251 replies to this topic

[Funkadelic Mayhem]

So this is not longer about all lights and is about the spider?



As it should be!

[Clit Beastwood]

Is everyone still using the on-screen graphics to determine what hits and what doesn't? Thats been clearly demonstrated not to work. Often the person shooting me sees hits while I see a near miss as I jump jet backwards to avoid fire (in my head it's all matrix-y). This is the frustration of HSR - so, what do we all prefer - having to lead targets due to lag, or having HSR show misleading graphics? It works both ways, I've been running around in a spider and randomly exploded from a shot someone took almost a full second before. Thanks obama. I mean HSR.

[Wolfways]

Funkadelic Mayhem, on 24 September 2013 - 07:15 AM, said:

he could not even keep his cross hairs on the spider most of the time. If he cant hit the spider with lasers, its lol he thinks he hit it with his ac's.

Looks to me like the reticule only turned red when he was leading with the lasers (other than the first few seconds where the Spider was moving away or standing still). Keeping them on the Spider would have been useless.

[Kensaisama]

Its all borked, lets hope they get it fixed soon, I grow tired of being a one man army in my light mech

As a light pilot I know how annoying I can be to larger mechs as well as how annoying lights are when I pilot a larger mech. It seems few people have experienced both sides of the fence. As has been pointed out, latency also plays a large role as well, there are times I get killed quickly when I am in a light, and other times not so quickly as hits don't seem to register properly, its during those times I am a legend in my own mind and I am invincible LOL.

[Gaan Cathal]

Ghogiel, on 24 September 2013 - 08:01 AM, said:

And yes there is ZERO indication that the hit was registered by the server>As explained by Thomas, the red reticule flash happens when the server confirms the hit, if the server isn't also confirming the hit, it doesn't happen.

Except that it quite clearly takes the damage (paperdoll goes darker) from the MLas at 0:59, but the reticule flash doesn't happen. Ergo, using that to confirm server hit doesn't actually seem to work (note, the paperdoll doesn't just flash, the colour changes).





Edit: typo correction

Edited by Gaan Cathal, 24 September 2013 - 08:34 AM.

[Ghogiel]

Gaan Cathal, on 24 September 2013 - 08:34 AM, said:

Except that it quite clearly takes the damage (paperdoll goes darker) from the MLas at 0:59, but the reticule flash doesn't happen. Ergo, using that to confirm server hit doesn't actually seem to work (note, the paperdoll doesn't just flash, the colour changes).





Edit: typo correction

It flashes as the targeting info comes up, prior to the lasers being fired.

[Gaan Cathal]

Ghogiel, on 24 September 2013 - 08:43 AM, said:

It flashes as the targeting info comes up, prior to the lasers being fired.

It's actually doing both. If you just watch the paperdoll corner, you see it appear with the damage glow on the RA and CT, then the glow stops, then resumes when the lasers fire.


Basically, the problem is not knowing whether to trust the paperdoll or the reticule, one of them is clearly wrong.

[Villz]

stjobe, on 24 September 2013 - 06:19 AM, said:

That's... A whole lot of missed shots and laser grazing. Was there anything in particular you wanted to point to?

I play with 270 ping so i have to shoot infront of the spider and it varies with ping fluctuation (jitter) i'm demonstrating how the damage indicator wont register red at all when i shoot the spider it has to be front of it i give up with you fools

[Villz]

Amsro, on 24 September 2013 - 06:17 AM, said:

I've even put my spiders in storage until this whole nonsense is cleared up, I refuse to let gimped coding take away from my spider piloting prowess.

I prefer to use the derp mode worst chassis and see how well I can fare. Originally it was the spider, now the spider is OP and hit detection free!.

Cheers to DV8 and others for looking into this in depth.

also i'm not dv8

**** dv8

[PEEFsmash]

So this whole thread is kindof stupid now because I posted it when HSR+Hitreg had been quite good for 1-2 weeks. Then, late Saturday>Monday has been bad as far as registration again. Not on lights but on everything as people have posted. Sunday night (for NA) was PARTICULARLY bad.

I was confident that PGI had solved the issue because it was solved for a couple weeks, but the network has been going haywire so....my whole endorsement of HSR/hit detection working is now on hold until that gets back to normal.

I guess this is what happens...things seem fixed, then they unfix =[

Villz, on 24 September 2013 - 08:57 AM, said:

also i'm not dv8

**** dv8

He said DV8 and others. Are you not others either?

Edited by PEEFsmash, 24 September 2013 - 09:02 AM.

[Villz]

Funkadelic Mayhem, on 24 September 2013 - 08:06 AM, said:

So this is not longer about all lights and is about the spider?



As it should be!

I have the EXACT same issues with spiders jenners and even cicada's although not as bad. The issue is to due to HSR not working properly as its "ping sensitive" according to the whiz kids. The spider has just such a narrow hit box its compounded. The faster the mechs moving the worse it is. Its 90x worse up close than at distance also.

