Simple changes to improve WF

[Shadowh4nd]

• Performance impact indication in display settings, or sort these into tiers from nearly cosmetic (film grain) to huge (ambient occlusion)
• Messages not sent are cleared after 1 minute, seriously. I have nearly sent a hundred messages written over top something I wrote earlier. Real awkward when I'm copy-pasting to a seller and there's any ol' sentence there from 30 minutes ago, meaning I now say it in whatever channel I'm in. 
• An option for players to disable being selected as host
• This functionality built in when player ping does not meet certain requirements (say > 250ms).
• Blocking also stops a player from being matched with you
• Vote kick (I know this has been said 1mil times)

[WhiteMarker]

8 minutes ago, Shadowh4nd said:

• This functionality built in when player ping does not meet certain requirements (say > 250ms)

Explain how it would work.
Ping to DE's server? Or Ping to you, if you are the client? How would DE measure the ping you have to another player? Don't think that would be legal.
What use is there for the ping to DE's servers if the ping to you is S#&$?

[Syasob]

56 minutes ago, Shadowh4nd said:

• Performance impact indication in display settings, or sort these into tiers from nearly cosmetic (film grain) to huge (ambient occlusion)
• Messages not sent are cleared after 1 minute, seriously. I have nearly sent a hundred messages written over top something I wrote earlier. Real awkward when I'm copy-pasting to a seller and there's any ol' sentence there from 30 minutes ago, meaning I now say it in whatever channel I'm in. 
• An option for players to disable being selected as host
• This functionality built in when player ping does not meet certain requirements (say > 250ms).
• Blocking also stops a player from being matched with you
• Vote kick (I know this has been said 1mil times)

1-2. Agreed. Settings and chat should get some attention of UI team. They're both in the queue I suppose.
3. What if you're the only one willing to do the mission ATM? Or if every player queued for it selected not to be the host?
4. There's an option to set a ping limit already. Also WhiteMarker already stated that ping between players can't be measured before the start of a mission.
5. It already blocks you unless there're less then 4 players in public queue for the mission.
6. This has been said 'no' more then 1mil times. Go search the forum for more info.

[Steel_Rook]

2 hours ago, Syasob said:

Also WhiteMarker already stated that ping between players can't be measured before the start of a mission.

Yes it can. Why do people keep making that claim? You can ping any IP at any time to get a ping measurement. Simply have the player's client ping everyone in the matchmaking queue and either sort them by ping or disqualify those with ping over the limit. Games have been doing this since the 90s.

[WhiteMarker]

8 minutes ago, Steel_Rook said:

Yes it can. Why do people keep making that claim? You can ping any IP at any time to get a ping measurement. Simply have the player's client ping everyone in the matchmaking queue and either sort them by ping or disqualify those with ping over the limit. Games have been doing this since the 90s.

Fair. Didn't know this.

But what then? A simple ping doesn't tell you how the connection and the system of a player handles a full session. It doesn't tell you if the chosen host will have a constant ping that good, or if it will go to S#&$.

[Steel_Rook]

6 minutes ago, WhiteMarker said:

But what then? A simple ping doesn't tell you how the connection and the system of a player handles a full session. It doesn't tell you if the chosen host will have a constant ping that good, or if it will go to S#&$.

That's true. A basic ping doesn't give you a measurement of packet loss or connection capacity, both of which do matter. However, it's also cheap, fast and easy to do on a large scale while offering a pretty strong indication of poor connection. You can alternately do what traceroute software does and ping players multiple times for a period of time and grab a minimum, maximum and average ping value. This gives you some idea of the other player's connection. Ping with a lot of "jitter" is typically indicative of significant packet loss, unstable connection or other network issues.

Warframe already does some of this behind the scenes, by the way, when it pops up the "Poor connection quality, are you sure?" message when joining a team. Which, incidentally, is why I believe the game already has the feature the OP's asking for.

[Shadowh4nd]

14 hours ago, Steel_Rook said:

That's true. A basic ping doesn't give you a measurement of packet loss or connection capacity, both of which do matter. However, it's also cheap, fast and easy to do on a large scale while offering a pretty strong indication of poor connection.

