devtrackers.gg
Sorry, but, are you talking about the 300FPS cap or the known "stutter" bug above ~180FPS?
stutter
Sorry, but, are you talking about the 300FPS cap or the known "stutter" bug above ~180FPS?
stutter
You don't actually want to do this. The game engine has a bug in it which causes a microstutter above about 180 fps.
Either way, it's hard capped at 300 FPS and you can't go higher.
Should be gone after next patch. Do not hesitate to let me know.
does level disparity even mean anything? I've played with some really bad high level players. high level just means they play a lot, not that they're good.
you might be with a level 50, but it could be a great players smurf account...it works both ways
You are right, I am speaking about skill when I say "level".
Overall the fairness of a game is pretty even
How do you measure this? I have only anecdotal evidence but in my experience games were only fair when I played in a 3 stack where the game started to match us with other diamond 3 stacks. And that's only like half the time, the other half are either a stomp from our side or we get stomped.
Solo queue is always, without a single exception either me carrying the whole game or we get absolutely stomped. There is no inbetween.
You are right, it is extremely difficult from anecdotal data to draw conclusion but it helps us to identify and verify those intuitions. We measure every games. What are the skill of people involved, the expected outcome of the match,how long you waited to get into a game, what was the team composition etc.
So in your case, I may look to your trajectory as a player (to try to put what you are expressing into metrics I can analyze). Then we would look at your "stomp", so certainly your 3-0, 0-3 matches. Does it really represent 50% of your game played, trying to see why it happened (what were the level of your team mates / opponent, premade / not premade ), etc. Then I would try to generalize your player profile if I can see trends in the overall population of people playing Arenas.
And it is how we can iterate and make it better, so thank you for your feedback!
Yeah, same here. Plus the two low level players usually have the same [CLUB] tag.
But I do believe this is what Respawn wants with their match making, so little you can do unless you get as good as people like Hakis who can just stomp game after game.
Not really. I am working on it.
Overall the fairness of a game is pretty even but we understand it is a bit tiring when level disparity is high in your team composition. It will improve.
One thing I've noticed is as a solo queue player, I almost always get queued with a duo. My guess is that's intentional, although it may have an unintended effect of producing a team withlittle to no comms, as the duo is communicating in discord. Three or four seasons ago when I last solo queued my way through plat, seems like I was teamed up with more solos with more comms and a generally less toxic environment. Wasn't a horrible experience solo queuing last season, but it was a step or two down from the better experience a few seasons back.
I love those points. It is an interesting theory. I will check what's what. As for the truth of the algorithm behind it, it really depends on what rank you play, the higher you get, people tend to party up, so the less solo there is.
How would eliminating the network aspect of the firing range put "a huge strain on the machine" ??
What he said + memory.
Great writing!
About the second part where you name me directly about server stuff.
It is definitely an interesting discussion. What I tried to convey in the article is that using tick rate as a metric is misleading. The notion covers a lot of different concepts with different meaning depending on the tech stack. I am not going to comment on other games because every tech is different and have different design goals (I believe the technical part is an instrument to empower designers).
What I can say is that there is no free lunch. We have 60 players that can land at the same spot of a map and fight with high fidelity engagement. Shall we sacrifice a frame or two for shoulder picking to support this scenario ? So be it. From a design perspective it seems more interesting in our context. We still support high fps clients and we rely on our lag compensation to ensure fairness. I cannot throw money to grow bandwidth on people's connection so this argument seem...
Read moreWith the introduction of streaming services like Stadia, xCloud etc, can you see this having an effect on how online-centric games like Apex work, and are designed, in the future?
Would imagine it opens the door for games to be designed with a vastly bigger scope in mind, be it more players or more advanced mechanics, as it effectively removes the client from the equation which formed a lot of discussion in today's blog post. Keen to hear your thoughts on that from a tech standpoint.
Second, not a question, but thank you for all the hard work you and the rest of the team put into the game, and for the blog post today. I love the game, and I can really tell how passionate you all are about it which is great to see. Think it's really cool you're willing to talk at length about the behind-the-scenes systems and the decision making around them. I bet there were some awkward conversations in the office about posting it!
I think the key point in your question is "are designed". I believe that if publishers want to make those experience compelling, a game needs to be designed from the ground up to account for the platform constraints. I am particularly thinking about input delay. For example, aim assist is the consequence of consoles offering a gamepad as interface with the player. You need to run the same train of thought when you design your player / game interactions.
Thank you for the nice words! We are happy to share.
I think they were asking about which legends you main lol
haha sorry, I mainly play Banga / Wattson.
I had 1 game with around 30 package loss last week, every enemy I had did not even shot me, they were standing still for 1 second every 5 seconds and my hits to them would count as normal hits. So even if the enemy was standing still (which was clearly because of my connection not because of the movement of him) my dmg was being registered. I can only imagine that I teleported and lasered like crazy for the others...
