[SUGGESTION] Recent master branch update concerning spawning

It has come to my attention recently there was a change made to some servers concerning a certain cvar.

g_spawn_furthest 1 is now activated. Meaning that the enemy will always spawn at the furthest point from your position. In the duel servers that I usually play in (SMB), its set to 50% of the time spawns are the furthest and the other 50% are random. This has now changed without any real talk or discussion about the command.

Quake live has an interesting system (not saying we should copy them at all):

1.)The system consider the opponent 3D position inside the map.
2.)Computes the distance from that position to all spawn points
3.)Picks the farthest 50%, with a minimum of four.
4.)Picks one of these randomly.

So it eliminates the closest spawn points to you then selects at least 4 of the points furthest away from your position. Out of those four, one is randomly selected for you to spawn considering your opponent position on the map.



I was wondering where the community actually stands not only on this change, but positional spawning as well. I couldn't find a good thread on the topic other then some talk about spawn kills. I realize in some modes this matters more than others. This post is mostly concerning vanilla duel.

What do you think the standard for Xonotic should be in terms of how and where we spawn on the map? Do you like the current change? Is always furthest okay or do you like the aspect of randomness in there as well?

More predictable the system = easier the spawn killing, quite simple as that. I like spawn killing, but not easy or predictable spawn killing.

There's always some luck involved in every spawn, some players start better from certain spawns and some not, some players straight out get the situationally advantageous spawn. Sometimes the previous down player gets a worse spawn after a mutual frag, which is pretty much a "fuck you" to his face after going all the trouble of getting your opponent out of stack and so on.. All the unlucky stuff that can happen, I think you know what I mean. So if this spawn system is a crusade of luck factors, it can only help so much.

When the system is random, it balances the luck factor for both players. It forces players to be aware at all time, of the worst things that can happen and oppositely the best things can happen. The player in control can still get spawn frags the better he is aware of different routes the players tend to take from different spawns and it sometimes puts players into awkward situations that require improvisation and experience to make the best of it. I've personally enjoyed this system (in Reversion cup servers) more than the predictable system where you always end up with a linear and a repetitive spree of conversion frags. And it's not even an advantage of map knowledge, rather spawn system knowledge. I don't see as much value in that.

If the general problem people have with spawns is spawning next to your opponent. Then making it not random makes only a little difference in larger maps in a long run anyway and at the same time it notably narrows down the progress of matches in smaller maps. So I think the random system would collectively make the least difference considering the fact that maps larger than Hubwalken are in the majority and it would at the same time keep the action in namely Hubwalken, more interesting.

I do enjoy spawn killing and I also believe it should definitely be a staple part of the game. Obviously prediction is a big part of duels such as, knowing spawn points, listening to pickup sounds, timing them, or shooting around a corner and calculating where the opponent might be. I don't see the fun or the skill or even the entertainment value of basically influencing pretty much where they spawn. I think having a system that allows for variety and some times uncertainty makes you be on your toes and allows for things that wouldn't normally happen. g_spawn_furthest 0.5 allows for spawn fragging to still be a part of the game, but not enough where you can get like 5 frags in 15 seconds because your able to manipulate a certain part of the map. I know Mirio changed the spawn points on stormkeep for this reason, but it can be very easily abused in Hub and Toxicity. Those are the only two maps off the top of my head but obviously with this new Duel Mapping Contest I think now is the time to solidify once and for all on what we want done about it.

I also can't say that that is my end all be all about spawning. I'm certainly open to other opinions on the subjects as I have yet really to play a lot with this setting.

g_spawn_furthest has been reset to 0.5 in Git master and on the SMB duel servers.

Random spawns: http://stats.xonotic.org/game/419977
Spawn furthest: http://stats.xonotic.org/game/456918

Spawn furthest: http://stats.xonotic.org/game/456918

0.5 though!

Smilecythe, you didn't just rape a person as part of a inhumane experiment to test balance???

