ago (
https://bitcointalksearch.org/topic/m.17760676). But this is just the first of many issues that the Masternode concept has.
Let's just assume for now that it's actually possible to check if a certain masternode provides the expected service. The network should check if a masternode has a full blockchain, is serving blocks to nodes who are syncing, is relaying transactions, is doing the privatesend coinjoin mixing and is performing the instantsend transaction "locks", ... This is not at all that easy to implement. It may not even be possible. It's not my job to prove it's impossible, it's the job of the DASH developers to show it's possible by implementing it in stead of just having a "Proof of Ping".
The problem with PoSe is that you can't differentiate between "lazy masternodes" and masternodes who have troubles with their connection. Currently, there is a 1 hour "grace period" for masternodes who are offline. If they manage to get online again within an hour, they don't loose their "spot" in the line for the masternode payouts. This is done to prevent DDOS attacks: if this grace period would not exist, it would be very easy for a big masternode owner to gain additional payouts by taking other masternodes offline. The higher the price of DASH, the more incentive there is to perform such an attack.
[Note that even with a grace period of 1 hour, this attack is still possible, you simple need to attack a node for a full hour. You can have some bad luck if the owner of the node notices the attack and manages to move his masternode to another server within an hour though.]
Back to PoSe... It's impossible to know if certain masternode has problems with their connection (for example because they suffer from a DDOS attack) or if this masternode is being lazy (i.e. not providing the service): Imagine a that the network somehow checks every 10 minutes if a certain masternode is willing to do privatesend. When the masternode is pinged to check if it's online and ready to provide the service, the network notices it's offline. Then the network can for example try again 10 minutes later to see if the masternode is now able to facilitate the privatesend. Again, it failed. The network will keep trying again and the masternode doesn't respond until his "last chance": at the end of the grace period, then the masternode facilitates the privatesend transaction.
So what happened in this example? Was the masternode just being lazy? Or did the masternode have connection troubles and managed to get back online just before the end of the grace period? Also note that all those failed checks slow down the privatesend process.
Conclusion: either the PoSe needs to get out of the window, or the 1 hour grace period needs to go. If DASH chooses the former, then it admits that PoSe is just BS. If it chooses the latter, then "The Great Masternode Wars" can start. Masternode owners will then have an incentive (depending on the price of DASH) to attack other nodes to gain a bigger masternode reward.