I agree, it's a quick and easy way to tell, and not 100% accurate. I see this problem after getting into laggy games every so often with my matchmaking ping limit set low. Feels almost like it grabs one ping and runs with it, even if that's a low spike.

How to HANDLE a host that suddenly goes 250ms+ mid game, well, that'd be a separate conversation because host migration is already in a bad spot.

18 hours ago, WhiteMarker said:

Explain how it would work.

Ping server (x) times over the course of (y) minutes. Average these numbers. Creates an internal hosting score. Sort these players as preferred hosts.

Repeat every (z) minutes in order to provide a current and accurate measurement

17 hours ago, Syasob said:

3. What if you're the only one willing to do the mission ATM? Or if every player queued for it selected not to be the host?
4. There's an option to set a ping limit already. Also WhiteMarker already stated that ping between players can't be measured before the start of a mission.

3. If you have selected "not host" in your settings and, in a totally hypothetical world, the entire queue also had hosting disabled, it'd be no different than setting your matchmaking latency down to 100 and everyone else being 200+: Wouldn't match with anyone

4. Doesn't do well and isn't from the perspective that I'd look from: Client to DE, and rating these over time. I am aware of the option to limit player to player within settings, and I think it does a poor job. I am asking for a change to this system to exclude people entirely from being selected as host while their ping is too poor.

[WhiteMarker]

1 hour ago, Shadowh4nd said:

Ping server (x) times over the course of (y) minutes. Average these numbers. Creates an internal hosting score. Sort these players as preferred hosts.

Repeat every (z) minutes in order to provide a current and accurate measurement

This is exactly why I asked.
Why would the ping to the server tell you anything.
A player from London, Ontario will have a stellar ping, because he lives right next to DE's servers. Doesn't mean he will be a good host for people from Melbourne, Australia.

1 hour ago, Shadowh4nd said:

I see this problem after getting into laggy games every so often with my matchmaking ping limit set low. Feels almost like it grabs one ping and runs with it, even if that's a low spike.

This happens because the game does what you said.
DE's server pings the player. This doesn't give any information about the ability to host...

[Steel_Rook]

5 hours ago, WhiteMarker said:

This is exactly why I asked.
Why would the ping to the server tell you anything.
A player from London, Ontario will have a stellar ping, because he lives right next to DE's servers. Doesn't mean he will be a good host for people from Melbourne, Australia.

Pings are done directly between players without involving DE's servers. The whole game's team infrastructure is peer-to-peer, meaning you have access to the IPs of everyone you're teamed with because you're connecting to their machines directly. Pinging them at that point is trivial. It's what multiplayer games have been doing since the 90s. Hell, it's what VOIP software does even now. You know - "Voice Over IP." Basically, Warframe's teaming works a lot like old LAN Party games, where one person hosts the game and everyone joins them.

 

6 hours ago, Shadowh4nd said:

3. If you have selected "not host" in your settings and, in a totally hypothetical world, the entire queue also had hosting disabled, it'd be no different than setting your matchmaking latency down to 100 and everyone else being 200+: Wouldn't match with anyone

You're running into the risk of really long queues that way, though. This is something most games with "intelligent" matchmaking run into, or at least the excuse their developers use. Overwatch developers have talked about how even minor restrictions to their matchmaker can bloat queue times to 20+ minutes in some cases. I believe Jeff Kaplan said this in regards to a "role queue," where DPS players were likely to face untennable wait times. World of Tanks has ha similar issues, although in their case there's also underhanded RMT bullS#&$ involved. Their attempt to limit the number of people screwed by the matchmaker by using a 7-5-3 system (7 low-tier, 5 mid-tier, 3 high-tier players) did that, but the result was matching players as low-tier almost all of the time, which is frustrating. Then there were machines with "special matchmaking" tiers, which ended up sucking players from the broader matchmaking and creating dead spots I think around tier 9.

The long and short of it is that smarter matchmaking systems almost always mean longer wait times. When it comes to what you proposed up there, I'd personally let the game keep you waiting for a bit to see if people show up, then matching you as the host anyway, despite your selection. Or at the very least giving you a warning that no match can be found with your current settings, would you like to search for "any." Right now, Warframe matchmaking is pretty fast because it's almost completely unregulated. The moment you start regulating it, wait times are going to start becoming an issue.