Team Radiant has pushed back dates before, but with the exception of infinite worlds they seem to be slowly checking off all the features promised and stretch goals reached in the Kickstarter. My prediction is that they’ll do a little test/trial realise it won’t work and that they have to completely redesign the game engine to fix initial design failures, then decide that it’s going to cost too much, then not bother with multi-player because that’s more fun than covering the cost of an additional 4+ years in development and bankruptcy. It’s just a huge expensive for very little extra profit. Expensive things (like retro-fitting client-side prediction after it’s far too late) aren’t economically viable - the cost of significant changes aren’t going to be recovered by more sales, because most of the people who will ever buy the game have already bought it anyway. Now think about “green lit” alpha games on Steam (where most people that will ever buy a game buy it before it’s finished/released). That’s very expensive (in terms of development time/costs). To do client-side prediction the networking protocol and all communication between client and server need to be designed for it, so to retro-fit it onto an existing game that doesn’t do it (like Stonehearth) you need to redesign virtually everything. The thing is there is no evidence that Stonehearth does prediction at all (and no evidence that there’s any attempt at lag hiding of any kind) and this is very obvious even just playing single player. You can read about this in many places, but this seemed like a good enough (but simplified/basic) introduction: player fighting games), and completely useless for something like Stonehearth where it’s extremely impractical to “roll back” the server state.įor a game like Stonehearth you need something called “client side prediction”. It’s designed for (and useful for) a very different thing (e.g. It’d create the illusion of “lag free” (even now on single player) which currently doesn’t exist. making hearthings walk faster to catch up to where they should be, etc). If the game was designed for multi-player then the client would pretend the hearthings stop moving in the wrong direction (and/or maybe start moving in a “guessed hopefully right” direction) as soon as you issue the “move to location” command (before it sends command to server), and then when it gets updated pathing info from server the client will try to hide the difference between what it displayed/predicted and the new pathing info (e.g. Now you’re looking at 200 ms to > 2.2 seconds of lag before the game responds to your command. When a game (Stonehearth) is not designed for it from the start and isn’t tested under those conditions during development you end up with the “do significant (expensive) changes or forget multiplayer” choice where most developers choose to forget multi-player because “profit 100 ms before the server knows you gave the “move to location” command, it takes twice as long for hearthlings to react (because there’s twice as much CPU load), and then it takes another > 100 ms before server can tell the client a hearthling started moving. It’s entirely possible if the game is designed for it from the start, and if the game is tested under these conditions during development. Nobody sees any latency there, I believe if Team Radiant followed the same model with todays tech they could easily do it. Supreme commander forged alliance (on FAF) is a prime example of having 12 players direct connected to each other with thousands of units and you get a smooth game even when you have 350ms ping to some players. I don’t think it’ll be too bad actually, I play another game online that you would have thought wouldn’t be possible on such scales but it is.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |