[NEWS] Server Performance Survey Summary
Server Performance:
First, we've released a hotfix (v8.2.5) last night to address some issues with our server self-optimizing system. I want to explain why we are spending time on this system.
Why is server performance important?
There is a direct correlation between server performance and players having a good experience. One of the elements of measuring server performance is how many times it calculates everything in a second, this is called the tick rate. Players start to notice inconsistencies when the server has a tick rate less than 20. Very low values will be very noticeable.
What is this self-optimization system?
One of the largest impacts on how a server performs is how many players are running around and shooting. By raising respawn times we can reasonably predict that there will (eventually) be less players on the battlefield at once. The self-optimization system currently increases respawn times if it detects poor server performance. If poor performance continues regardless of the increased respawn times, the server will then force a map vote.
Is there something better than increasing my respawn time?
Yes, the best thing would be to not allow anyone to respawn until the server's performance is acceptable. Keep track of how many players can be alive before the server struggles and then force that to always be present. Obviously we don't want to take this approach. There are also other ideas we are considering, but I won't go into that.
Why not just whitelist servers that can perform well for its slots?
That would be the modern approach. Squad™ and many other titles are doing just that. We have a very small community, very few servers, and there are also downsides to the server whitelist system. It disrupts freedoms of server owners, it has to be updated, and doesn't attempt to improve performance if poor performance is actually detected.
Why not just get a better server?
Well there are some issues with that. Which is money and usage. There is no point in getting a better server if by chance another server was instead seeded and is populated. Suddenly, you realize are paying a lot of money and no one is using it, this has been a major problem. There is no incentive for server owners to offer a better performing server, when players just join whatever server has players. I'd imagine this is why the server white list system is useful for more populated games.
However, we plan on getting a high performance server based in London soon. This server should not have any tick rate issues, even with 64 players on Bridgehead! When that happens please help pay for it by becoming a patron here: https://www.patreon.com/theel
We will continue as needed to create a system that gives incentive to players to play on the best performing servers.
Survey Summary:
Overall, how do you feel about the WIP maps?The consensus shows that players support the WIP maps. Players see them as a good thing for DH and are willing to tolerate their downsides. I cannot find one player who said they hate WIP maps. I think this means we made the right choice and I'm glad we finally made it. You guys can thank Kashash "ksh" for pushing for the idea. I think the next step is to create a way for community levelers to easily submit updates to WIP levels. But I also want to work it out, so they know which WIP levels are being worked on by DLG (us developers).
How do you feel about Rhine_River_Clash? How could it be improved to be more fun? What do you feel are its biggest issues?
The consensus shows that its still very early to make any big claims about the level (a lot haven't even played it yet). But those who have, said it needs more cover, more forests, it's too large & open for infantry, river is too crossable, requires a full server, and the towns leave much to be desired.
Rank your most desired levels to be updated:
In first place, the most desired WIP level to be updated is "DH-WIP_San_Valentino_Advance"! I found this surprising because I honestly haven't a clue where I'd start. In second place, we have "DH-WIP_Maupertus_Push" which I'm not surprised about, because I've been asked to update Maupertus many times in the past. Other notable mentions include: Black_Day_July, Arad_Advance, Makhnovo_Advance, and Arnhem_Bridge
What should we do about levels that obviously harm the game because players leave consistently?
The consensus is to NOT remove them from the game, but let servers take harmful levels off rotation until they are updated/fixed. And I agree with that.
What level sizes should we focus on?
65.35% voted to focus on large levels
28.85% voted to focus on small-medium levels
5.77% voted to focus on tiny levels
Well I'm not at all surprised by this, but truthfully it might seem like we don't focus large levels. The reason being is large levels take a lot more time and effort to update, overhaul, and create.
Would you be interested in "fight yard" maps?
This turned out to be roughly 50% yes and 50% no. I think this means we need to focus our efforts on all levels and sizes and try to balance that the best we can.
Which do you prefer for a low-medium population (10-24 players) server?
60.78% said Push
19.61% said Clash
19.62% said Advance
This makes total since because the majority of Clash and Advance maps are large, too large for low population. But I think that is their nature, we can experiment with a "Squad vs Squad" smaller map, but I think in the end Push will come out best for low population.
Which do you prefer for a high population (50+ players) server?
76% said Advance
12% said Push
12% said Clash
Clash is newer and a lot of players consider many Clash maps to be "Advance" still as they are similar. For instance, Hattert, Cambes, and Nuenen are Clash maps. Therefore, I think the question was not well worded. We should probably include the Clash votes with Advance, which gives a large majority to Advance/Clash for high population. That said, we are noticing some annoyances with Clash and will be making a significant change for Clash. For more information: https://github.com/DarklightGames/DarkestHour/issues/973
What is your least favorite level?
The most voted least favorite level goes to... Kasserine_Pass this is not surprising. Other notable mentions are: DogGreen, Gorlitz, Dom-Pavlova, and Donner.
I'm surprised these levels were only mentioned once or not at all:
Simonskall (no one hardly votes for this level, so I just assume no one likes it)
BridgeTDM (my personally most hated level, had 1 vote which was mine)
Kommerscheidt (sadly this level has just never played well, had 1 vote)
I have yet to make my own Normandy Beech map, which I think one day will happen, just a matter of time.
Summary:
The purpose of this survey was to gather how players feel regarding WIP maps, what maps they want to see improved the most, and where to focus efforts. I know that players love or at the least accept WIP levels. We know which maps players want to see improved. But, where to focus our efforts was not really answered. I think realistically we just need to keep moving forward with a reasonable regard for low and high population play. Thanks to everyone who participated in the survey.
See you on the battlefield,
Theel