It felt like a big advantage but it only occurred once and could very well be an on-time thing, but I wanted to tell you about it.
Thank you Lex for the feedback. My gut feeling would be that everyone was experiencing massive packet loss due to an infrastructure problem. Do not feel too bad about it.
Off-topic:
- What's a Lead Engineer's Mains in Apex?
Lead means I have a team. Everything I am speaking about is the result of a collaboration between great individuals!
Is the Battle(non)sensevideo still relevant to this day https://www.youtube.com/watch?v=9PfFPW9a90w
I think most points raised in the video are explained in the blog post:
Some high numbers are also explained simply by the fact that we use an authoritative-server model. For example for damage, we wait for the server to return the exact number to give you an accurate representation of your opponent health. This makes tracking our server response speed based on the numbers you see a little misleading (which seems to be how Battle(non)sense is doing it.
The downside of our approach is that you have to wait for the server response after shooting someone to see the numbers, which inflates the time to display it on screen. But the upside is that the numbers are accurate, and you can be certain that your opponent lost life, and you can base your strategy on that information.
You can see the hit registration issues this way too, and at least it is factual data. Differe...
Read moreWith SIDs, does that also mean it's possible to transfer a game instance to a different server on the same rack? For example, if a lobby is slow-mo and its detected, could there be a system put in place to transfer the game to server 3 if server 7 is lagging? Are there any roadblocks that would prevent that, or can the lobby just be paused for the time being while the game instance loads on a different server and redirecting player traffic to that server? Could be a really cool way to prevent slow-mo lobbies from happening entirely because they're extremely frustrating to get in ranked and just lose 60RP for something that wasn't your fault. Really happy about all the interesting new tech going into helping the server side!
It is hard to move around game servers because they consume a lot of memory. Usually, slow-mo is due to server not running up to spec. We hope, even if some suffer a loss because of it, that after a game those servers will not be re-assigned to new players.
As part of the blog post you mention about balancing the experience for both the "low" and the "high" ping person. But one situation I am wondering about if Respawn will do anything about are people with ping so bad that all other players in the match literally see them teleport and cannot hit them as a result as they are teleporting from location to location across your screen. Is Respawn looking into this and going to resolve this issue?
Player teleporting have definitely bad latency but also certainly packet loss. When you have dramatic packet loss, it will mechanically also raise your latency over a threshold that the game cannot compensate. While I do not believe it can be exploited to the high latency player advantage, I agree that the game feel / look broken. We will look into it, maybe our current solution is too defensive.
Probably a stupid question, but have you playtested the mode personally?
Curious to hear your impressions on it compared to BR.
I'm excited as hell for it ever since the leaks. 😁
Worked a lot on the mod. It is very hard to give you an objective opinion of your own baby (of course I love it). I hope you will enjoy it too :)
All I can say is.... thanks. But I still have my comments:
Regarding ticks: would you mind explaining why we cannot use Battlefield series (which is also published by EA) or even other games that have high number of players and claim to have high tickrate servers? Is it due to the lower TTK means they don't have to process as many information overall?
Regarding the walls-hide situation: Wouldn't it be okay to just go with the flow and award Low-Ping Players? I mean it was the way since Quakeworld (yes, I am old), where having better IRL equipment will give you advantages. (And in fact you acknowledge quite a bit of games did that)
Addressing the first part of your question: At the end of the day, our job is to fulfill the designer's vision on how the game feel given a set of constraints. Apex supports 60 players dropping at the same spot while guaranteeing bandwidth / latency. You can see that a lot if you watch our competitive scene for example. A lot of teams play the placement and get to the final ring. Battlefield works a little bit differently. Without getting into their netcode (not my game!), each game makes different choices and accepts different trade-offs.
Since Apex legends has just reached 100 million players. Where are the servers for Africa?? We've been playing the game with 190ms.
To open a new datacenter we need to be sure that the number of player will be healthy enough to support matchmaking. Sadly for the moment it is not the case. For example, our lowest population datacenter is in Australia. South Africa account barely for 1/3rd of that.
Arena matches will probably be better connection-wise just by virtue of having fewer players. It's a lot easier to match 6 players of similar skill with a good connection to a common server than to do the same for 60.
This is completely correct! It's not that we'd never consider it, but we think there are higher-priority changes we could do that would help both BR players and Arenas players. I'm very excited to see people's feedback on the new mode!
In the article you mention that getting to the root of packet loss/big variations in ping, is difficult and requires users to run "network traces". Is this something that is available for all players or just the pros and content creators?
If its available to everyone, how would one go about reporting that they have a netcode issue and possibly have respawn track down the cause of the problem?
It is for everybody. We have automated the packet loss collection from the server. We will be able to react faster on the data center side. The issue remains between you and your ISP. We cannot do much on that front sadly.