Forum Discussion

DoYaSeeMe's avatar
6 years ago

About netcode

For those that don't know already, Battle(non)sense gives pretty good explanations regarding netcode:

Take your time to read these, as they provide good reasoning for some of the major  issues with Apex.My conclusion is that there are no easy fixes for the network related problems with Apex (disconnections, hit registration, lag compensation, etc.) and most of those that are or will be applied may come with compromises (maybe slower movement, laggy weapons, occasional stuttering, disconnections decrease in quality, etc.).

10 Replies

  • @Fr3dY2 These things are rarely a choice. Compatibility between the engines and systems used for big projects such as a triple A game is a huge issue, devs are usually stuck with what works. Switching to other engines or systems or writing code from scratch can take many years and is not viable.
  • Fr3dY2's avatar
    Fr3dY2
    Rising Ace
    6 years ago

    But does TitanFall 2 suffer the same issues then?

    It hurts seeing Apex at the top of the delay comparison screen 😢

  • Servers getting killed in solos when the matches were filled with Wattsons, Gibraltars or Bangalores are a good indication that there is a huge amount of data to be chewed. Considering this, I can understand why the code net and leaf errors occurrence increased after they announced the increase in tickrate. The amount of data transferred between the client and the server needs to be optimized as much as possible, and we can see some signs that they are trying to find the best compromise. For example, if we look at the death recap, the values probably come from the client, in order to take some strain off the server, sometimes resulting in inconistency with the actual situation (losing a fight although you apparently did more damage and should've won) .This might also explain why grenades and damage are not tracked properly, why sounds are not heard at times.

  • @Fr3dY2 I think Apex is using pretty much the same servers Titanfall 2 used. For Titanfall 2 they were good enough, as there were way less players in a certain match, also less stuff to track, much smaller maps, less sounds, etc..

    The comparison is also not very balanced. Overwatch for example has only 12 people on a map, it doesn't have inventory to pick items, most of the weapons are hitscan and have a low firerate, movement is a bit slower, maps are smaller, there is less detail on the 3d models. A 60-player BR in Overwatch, on a larger map, would probably get much closer to Apex's and Blackout's values.

  • Fr3dY2's avatar
    Fr3dY2
    Rising Ace
    6 years ago

    Let them hire someone at Epic to implement Fornite's netcode then... there are 100 players per match ☹️

  • @Fr3dY2 even if, by magic, the netcode of Fortnite would be perfectly compatible with the rest of the code in Apex, I doubt there would be the same values as you see next to Fortnite in the chart. Also, it would probably cost a fortune just to get to see that.
  • I was looking for some movement speed data and found that  Overwatch has an average of 5.5m/s and Battlefield has sprint at around 6m/s. Apex reaches 7.4 m/s, while sliding can reach a whopping 11.9 m/s (and probably impressive values for stim boost, double timing, grappling and ziplining as well). So, if we consider the same latency compensation and tickrate values for all these 3 games, the hit registration issues would still be significantly more noticeable in Apex. 

    Didn't look at bullet travel times yet, but I believe it's hitscan in Overwatch, which is way faster, and in Battlefield bullets seem faster in the videos iI watched so far. In the end, I think Apex is furtherly disadvantaged by the more realistic model they chose for the projectiles.

    My conclusion is that it's not the netcode to be blamed for most of the big issues with Apex, but the much loved speed of this game.