Why no one addressed the elephant in the room

[Reinhardt_Azureheim]

Honestly, to me the hit detection changes have actually been fairly beneficial and allowed me to do things I couldn't do otherwise, like moving AOEs before they go off so they snapshot better. I also do not have any issues with the Purify reaction time because it is nigh-instant for me, same for Guard - I guarded vs Primal Rend about 0.1s before impact the other day.

I am not discrediting the experience here, but there is clearly a factor they completely ignored when designing these changes - LATENCY.

For context I play at 10ms (10 Ping) as I live in Germany connecting to a german datacenter (Light + Chaos) through a VPN to ensure I get a direct connection with basically no hiccups. I literally do not have "any" issues regarding hitreg and anytime I miss, it's actually my fault.

Inversely, this means that this netcode is working, but is absolutely brutal at latency exceeding 50-70ms. Things like stuns hitting after pressing Guard making Guard go on cooldown, Resilience not protecting against follow-up CC in time, AOEs not landing (particularly position based AOEs like Holy Sheltron bomb, Impalement, Bioblaster, etc.) are MAJORLY affected by latency, as you are actually no longer snapshotting at a single point in time where you know the player is for sure, but you have to track them where they actually "might be".

It will favour you at low latency, it will be basically soft-gambling at high latency. Given North America Datacenter location(s) generally being unable to favour the entire region at once and JP being a relatively small region (geographically), it makes sense they did not consider high latency when designing the netcode changes.


"Technically" how it works is just different from what we are used to and can be learned, but our network infrastructure world wide does not allow for low latency gaming worldwide, let alone regionwide. As long as this is a factor, this will remain a problem, therefore this netcode without adjustments is flawed.

[Loggos]

"Technically" how it works is just different from what we are used to and can be learned, but our network infrastructure world wide does not allow for low latency gaming worldwide, let alone regionwide. As long as this is a factor, this will remain a problem, therefore this netcode without adjustments is flawed.

I agree with you that latency is a major issue but I would like to dispute the aspect of "it can be learned" a little because at least for me (also living in Germany) the hits have not just shifted from instant to delayed activation. Many actions have become straight-up unpredictable (for me).

• For example, I heal up in time (according to the feedback on my screen) and I still get killed by a nin LB in one situation. And in another situation I survive.
• Sometimes I finish my casts even though a person has long run out of my LoS and it still hits them. Then again, sometimes the same cast is canceled when the target runs out of range.
• Sometimes purify/guard works normally, sometimes I - and I am not exaggerating - have to press the button several times until something happens.
• Sometimes my LB goes off at the press of a button, sometimes it's completely unresponsive for several button presses even though I am explicitly not doing anything else and there is no GCD rolling. Really, really maddening when it happens in very critical situations where there's really no room for this kind of random misfiring.
• Sometimes in a match I can chain an oCGD to a GCD in advance and at other times my oGCD button press isn't registered. It doesn't matter if it's an offensive oGCD, a defensive one like Tempera Coat or my palette switch. It is legit random when they work and when they don't. I've paid close attention to it over the last weeks and it has been a source of frustration until today.
  (I also tested it on a dummy and there it's a lot more reliable but it still doesn't work sometimes. I suppose in matches where there is a lot more data from up to 10 players that has to be computed at once, actions "getting lost" in the queue of everything that's taking place becomes more extreme, esp. with bad latency.)

The thing is that this makes my muscle memory unreliable because what used to work automatically now may or may not fail me. This can really throw me off and yank me out of my rhythm/my flow. It's difficult to adapt to that because sometimes your muscle memory is right. So you get very mixed/paradoxical feedback/(automatic) conditioning over the course of several matches.

I know that the netcode pre 7.1 wasn't perfect and some issues we are currently experiencing already existed before. But it is my impression that the problems have been exacerbated a lot with the update. I don't remember being nearly as frustrated by the unreliability of my actions before the update even though my latency should be more or less the same.

If it was just a different timing I could adjust. But the fact that I have no way of predicting which situation I'm going to get makes planning ahead much harder because the execution of my actions turns into a bit of a gamble.