There have been times in the past where they didn't bring the servers down on their regular periodic basis and a lot of players started experiencing issues... Servers are not magical entities; they are computers and they end up with memory fragmentation, tons of disconnected handles, disk errors, and so forth.; a "megaserver" setup is almost certainly worse than most, begging the question as to why any company would go that route... Those sorts of problems went away when ZOS finally did bring the servers down for a reset. That being the case, we have ample evidence that there are server issues. Server code is still code...

I'm on the US server; as an old dog who's been in this racket, I'm pretty sure what we're seeing isn't lag per se. The netcode for ESO is a problem; the original code -- which worked well -- was modified because of hackers. The implementation they're currently using requires the engine server-side to acknowledge pretty much everything you do on the client.

Which isn't necessarily a smart thing to do. Nobody likes cheaters, but an ACK-NAK system isn't the smartest way to address that problem.

Things like having to press the "Q" key twelve times to get a character to pop a potion... or weapon swap becoming an unreliable function again... or block/dodge not happening when you need it... or freezing in place when you need to move... all the things that get your characters killed and frustrate the hell out of us... the client sends a message to the server and the server has to send an ACK back.

Which is a small message... however, it's most likely sent UDP which is not guaranteed delivery. If they used TCP/IP -- a grossly inefficient protocol with a rigid messaging structure -- in order to get guaranteed delivery, the bandwidth requirements would eat them up and present a whole different set of problems... like how to have more than two players (I exaggerate, but that's the crux of the problem).

With UDP, you do lose messages up and down the network links because it isn't a collision-free environment. Your client and the server almost certainly use a message counter that is used to determine if a message has been missed at either end and a RETRANS request is sent if the recipient determines that it's missing something. That's how the netcode gets around the non-guaranteed delivery problem. Lots of communications systems use similar schemes because it's a no-brainer and it's a thousand times more bandwidth efficent than guaranteed delivery systems... [It's actually nothing more than basic freshman year queuing theory, but network engineers like to pretend it's rocket science.]

So you can have a wonderful ping and that number is absolutely meaningless. And the experiences of different players are not going to be the same (so the ZOS-apologists need to stuff it).

There's still processing going on at the recipient's end in prioritizing the handling of messages. And that is where they start playing optimization games. They're not changing code (they'd have to bring the system down for that), just event prioritization. Set up correctly, they can modify the priorities of events server-side without ever bringing the server down by simply setting them up in a flat file.

The problems we are experiencing appear to be due to that.

Guesswork on my part; but educated guesswork. I've been out of the business and retired for a while; but given the inconsistencies of game performance we're seeing right now, that's my bet.

Just logged in and my ping is around 70 now, first time i have seen it that low in well over a week!