I think that system should be made such way to avoid repetitive spawn killing of one player by another. I cant forget the blood-boiling experience of spawning 3 (!!!) times in a row in the same spawn, each time being identically spawn killed by the opponent.
I think that there should be a mechanism that would prevent a player from spawning in the same spawn as in the previous time, to both prevent getting repeatedly pwned or gaining an advantageous position again.

Random spawns: http://stats.xonotic.org/game/419977
Spawn furthest: http://stats.xonotic.org/game/456918

Holy shit. 55-5. I dont think I can even kill a bot that many times in 10 minutes.

That match had absolutely nothing to do with spawnfurthest, it was the basic 0.5 with 2 duellers with absolutely gigantic skill difference. But yes good change, 1.0 was a joke, there was nothing you could do on some maps on duel modes. I'd still vote for completely random with a twist, you cant spawn to a spawnpoint closest to your opponent. I think its like that on UT. Maps are divided to areas, one map has areas like "Basement, Ramp, Top, Rocket". and if i'am in Basement, opponent will not spawn there, but can spawn anywhere else randomly. I have found that as the best spawning system yet to date and i find it really good for TDM gameplay.

So ye, i guess my vote goes for "Other", then

---

And ye, its very strange if we have to adjust maps to fit to a predictable and strange spawnsystem (talking about spawnfurthest1.0). Maps like stormkeep or aerowalk. Blame the maps, not the spawnsystem, quite strange eh, isnt it you mappers? Its like communism, on paper its like a dream but it just fucks everything up in practice.

---

Ah ffs, still one thing to add for that spawn system ... on UT you also cannot spawn 2 times on a row to same spot. So otherwise completely random, but not the closest spawnpoint to your opponent and not 2times to the same spot! Its hard to write this shit in english for me lols.

Ok im finished.

Quakelive spawn system is awesome and there is no shame in copying it.

0.5 though!

That match had absolutely nothing to do with spawnfurthest, it was the basic 0.5 with 2 duellers with absolutely gigantic skill difference.

50% spawn furthest is still 50% spawn furthest

@Thimo: Without any catalysts like that our games used to be so much better on any maps alike, don't you agree?

Fuck the furthest and biased furthest crap, random + not twice in same place would be master race.

Doesn't avoiding same-place spawn decrease unpredictability? Just for the sake of discussion, say, a small map has 2 spawn points; we can then 100% predict where the next spawn point is (i.e. the other one).

As for furthest points, they sound predictable. With some many maps out there, is it possible to have teleports make furthest points actually close? Obviously some maps are better than the other, but do we want Xonotic's spawning algorithm rule out some of the popular maps?

Avoiding same spawn gives some predictable information, but that information becomes less significant the more spawns the map has. It's certainly less predictable than a 50% chance to spawn to the opposite side of the map.

It's a good point, however I can't think of any genuinely good 2-spawn maps or people so illogical who would first of all actually enjoy them (competitively or casually) and then be convinced that the map would play better with a different spawn system. Silly maps like that which people create for experiment or for nonsensical fun should be anything but a concern in front of gameplay decisions like this.

And yes, there are maps with teleporters and warpzones that shortcut people to the "furthest" spawns and that is one of the main problems people have with this system, it renders way too many maps exploitable in a way or another.

The question is whether we actually consider "spawn prediction" a skill that adds to the depth of the game. For me personally spawn killing is not a particularly fun aspect of the game, but I may be biased since don't have that skill in the first place (also it felt too lame to me to attempt to acquire it).

So I would appreciate moving away from a spawn system that effectively increases the probability to spawn at a specific spawn point. I like the idea to avoid spawning too close or in the same spot twice in a row but otherwise randomly. By keeping the pool of eligible spawnpoints to choose from randomly as big as possible we can effectively prevent spawn prediction. Excluding the "worst" spawn points while leaving the rest at random turns out better for the fragged player since the benefit of spawning at the "best" spawn point is more than outleveled by the experienced player's prediction of you spawning there.

