IGN N-Podcast 60 and Summary - How Wii doesn't even compare to Current Gen.

[cAPSLOCK]

LordTheNightKnight said:

cAPSLOCK said:

routsounmanman said:

LordTheNightKnight said:

routsounmanman said:
Do you guys know the meaning of ping? It would destroy NSMBWii online. It could be a 8-bit game and still lag due to slow ping.

I try to be up for learning new things. What specifically would be the problem?

Ping is the time required to respond to a call between the client and the server / another client. You can have the fastest connection in the world and still have very low ping, thus, when you browse a page, you see no difference from the person next to you with a 1Mbps connection.

Mario platformers require quick reflexes, thus low ping, response time. The network system is not fast enough for a game like Mario, or other split-second decision games.

I'll expand on this by saying shooters are a completely different category from something like a platformer, since shooters rely heavily on prediction in their coding since the days of DOOM on 56k.  Look at how the characters physically interact in NSMBW, now show me a shooter that has that level of interaction between players. Again shooting doesn't count because it's (invisible) projectiles relying on prediction, which also leads to so many "oh that's a bunch of BS!!!!" moments in online shooters. Compare local FPS gaming (LAN party or 4 player split screen) with online, you'll see a massive difference.

So it's a bunch of extrapolation, not collision?

Also, what about racing and fighting games?

Well, a racing game is easy enough to predict, and if all else fails just go on times. Mario Kart Wii, in spite of what's goin on, is insanely easy to predict. Think about it, a vehicle doesn't move suddenly in 1 direction, it goes forward and turns, there's no sidestepping cars or karts in any racing game.

Now fighting games, I imagine those are pure hell to code online for.

I did find this, though:

Been thinking about the comments on how a deep fighting game like Virtua Fighter couldn't be done online due to the latency issues.

"Well, I'm not a programmer, but I thought about the characteristics a bit versus a traditional 8-16+ player game (say, FPS or racing).

-A fighter would be 1 on 1, and would thus allow a larger packet stream per player (say, 10-15k) versus an FPS which tries to keep each box under 5k or so.
-Generally speaking, barring bad routing, typical p2p broadband anywhere in North America is ~100ms ping time. Could there be more aggressive pushlatency to drive this down to ~30ms?
-Many FPSs rely on hitscan characteristics for gameplay (counterstrike, ghost recon, some unreal weapons), whereas a fighter is based upon effective collision detection. Is this a problem?
-Would this larger theoretical 10-15k/sec network code be useful for driving down effective latencies and creating more advanced prediction?"

and

"I'm sure good netcode can greatly compensate for latency in a fighting game. Unlike a FPS, where the majority of weapons are hitscan, any move in a fighting game takes a short while to pull off. If it takes your character 600 ms to kick someone, and your latency is 100 ms, good prediction should take care of it with no problem. Slower paced fighting games such as OMF (one of the few playable Internet fighting games) are surprisingly tolerant of lag. It's probably not as bad as a Firearms or Day of Defeat, with all hitscan, highly lethal weapons and fast movement speed..."

 

So I guess the question is, what can you predict when playing a platformer? Especially one with 4 people bouncing off each other constantly? Fighting games I think would be a bit of a problem, but I also wonder how good they are online.