Central questions to the viability of traditional MP

[dmcheatw]

1) are the (land) maps larger?
2) are the maps balanced?
3) is there a way to share third party maps? and if not how many maps does the game ship with?
4) are there battlefield objectives or attacker/defender deisgnations?
5) can you see the number of disconnects or unfinished games a player has?
6) can you see pings?


we already know there will be a chat lobby (major plus) but there won't be centralized servers (major minus), but the rest of the questions above i've not heard either way on, and every one of those questions is very fundamental to a satisfying MP experience. if anyone can think of any more major things they want to ask feel free to add them here.

[aletoledo]

Top on my list of questions is if they solved the inability to save a replay if the host disconnects first.

[GrnEyedDvl]

Top on my list of questions is if they solved the inability to save a replay if the host disconnects first.

Uhh I doubt it. After all, all of the in game variables are on the host machine.

[daniu]

Uhh I doubt it. After all, all of the in game variables are on the host machine.

Yes, but they are on the other players' machines as well

[GrnEyedDvl]

Well not exactly. All the calculations are done on the host machine, and its streamed. When you are data streaming the idea is to keep everything as real time as possible. So all the "stuff" on the host machine is converted into simple bits, streamed to the conneting machine and then converted back.

The problem with trying to save something mid stream is that you will have objects that are out of synch. One of the jobs the host does is try to keep everything in synch, so if the host drops, the stream dies and synch cannot be maintained. Especially with stuff that moves in terms of milliseconds anyways, like musket balls.

So if the client tried to save all the information it had, things would not work out properly. The client machine sends information to the host machine, which then processes it and sends it back. Since this stream is constant, information is always flowing both ways. But what happens when the client machine sends data, and the stream is broken before it receives data back, on somthing like a musket ball?

The host thinks its in one place (or maybe not even fired yet), the client thinks its in another. When there is a discrepency like this, the host sets the rules, as long as there is a connection. This is why people see "lag" in games. Because the host and the client are not in synch. When someone "jumps" across the screen in a FPS, its because the host is forcing the client to update that objects position.

But there are always built in timeouts for connection drops. Lets assume its 10-15 seconds for NTW though I have no idea what it is.

How many calculations can occur in 10 seconds when someone drops from the game? The client is sending out tons of packets that need to be processed, and things keep moving on the client screen according to its own calculations.

So which information would you save? Nothing is accurate at all, and there are actually several places where the data is different. Whatever is in local RAM is different from the stream, and the front of the stream is different than the back of the stream which is different from the middle of the stream.

It simply isnt feasible and is an unreasonable request.

[daniu]

Well not exactly. All the calculations are done on the host machine, and its streamed.[...]
It simply isnt feasible and is an unreasonable request.

It's very feasible and not unreasonable at all.
Since it's possible to save if the game finishes without the host quitting, the general possibility exists. Even if the host quits after the statistics are show, the game can be saved on the clients so the data is present there.
Because the "Quit" command is not a disconnect or other unexpected occurance, the host software can react to it by simply sending the information that is still missing on the clients and shut down the game after that.

And even without this mechanism (i.e., on unexpected disconnects), the client could easily save the data up to the last synch point received by the host.

[dmcheatw]

can someone please give just some brief yes/no answers to these questions in the OP? in my judgement they are abosultely essential questions in determining whether or not napoleon multiplayer will be fun in a year...

[LEGIO_Desaix]

By the way in previous TW series games we had the chance to find a temp.replay file in the temp directory even in case of crash of the host. Cannot understand why this file has been removed.

[LiN]

Is the ranking system still the same? Are all the decent players in unranked games again?

[hannibalic]

1. yes
2.not really
3.not sure but i think so
4.no objectives, there is attacker and defender though
5.no
6.no

[dmcheatw]

hannibalic u were with me last night when we assaulted the stupi hill in waterloo. i told you not to push the hill....

but thank you for answering! i agree those answer are right, but there are no attacker/defender designations. i mean when the game loads it says "attacker" and defender" but there are no concrete objectives and the attackers don't get more units, the defenders don't get any defenses. it's just words.

[Lockhard]

LOL, CA said they will not improve all those. Be thankful if they fix some emerging bugs, befor they move team from support to another TW game.