I’ve been doing some research into how I can reduce input lag when playing TBC Classic and I’ve run into a few curiosities.
Primarily, I encountered the CVar SpellQueueWindow which is by default set to 400.
Looking at previous posts on this subject, they were all from Classic WoW, pre-TBC, pre-Legion Client and pre-batching removal. Those posts state that changing this variable does nothing due to it being a variable for the Legion client, and it may only appear to be an improvement. Is this still the case?
At the start of TBC, the Classic client as far as I am aware was updated to the Legion client, and spell batching was removed. In my experience from changing the SpellQueueWindow variable from 400 to 200 (my latency plus 150 - ubanster) my experience as a hunter feels significantly smoother with far less glitchiness in the swing timer I use and the latency (input lag) at which I perceive my button presses to respond with actions.
Is this the pseudo improvement mentioned in the previous posts regarding this subject? or with the changes to client and batching is this an actual improvement?
If this is an actual improvement, what number should I actually aim for? Referring to this ubanster article: fix-wow-key-input-lag-latency/ it says to add 150 to your latency, but it is from early 2018. Where does the 150 come from and has that changed at all with the removal of batching?
As the 50 from my latency is in ms, presumably the additional 150 is also ms and the CVar is milliseconds. What does this additional time value represent? Is it a constant (from server processes), a measure of reaction time, or keyboard latency? Something else?
Note: none of the commands on the article were working for me, I had to use the AdvancedInterfaceOptions addon to adjust the CVar.
Setting it too low and you’ll leave lots of gaps between globals or be forced to spam your finger super fast (within 0-50ms etc). Recommend PVE’ers keep it at default 400 to reduce carpal tunnel (you can press slower/less accurate and still get the global to go off immediately because the spell is queue’d). For PVP/arena, not sure, but I think the +150 over ping is a decent idea.
The Spell Queue Window does not affect your input latency but provides a mechanism to circumvent input latency all together. You can queue a spell to the server before your global cooldown is refreshed, meaning the server will queue it to be cast immediately after.
So if your window is too large it may feel like you are getting more lag because you are queueing up spells from a previous input and your current input doesn’t register.
On the flipside if your window is too small, you will never queue spells and will always have your input latency added onto your spell casts.
Healthy window is 90-120 for most people, 400 for pve where you are spamming 1 button.
This will help I think, but everything said in this thread is on the money basically.
/dump GetCVar("SpellQueueWindow")
400 ms latency default… =/ this is poop if expect your spells to go off when intended in some cases.
For example if your latency to the server is 100 ms you can reduce the spell Q latency to whatever you like by the following command line.
/console SpellQueueWindow 100
This will change your spell Q latency to 100ms, but say for example you have a 100ms connection latency then maybe its a good idea to stay above your latency because in some ways the spell Q thing helps with that a little bit in theory; at least how I understand it.
You can turn it off entirely by replacing 100 with zero or whatever value you like.
The other place spell Q works out great is when you’re chain casting frost bolts for example, and you want them to stack up perfectly; with the spell Q its possible to do this flawlessly without any real skill. But if you have too much spell Q like the default 400ms it can delay things like counterspell or POM sheep.
From what I understand the Spell Q is harmful in too high a value and too low a value; so some tuning on your end will be required based on your play style and class.