PEEFsmash, on 24 September 2013 - 08:58 AM, said:

He said DV8 and others. Are you not others either?

edit in my vid to the op ty

[Gaan Cathal]

Villz, on 24 September 2013 - 09:01 AM, said:

I have the EXACT same issues with spiders jenners and even cicada's although not as bad. The issue is to due to HSR not working properly as its "ping sensitive" according to the whiz kids. The spider has just such a narrow hit box its compounded. The faster the mechs moving the worse it is. Its 90x worse up close than at distance also.

How much jitter do you have, out of interest?

[Mehlan]

Peef, it's still valid....really need to know thins like ping etc anyway. Performance for a local is still going to differ from someone in say....Podunk Portugal.

[Villz]

Gaan Cathal, on 24 September 2013 - 09:04 AM, said:

[/size]

How much jitter do you have, out of interest?

well that varies ofc

Luckily aus -> West coast routing is great usually 1 long cable from sydney to san jose the variation is usually 260 - 280 stable but sometimes i can go as high as 295ms

It gets worse at oceanic peak time which is from 5pm - 1pm our time but this video was recorded at approx 1am my time. So definatly not network congestion on local side

Edited by Villz, 24 September 2013 - 09:14 AM.

[MaddMaxx]

Let's not forget, that the Internet is a maze of connections and while robust, and can self correct traffic flow quickly, from day to day, natural disaster to natural disaster, save the oceanic cable lines, you routing may change as much as 2 to 3 hops depending on where you are coming into the PGI system from.

What many want PGI to do is FIX the Internet. Christ himself (or your deity of choice) couldn't fix the Internet.

Run trace route to PGI's server address and see how many hops you get and packet lose values. Do this once a week and see if your stable or if it fluctuates.

[Gaan Cathal]

Villz, on 24 September 2013 - 09:05 AM, said:

Luckily aus -> West coast routing is great usually 1 long cable from sydney to san jose the variation is usually 260 - 280 stable but sometimes i can go as high as 295ms

Mehlan, on 24 September 2013 - 09:05 AM, said:

Peef, it's still valid....really need to know thins like ping etc anyway. Performance for a local is still going to differ from someone in say....Podunk Portugal.

This is making sense, to be honest. If people are having issues with HSR due to ping-jitter, which we've known has some effect for a while now, then light mechs, particularly those with narrow, gappy, hitboxes like the Spider, are going to show it more. It's a lot easier to notice when it puts a round off target than when it puts it in the wrong hitbox, for one thing. It would also explain why there's so much variance in observation (from "all the damn time!" right through "sometimes" to "what HSR bug?"), which will compound easily with the ever-present issue of observer expectation bias (some people are going to be calling misses mis-registrations, let's face it).

I'm suprised if it's manifesting with a jitter of 20ms though, that's not really a huge amount in the grand scheme of things, although mind you I imagine most people aren't in the habit of checking their ping in the middle of a brawl so might be missing bigger situational jitter. Pity we don't have the option to have our ping displayed constantly like we do with FPS, that would make tracking down these issues a lot easier.

[Villz]

MaddMaxx, on 24 September 2013 - 09:16 AM, said:

Let's not forget, that the Internet is a maze of connections and while robust, and can self correct traffic flow quickly, from day to day, natural disaster to natural disaster, save the oceanic cable lines, you routing may change as much as 2 to 3 hops depending on where you are coming into the PGI system from.

What many want PGI to do is FIX the Internet. Christ himself (or your deity of choice) couldn't fix the Internet.

Run trace route to PGI's server address and see how many hops you get and packet lose values. Do this once a week and see if your stable or if it fluctuates.

Making the system of HSR (a remedy to solve latency issues) dependant on having perfect routing for it work seems kind of smart right?

The truth is they have already said in the past they were planning to goto a HSR system that didn't rely on latency. Which would kind of seem like the smart way to go about it in the first place >.>

[stjobe]

Villz, on 24 September 2013 - 08:54 AM, said:

I play with 270 ping so i have to shoot infront of the spider and it varies with ping fluctuation (jitter) i'm demonstrating how the damage indicator wont register red at all when i shoot the spider it has to be front of it i give up with you fools

Believe it or not, it was an honest question. I was at work and couldn't frame-by-frame the video (since I should be coding...), so I just browsed it once - and I saw mostly misses and glances. Since you made the video I wondered if there was a particular moment in there you wanted to highlight.

[PEEFsmash]

stjobe, on 24 September 2013 - 09:44 AM, said:

Believe it or not, it was an honest question. I was at work and couldn't frame-by-frame the video (since I should be coding...), so I just browsed it once - and I saw mostly misses and glances. Since you made the video I wondered if there was a particular moment in there you wanted to highlight.

The problem with Villz's video is that we never see the paper doll of the Spider until the very end, so we don't know if the damage was applied correctly or not.

[Roland]

Being able to identify issues like this is yet another reason why private lobbies would be useful.

Back in MW4, we used to extensively test {Scrap} so we could better work with the FASA guys to identify and fix bugs.

The one environment where this is easy to do in MWO, is the test arena, where it is specifically not using the same game logic as the real game, thus making testing in that environment pointless.

In addition to private lobbies enabling all kinds of things which players want, like league play, team practices, etc... it'd also help us identify and reproduce bugs in the game much faster than we've been able to thus far.