Why GGPO/Rollback code is important to you. NEW: Current (very strange) XBL network results

i wont even shame my connection by posting it here, all i know if AE on PC is nice and stable with europeans, im hoping for better from SG :wink:

http://www.speedtest.net/result/1871595738.png
http://www.pingtest.net/result/60159741.png
Not wired.

http://www.speedtest.net/result/1871605487.png
http://www.pingtest.net/result/60160074.png
Wired.

I am so confused right now.

I like the why is it important part of this thread.

I like choosing to react to jumpins by doing nothing rather than out of fear my anti air will be delayed.

I’ve never liked Speedtest.net. I always seem to get varied results (from 12-28 DL)

http://www.pingtest.net/result/60214360.png

http://www.pingtest.net/result/60216181.png

Hrm… I think I should be fine. I used to play league a lot and it ran okay. Online should be fine.

http://www.pingtest.net/result/60228259.png

http://www.speedtest.net/result/1873855992.pnghttp://www.pingtest.net/result/60234398.png

I think we need to remind people to divide their ping times by 16.7 (16.7ms = 1 frame). This should at least, give folks an idea of how many frames of delay they need to set when playing against those in their area. Also, if you set the test server to somewhere close to where your opponent is, you can get the amount of delay for a match with them too.

Yeah this is pretty good to go by. I hope the recommendation feature in the game is accurate enough that we may not even have to account for that. Let’s see when the game hits.

What exactly is the point of sharing speedtest/pingtest results? Bandwidth (speedtest) is borderline irrelevant for 2-player fighting games since as long as you have any kind of broadband, you’ve got more than enough bandwidth to handle the game. Testing ping to a random server won’t tell you much at all about what your ping will be to other players, either.

In order to measure the latency of your connection, you should be using something like the Pathping command in Windows (or your operating system’s equivalent) to trace the route from your home network to the person you’re trying to connect to. Example output:


C:\Users\Patrick>pathping -q 25 google.com
Tracing route to google.com [72.14.204.113]
over a maximum of 30 hops:
  0  Patrick-Laptop [192.168.1.101]
  1  192.168.1.1
  2  cust-gw-128-1.troy.al.troycable.net [216.93.128.1]
  3  10.77.0.1
  4  10g-1-1-1-3.atl.troycable.net [64.130.117.163]
  5  ae1-216.atl11.ip4.tinet.net [216.221.158.21]
  6  xe-7-1-0.was14.ip4.tinet.net [89.149.184.42]
  7  72.14.219.89
  8  216.239.46.248
  9  66.249.94.54
10  iad04s01-in-f113.1e100.net [72.14.204.113]
Computing statistics for 62 seconds...
            Source to Here  This Node/Link
Hop  RTT    Lost/Sent = Pct  Lost/Sent = Pct  Address
  0                                          Patrick-Laptop [192.168.1.101]
                                0/  25 =  0%  |
  1    0ms    0/  25 =  0%    0/  25 =  0%  192.168.1.1
                                0/  25 =  0%  |
  2  10ms    0/  25 =  0%    0/  25 =  0%  cust-gw-128-1.troy.al.troycable.net [216.93.128.1]
                                0/  25 =  0%  |
  3  14ms    0/  25 =  0%    0/  25 =  0%  10.77.0.1
                                0/  25 =  0%  |
  4  27ms    1/  25 =  4%    1/  25 =  4%  10g-1-1-1-3.atl.troycable.net [64.130.117.163]
                                0/  25 =  0%  |
  5  32ms    1/  25 =  4%    1/  25 =  4%  ae1-216.atl11.ip4.tinet.net [216.221.158.21]
                                0/  25 =  0%  |
  6  46ms    0/  25 =  0%    0/  25 =  0%  xe-7-1-0.was14.ip4.tinet.net [89.149.184.42]
                                0/  25 =  0%  |
  7  42ms    0/  25 =  0%    0/  25 =  0%  72.14.219.89
                                0/  25 =  0%  |
  8  45ms    0/  25 =  0%    0/  25 =  0%  216.239.46.248
                                0/  25 =  0%  |
  9  ---      25/  25 =100%    25/  25 =100%  66.249.94.54
                                0/  25 =  0%  |
10  43ms    0/  25 =  0%    0/  25 =  0%  iad04s01-in-f113.1e100.net [72.14.204.113]
Trace complete.
C:\Users\Patrick>

Replace “google.com” with the IP address of the person you’re trying to connect to. Have him use something like whatismyip.com to figure it out if he doesn’t know it. You can omit the “-q 25” or change the value, depending on how thorough you want the test to be. This pings every single router on the path between your PC and your destination. If the ping skyrockets between any 2 nodes on the path, then that’s where the problem lies. If it happens between one of the first few nodes, such as somewhere between your PC, your router, and your local ISP, then the problem is likely on your end. If it happens at the very end, then it’s likely on his end. If it happens in the middle, then you’re shit out of luck.

If I’m not mistaken, you should divide your ping by half when it comes to GGPO to get an accurate picture of how long it takes data to get from one end to the other. Ping measures the time needed for a packet to make the trip to the destination and back. GGPO doesn’t require a round trip.

Hmmm…yeah this makes sense since you don’t really need that much upload just to play against somebody online. I doubt getting past a MB in upload that it really matters that much how much upload you have. From there it’s probably just distance and the way the ports are forwarded and other online stuff that I was hoping someone more knowledgeable would post about.

That’s kinda why I added in “if there’s other issues then we’ll have to look further” because I know that even I don’t know enough about networking to know everything that will clog up the game. If you need so little upload that posting up speeds and pings is irrelevant…then that really makes things easier I guess.

I’ll go ahead and put this in the 2nd post and we’ll see how that goes once the game comes out. I think the GGPO delay recommendation should take care of a lot of issues if the upload speed doesn’t really account for that much.

I just wanna play Fightman Game: INJURY MAXIMUM:nunchuck:

It’s the best shit since Shaq Fu.

http://www.pingtest.net/result/60245917.png
http://www.speedtest.net/result/1874166927.png

These are my results. I using a wireless setup for my 360, no I cannot go wired as most people love to blurt out to wireless users.

My experience with 3SO was just awful, and the online portion for SF x TK has been inconsistent. Some games I’ll have minor/acceptable lag, buth other times it’s a teleport fest!!

http://www.pingtest.net/result/60252794.png

http://www.speedtest.net/result/1874392703.png

I just realized my ping isn’t so bad considering the distance is 2.5k miles.

Yeah, speedtest isn’t really too relevant to game performance, as long as both directions are above 0.5-1Mbit/s. Pingtest is a fairly good indicator of gaming performance though (as long as you test a server near you, say below 200 miles). The idea being, that the pingtest server has a good connection, so most of the latency would be your connection or your ISPs network.

What sixfortyfive posted (running traceroute) is a good way to get more exact results though, and also diagnoses where the problem is. But that requires you to know who you’re connecting to.

Yeah I feel if anything at least a ping test should do something basic for giving you an idea. It may not be exact but I can’t imagine it’s terribly off either. A nice way to rough estimate what it would be around. Your ping to someone 2,000 miles away from you is just not going to be as good as someone 200 miles away that’s pretty for sure.

If you wanna know for sure for sure the game will tell you any way. What sixfortyfive seems best at helping diagnose where a problem would be if you’re getting hard lag. Like if you’re playing someone that isn’t that far away from you and you’re still having issues, the pathping would be helpful.

http://www.pingtest.net/result/60365740.png

http://www.pingtest.net/result/60626663.png

Looks like I’m playing offline…

PS. I was using a wireless connection…