Toribash
Original Post
2v2 | 4 player toribash | more than two players
PLay with more people, battle of teams...

Originally Posted by Pikadex
Just a quick idea that one of my friends made: simultaneous battles.
It would involve multiple small squares/circles/dojos as "mats", spread apart, where in each one, two players would go up against one another. If you are to land outside of the area you were placed in (including another mat) you would be eliminated. It could also function as a ladder; the victorious players would move on to the "next round", where it would continue up until there are two left, where they could fight to the death.

Another idea that I made, derived by the extra people battling at once, is team battles. Team battles, as you'd expect, involves 4+ players, equal amounts on each team, pitting against one another. Once all players on a single team are eliminated, or the time runs out, the team with the remaining players (or more points) would be victorious.

One final idea I have is "multi-battles". It involves 3 - 4 players, all against one another. This one could lead up to great chaos, and, other than the extra players, it would function as a normal battle. It could also be used together with the idea of simultaneous battles and/or team battles.


Originally Posted by Blam View Post
Been suggested many times, and was actually possible for a while, it however proved really buggy and slow.

Maybe eventually we'll see it but most not some time soon.

Originally Posted by SkulFuk View Post
^ That...

...or as hampa said on the live steam - it wasn't fun.

Last edited by Lazors; May 6, 2014 at 07:13 AM.
Originally Posted by CombatBot View Post
I think everyone is missing the obvious here.

Version 3.72 of toribash was able to handle 4 players. I tested it myself, no lag, no crashes, no overhauls. I made a friggin replay with it (quite awhile ago, dont have it now) and no problems.

http://forum.toribash.com/showthread.php?t=192853

I didn't have toribash back in 3.72 but I can imagine the physical memory was far less? And the graphics were below, I might download 3.72 and compare how much physical memory both versions use
Originally Posted by CombatBot View Post
I think everyone is missing the obvious here.

Version 3.72 of toribash was able to handle 4 players. I tested it myself, no lag, no crashes, no overhauls. I made a friggin replay with it (quite awhile ago, dont have it now) and no problems.

http://forum.toribash.com/showthread.php?t=192853

As much as I know the problem of it wouldn't be the singleplayer but the multiplayer.
Originally Posted by Zockinator View Post
As much as I know the problem of it wouldn't be the singleplayer but the multiplayer.

The fact its multiplayer shouldn't make much difference in CPU and memory uptake, so if it works in SP it should be fine in MP
you are talking about toribash as if its some kind of resource-intensive program run on supercomputers. the only "data" this game works with over the net is where everyone's limbs are located.. my post has more data than that. the reason why toribash lags sometimes is because you have ray tracing or reflections on.
Originally Posted by HentaiSatan View Post
you are talking about toribash as if its some kind of resource-intensive program run on supercomputers. the only "data" this game works with over the net is where everyone's limbs are located.. my post has more data than that. the reason why toribash lags sometimes is because you have ray tracing or reflections on.

Your post does not contain nearly as much data. There is also the matter of all the textures animations and collisions, it does take a lot of physical memory.
Originally Posted by SmallBowl View Post
Your post does not contain nearly as much data. There is also the matter of all the textures animations and collisions, it does take a lot of physical memory.

textures are being streamed throughout the game until they are loaded, they by no means affect performance. you can play an entire game without the textures loading and only have them appear at the end of the round. collisions and animations dont take any substantial physical memory.. only some processing power

there are 20 joints with 4 settings to them each. a move you send to the server looks like this:

000122130011233221110 -at the worst case this is 80 bytes

this is the only data you send and receive from the server by the end of each turn, along with the positions of each tori at the end of the server calculation. again, my post has way more data than this, unless you still want to argue otherwise...
your post clearly does not have more memory, feel free to explain this then:


The position of all the joints as they move is recorded, but not just at a fixed point all the coordinates of your ghost's joints are recorded, as are the velocities and their status (held etc).

Each of these values (excluding status of joint) takes up 7 s.f. if i remember correctly and there are a LOT of values, and must be using floats, so in a 64x system that is 8 bytes per data value. with the number of data values there are, this multiplies to many, many, many times more than your post, which I assume is using ASCII characters which only take 1 byte per character.

and we have even got to collisions and animations yet.....
The position of all the joints as they move is recorded, but not just at a fixed point all the coordinates of your ghost's joints are recorded, as are the velocities and their status (held etc).

