i dunno about net_port 0, but sometimes when i spam "net_port -1; net_restart" my ping goes down..

net_port -1 is the same as net_port 0, each net_restart is binding new random port. But net_port 1 means udp port 1, really.

thanks for the effort

I disagree about scoreboard ping being inaccurate, when my ping drops from 45 to 30 after finding the 'good port', I'm really losing 15 ms, it's blatant when using PG, I'm hitting 5-8% more

As I wrote in article it just your subjective feeling. We cannot -measure- does "good port" really affects on gameplay or not.
Subjective feelings are acceptable but couldn't be used as an objective argumentation.
So I wrote "We suspect" cause we cannot prove it somehow.

DieZel, this might interest you :

http://www.esreality.com/?a=post&id=2548041#pid2548581

yeah, I read that.
Few players agreed with you, there could be something in the middle affecting routes or routing policy depending on client's udp port. But we also tested on "clean" system when client connected to the server on the "direct wire". I mean we used the network that does not have any balancers or routing policies. And problem affects clean laboratory system too. So I have to conclude balancer could provide ping differences in some cases, but net_restart problem clould be reproduced many times with predictable results, and I am sure the problem clearly has another effect and it lies on quakelive client side.

>> So I have to conclude balancer could provide ping differences in some cases, but net_restart problem clould be reproduced many times with predictable results.

And what are these predictable results? That ping difference that could be reproduced? 4/5/8/10ms?
And what are these unpredictable results? That ping difference that can happen due something in the middle affecting routes or routing policy as you said. That difference that can't be used as an objective argumentation because you don't have tools to measure it and don't accept that players feels it.
And what we care more as players? That tiny but stable difference in latency we get on every server by net_restart or some higher random added latency on some servers?
Ask this this questions for yourself.

Predictable results: I can manage indicated ping in the game whatever milliseconds it could be. I met no upredictable behavior during investigation. All the facts I found are explained. My goal was to explain what happening when you're calling net_restart but I didnt even want do anything to make game better just because it is quite impossible. I cannot provide any care to players accepting personal feel of everyone as physically observed and reproductivable fact. If you have some let we do the new investigation.

>> I met no upredictable behavior during investigation. All the facts I found are explained.

You Know Nothing, Jon Snow.

A "clean" system? "Direct wire"? So basically tested on a local network? If so, how could there be any latency? Or did you use a VPN with one endpoint on the client machine and the other on the server machine? Or maybe some kind of program to just warp/unwrap UDP packets into bigger UDP packets so the header doesn't change with a net_restart?

Direct wire. Lagometer is obviously changing even ping is 0-8. Usually on LAN you can see no "green" part of lagometer. With "bad" port green particles are shown up.

So even on LAN you can get a "bad" port which increases your ping. I am surprised.

you can reserve port ranges the hard way

https://technet.microsoft.com/en-us/library/c...10%29.aspx

command line is on http://www.riedmann.it/blog/?p=110

suppose to be a front end https://sourceforge.net/projects/trbdk3-netsh/

me and FlashSoul - thank you for comments, I've added a few lines in guide based on your words.

DieZel <3

I skimmed fast through all the info but as an avid net_port users since beginning of standalone I can say this:

Since net port could get me below 70 ping sometimes I can honestly say it had a HUGE effect on me when it gave me a lower ping. This was very noticeable because under 70 and above the ql netcode acts very different (i suspect unlagged kicking in or something)

sometimes it could be very erratic, i remember getting 100 ping with some ports and certain ones could drop me to like 70 and these pings were very real and easy to notice.

Ping means nothink.

more then likely just clears the current snapshot, last sent and received time