GDFireFrog

This is fixed in 469b, going out now.

We've fixed this in 467b.

You still have this issue in 466d? (Check the link Raven posted for context)

Update: We just released a minor version (466d). This keeps the same improvements mentioned in the patch notes but adds a small grace window of time that also allows rocket-jumping with the previous shoot-first timing. From our testing this also improves rocket jumping in this scenario compared to the situation pre-466. (Update update: Sorry for update right before finals, eggwp, we have been testing this all evening and forgot to check.)

Yes, of course. Go ahead.

We had an issue rolling out this update, the new mode should start working in about 10 minutes. (From 21:05 CET*)

We are having some server issues after the update, the game will be down for a bit. I'll update when we have a ETA. (Also forgot to put say in the patch: shout-out to Netborg for lots of testing and suggestions and bug reports related to performance.)

The note is a bit misleading, yes. The waypoint is indeed a sticky one.

I asked somebody in this thread with issues but he's not having those issues at the moment. I need somebody who's experiencing the issue right now to DM me on discord or on reddit or send me the log of a tracert/tracepath/WinMTR to the IP 109.230.214.11. We can drop the provider for this location, but first I need to see the issue to give them a chance to fix it.

Looking into this. Will update here.

Yes, sorry, left some WIP in accidentally and we didn't notice it during testing. Pushing a fix in 5 minutes with 417c.

Fix is up in 412.

That may be unpractical because of the Eggbot dimensions right now. But also, good players are already hitting 70%, things may get a bit out of hand if we keep making it easier. (As an anecdote, somebody did a test setting the metrics to be more like QL and they were hitting 80%, about 25% more of what this person hits in QL.) But in any case, we have identified an issue (see my reply to Ardi46) that may be affecting people's impressions. The PNCR effect is delayed by what seems to be at least one frame, this may lead people watching videos to read some things as a hit that are not. So we are going to first get this issue out of the way and see if the feedback from these players improve, no need to consider anything drastic just yet.

Thanks, I'm going to focus on hitscan registration for now, and you have only provided a clip of that. In that pncr clip where you think it's going through this is where you actually shot: https://imgur.com/a/9rKcS0w

You can tell if you play it frame by frame because the pncr fire animation nudges the pncr upwards a little bit when the attack is triggered. Because you are moving the mouse to the right and the particle effect is delayed one frame this gives the impression you hit, but you didn't. If this is just about the particle effect being delayed one frame it may explain why some people feel registration is bad even thoug...

The last time any change was made that could potentially affect hit registration was about 2 weeks ago, and this was announced. We are not doing any kind of unannounced A/B testing on the regular as you suspect. Every time netcode or anything related to netcode has been changed, it has been reported in the patch notes. The person who pointed me to this thread asked me if maybe we do secret A/B testing to remove bias. Just to be clear, we don't, and I don't think we'd ever do that with netcode. If we wanted to do that kind of blind testing with more people than we usually test with, we'd put together a user group. We are not gonna be doing those kind of unannounced changes, especially, when tournaments are happening.

Let's focus on "but the reg seems random as hell". We do test hitreg quite a bit, analysing the data to make sure everything is working as expected. We also record video from our tests, note the failed shots that could potentially have actually been hits, then w...

This was also fixed in 406d.

The precise level of curation is something that we'll have to discuss internally, but I can't see us wanting to enforce uniqueness of concept. We'll probably be fairly hands-off, with the exception of highlighting work we may consider outstanding in some regard.

This kind of stuff plus what _sohm mentions in his comment should be coming shortly.

Hey, thanks for reporting it. This was fixed in 406d.

We'll do both. Thanks!