incorrect and irrelevant, they are not all recorded, theyre simulated and dont get transferred between client and server. like i said, the only thing that gets transferred between client and server at the end of each turn is each toris joint status (to the server) and each toris final position and velocity (calculated by the server, transferred to the client). the "simulated" calculations done by the client to show how your tori moves doesnt affect the server and takes little to no physical memory to calculate (you only need variables for position and speed) and doesnt use a whole lot of processing power.... do you understand what i am saying? physical memory plays no role in this.
Each of these values (excluding status of joint) takes up 7 s.f. if i remember correctly and there are a LOT of values, and must be using floats, so in a 64x system that is 8 bytes per data value. with the number of data values there are, this multiplies to many, many, many times more than your post, which I assume is using ASCII characters which only take 1 byte per character.

even if you assume that all the joints have x y z + x y z velocity floats (and not another more efficient way of calculating tori positions), thats still less than 1kb of data to be transferred over FROM the server, client only sends joint status, otherwise this game would be easily hackable. your entire argument is about physical memory which has nothing to do with this... not to mention that my post does in fact contain more data than what you said lol.
and we have even got to collisions and animations yet.....

they are all a product of the same data we mentioned, not a whole new thing.

once again you speak of toribash as its the most complex, resource intensive simulation of moving dummies ever made.


oh, and this..?

the result of ray tracing, shaders and reflections. disable those and your chrome browser will use up more phyiscal memory than toribash....
heres my toribash with no ray tracing or reflections (even looks better imo)


if youre curious to what uses most of the memory in the game, rest assured that its not the calculations from gameplay or the relationship with the server. its mostly the memory used up by the games base library dlls such as SDL, lua and oggs..aka things like graphics, sound, etc.

edit: i hope at least other players drew something from my posts, you specifically seem to be stubborn. your next post illustrates your ignorance and limited understanding of the topic; to even suggest that replays share the mechanics of the actual gameplay is silly. your entire argument could be compared to a someone not wearing clothes out of an irrational fear for his body collapsing under the weight. i dont want to continue replying pointlessly.
Last edited by HentaiSatan; Apr 8, 2015 at 03:42 PM.
Originally Posted by HentaiSatan View Post
incorrect and irrelevant, they are not all recorded, theyre simulated and dont get transferred between client and server. like i said, the only thing that gets transferred between client and server at the end of each turn is each toris joint status (to the server) and each toris final position and velocity (calculated by the server, transferred to the client). the "simulated" calculations done by the client to show how your tori moves doesnt affect the server and takes little to no physical memory to calculate (you only need variables for position and speed) and doesnt use a whole lot of processing power.... do you understand what i am saying? physical memory plays no role in this.

All the positions are recorded, thats how the replays work.
Also physical memory isnt about what is transferred to and from the server its about the RAM being used up in your computer.... So physical memory plays a very large part

even if you assume that all the joints have x y z + x y z velocity floats (and not another more efficient way of calculating tori positions), thats still less than 1kb of data to be transferred over FROM the server, client only sends joint status, otherwise this game would be easily hackable. your entire argument is about physical memory which has nothing to do with this... not to mention that my post does in fact contain more data than what you said lol.

How would it be hackable if more memory was transferred? That makes no sense, also no it does, ASCII characters are 1 byte each. so the post is less that a kilobyte.

they are all a product of the same data we mentioned, not a whole new thing.

once again you speak of toribash as its the most complex, resource intensive simulation of moving dummies ever made.

It may not be the most resource intensive simulation of moving dummies but its still fairly complex and resource intensive...

oh, and this..?

the result of ray tracing, shaders and reflections. disable those and your chrome browser will use up more phyiscal memory than toribash....
heres my toribash with no ray tracing or reflections (even looks better imo)

I do not use ray tracing or reflections...

if youre curious to what uses most of the memory in the game, rest assured that its not the calculations from gameplay or the relationship with the server. its mostly the memory used up by the games base library dlls such as SDL, lua and oggs..aka things like graphics, sound, etc.

All of which would be using more memory if more players were playing at the same time.



EDIT: the source of the memory uptake is not important as more memory would be used with more players however you spin it.
Last edited by SmallBowl; Apr 8, 2015 at 02:14 PM.
Originally Posted by HentaiSatan View Post
you are talking about toribash as if its some kind of resource-intensive program run on supercomputers. the only "data" this game works with over the net is where everyone's limbs are located.. my post has more data than that. the reason why toribash lags sometimes is because you have ray tracing or reflections on.


The fact that toribash's engine was not created to have more than 2 toris at once makes the game extremely unstable during 2v2 matches or whatsoever. It's not about you having a super computer or not, it's the fact that the game crashes so easily with that option that it's not even reliable.
Hampa already said he will NOT change the engine to make 2v2 matches decently playable, and if he ever actually makes a new engine, he'd rather do it to a game for much more than only 4 players....
:^)