Well, as long as the updaterate and cmdrate are the same or close to each-other and 30 or above, and 66 or below, with a "sane" rate, the experience will be good. you hinted already: ping 100 or higher is really bad for experience (personally, I start to notice it on ping 60 and higher). But that's ping, nothing to do with these settings.
I agree that some servers have really bad configs in that regards for players, the min and max values are to allow a degree of freedom to gamers, not pushing them too far, and not allowing them to be too low. Or servers that have 10 instances running on a dual core, with a cap limit by their hoster. Unfortunately, there are also still a load of countries where dial-up is still the latest technology they know. So while here in the west we have the luxury of good connection, a friend of mine in some east European country has to call his shaman bout everyday to get his ADSL router working for a measly 4096/256 Kbit. >________________________________ > From: dan <[email protected]> >To: Half-Life dedicated Linux server mailing list ><[email protected]> >Sent: Thursday, 18 October 2012, 15:38 >Subject: Re: [hlds_linux] lagg problem on ubuntu 12.04 > >On 16/10/2012 11:56, Drogen Viech wrote: >> Why can't you just leave the cmd- and updaterate alone? It seriously >> pisses me of when people think "higher == better" - the default >> settings of updating and sending 30 times per second are just fine. > >I'd say they definitely matter. There's no higher==better of course. >I think they should be 66 - the tic rate of the server. So fixed values. > >Higher shouldn't, aiui, achieve anything. But is there any sane reason for >them to be lower? Unless there's >some country with a significant dialup population (but their experience is >going >to suck anyway) or if someone can show (in a robust way) that the lower values >make >no difference. > >The point is, I don't understand why you can even change these values in the >first place. >Why aren't they constants? And 66? > >The same with rate. <= 30000 and its anecdotally not as good. That's over >several >thousand hours of playing. I can tell without looking if I join a server that >has either capped max_rate or >forced it to a particular value compared with the client setting I have. > >But, equally, there are a set popular servers with sv_min_rate 90000 for some >bizarre reason, and they seem just as broken as >the ones with it set low. (It's not clear whether their performance issues are >because they have hundreds of servers though or >whether it's the actual config causing them) > >But, I don't really understand why the defaults are what they are. >The defaults should be right in the sense that >- There should be no need for servers to cap anything based up a feeling they >have that clients could break something if they don't. Or that without the >caps it would give some players an advantage over others if they aren't aware >of the client settings. >- There should be no need for servers to change things today because the >defaults are years out of date. >- This should apply to the client settings too. if I can change my settings >and get a better experience - especially if that gives an advantage or edge >over other players, then either the defaults are wrong and/or the fact these >settings are hidden is wrong. > >Although I'd tentatively agree with you that server admins pissing with them is >one of the worst aspects of 3rd party servers it's not because they make them >bigger. >It's usually the reverse. > >Especially now more of Valve's games are forcing matchmaking and, as we saw >last night, as >Valve are trying to slowly hammer a wedge into the "punish you for leaving >early" initiatives. > >Join a server with sv_max_rate 25000 and I'll usually leave. Perhaps >that won't affect mann up (but Valve's French servers are a performance joke >so I hope they aren't punishing >anyone for leaving them) similarly there are umpteen admin foobars and no-nos, >halfwitted plugin use >that the browser and (obviously matchmaking) does not tell you or let you >filter out before you join. > >Hence, it's no wonder lots of people leave games after finding themselves on a >server like that. Valve's French servers could barely manage 12v12 and cannot >run boot camp MvM - so I'm sure some of this will affect mann up servers. > >It would make sense if Valve are phasing out server browsing and having "you >must finish a game" initiatives, to improve the server matching so it takes >into account not just ping but these parameters too, and plugin use. A good >first step might be getting sane defaults in place if they genuinely aren't >sane now (and I would say anecdotally setting cl_updaterate and so on to 66 , >rate to 60000 and cl_interp to 0 (which gives you typically 30.0ms instead of >100.0ms but some server admins have managed to break this) is a much better >experience and an advantage over anyone who uses the defaults. > >So either I'm wrong or there are shedloads of badly configured servers out >there hurting the experience of their players and, obviously, Valve's client >defaults are not helping either. > >Perhaps the real issue is, no one has any real idea what the values should or >shouldn't be bar from their anecdotal experience from trying the client. And >the page written by valve talking about how the network compensation works >that touches upon some of this. Which sounds reasonable at first but, as >anyone who has played with a ping >=100 will tell you, in practise it is a >load of hot air. There is no compensating for a high ping or at least, Valve >certainly haven't implemented it - if your ping is high, the game sucks. I >believe that's equally true of these other values too like rate and cl_cmdrate >and so on - and that the better you are at the game the more sensitive you >will be to them. > >-- Dan. > >_______________________________________________ >To unsubscribe, edit your list preferences, or view the list archives, please >visit: >https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux > > > _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: https://list.valvesoftware.com/cgi-bin/mailman/listinfo/hlds_linux
