Highly discussed topic, nice to hear that Valve had noticed it. So in
the future IP changes are not longer the death sentence for servers...
Thank you very much for this response. It is quite helpful to know that
Valve is contemplating our suggestions.
On Sun, Jul 10, 2011 at 5:35 PM, Fletcher Dunn
<fletch...@valvesoftware.com>wrote:
If you want to know if matchmaking is working or not, looking at your own
server will tell you very little. You have to play the game. Our current
goals for quickplay are pretty limited: to match players (especially new
players, who we expect to be the ones using this feature) to the best server
for them. In other words, we define success solely from the client's
perspective. We don't have any criteria that attempts to optimize results
from servers' perspective, with the exception of reducing unwanted traffic
(in the form of query pings or quickplay player traffic in general).
Last weekend (July 2) all of the Valve servers went down. This undoubtedly
sent a lot of quickplay traffic your way. Then, on Monday, they came back
up. This almost certainly resulted in less quickplay traffic coming your
way. Our overall player counts are reduced since the initial surge, but we
still have a pretty large number of Valve servers up. I'd guess that ratio
of Valve servers versus players using quickplay is the single largest factor
in determining how much quickplay traffic comes your way.
On the other hand, last week's update reduced the scoring advantage that
Valve servers get. We are still tweaking this. We may make the scoring
advantage a function of player experience, to get totally new players on
Valve servers, but over time transition them into the general population.
Payload is the default game mode choice. You might get more quickplay
traffic on payload maps.
We are experimenting with certain rules changes being score penalties
rather than deal-breakers. (nocrits is the first test.) But right now,
there are so many good Valve servers available, unless your server is
vanilla with a good ping and 15-20 players in it, most clients will score
the Valve server higher anyway, so pretty much any scoring penalty can
effectively rule you out. This is just a function of the number of players
and the number of (Valve) servers.
The fix for the replay off-by-one player count bug should be in the next
update. We have received some helpful reports about zombie players and
hopefully will have that fixed, too. We also read the suggestions for how
our codebase could be improved and features implemented more simply. Rest
assured that those suggestions are being given all the attention they
deserve.
The reputation scoring system will take some time to tune. Currently it
can be dwarfed by the other factors. We are not certain it is the best long
term solution, perhaps creating a more persistent server identity at the
Steam level is better. For example, so you can change your IP and not lose
all of your loyal players who added your server to their favorites list. We
have heard you loud and clear on this point and it is something I personally
am very interested in working on. Having a less anonymous system will
benefit good servers and players and is a good thing all around.
Your humble servant.
- Fletch
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