While thinking a bit more I wondered whether it would be possible to introduce a flexible spawn system like Overkill mod has where you can actually spawn anywhere in the map. Then we could define areas in maps where you can't spawn (such as in lava or in open space) and have the spawn position vary completely randomly, maybe influenced by regions only. I'm not sure this is a good idea for competitive gameplay though, since it might be harder for people to immediately recognize their position in the map upon spawn if you don't stick to fixed spawn points.

Overkill spawns players near their team mates, not randomly anywhere.
Random spawns would open up the possibility of spawning players inside walls that aren't fully solid (e.g. the floor of dance, which caused some grief back when I was experimenting with a zombie apocalypse mutator).

For public matches, "spawn zones" could work very well though, worth a try if someone is willing to implement it (tZork made something like that for his SNAFU gamemode if I recall correctly).

When referencing Overkill spawn system I meant to refer to the fact that the spawn is independent from a dedicated spawn point. I know it puts you near your team mates :o)

Edit: OMG this is my 2000th post! Thank you all for your endurance (or ignorance, as the case may be)!

I see people talking about not wanting the same spawn point twice in a row. I was actually working on a post about that the other day but got distracted! :O

Most of the spawn modes in Quakeworld are like that, and I've never had a problem with spawns in that game. Most of my games were on either KTX or KTX2, not sure which. http://wiki.quakeworld.nu/Spawn_modes

I find KTX2 especially interesting:

"First a random spawn point is chosen. If this spawn point is the same spawn point where the player spawned last time, a new spawn point is chosen again and the player is spawned there, even if it's the same spawn spot again.

The probability of spawning at the same spawn spot is 1/(N*N), where N is the number of spawn points on the current map."

This just makes it highly unlikely that you will spawn in the same spot twice, but still leave it open as a possibility.



I think g_spawn_furthest 0.5 strikes a decent balance between the predictable doldrums of furthest spawns and the chaos of random spawns, though I would prefer it completely random, or with the non-consecutive stipulation. I see it as a personal preference more than anything. I like that such an option is available in-game, and that tournament organizers will pick whichever they think is most suitable.

How about approach this from a different perspective? Spawn kills frustrate us because the victim can't do anything about it. How about an option to allow the player to move, but not to shoot (or it doesn't cause damage), nor to take any damage, for a period of time, say, 0.5 seconds? This way, the victim has a choice to try to escape from spawn kills. We can think of the spawn as the bodily substance not solidified yet so it's physically impossible to be touched (although we do need to allow it to interact with the ground/wall).

@BuddyFriendGuy: We already have this option. I don't know the cvar, but players refer to it as "spawn protection".

For what it is worth I'd prefer Smilecythe's suggestion to be implemented the most.

Thanks @Antibody. I looked it up and It's g_spawnshieldtime. Doesn't that solve the problem?

(update: never mind. I tested it. It doesn't solve the problem. If the seconds are too short, it's not enough time to escape from the devastator's range. If the seconds are too long, it would be too unfair because the spawned player can still pick up stuff while being invincible, which I think may need change, too.)

Thanks @Antibody. I looked it up and It's g_spawnshieldtime. Doesn't that solve the problem?

No one likes to be in the receiving end of spawn killing, but spawn killing is a traditional and important game dynamic in a competitive arena shooter. The point of the spawn system was to make spawns fair, but not a leisure trip that you can just happily walk out of. Dying is supposed to hurt your game and progress, not to reward with an opportunity to come back even in score. I love to spawn kill, but not when it's pointlessly easy.

You only see spawn shields used in casual servers which mostly are all instagib or overkill.

Yes, on public servs where ppl plays for fun, from beginner to advanced skill level, it partly do solve the problem and im all fine with that, but on "casual" servs only, as Smilecythe stated it. But spawnshield is really not the option on competitive games and modes, unfortunately

Thanks @Antibody. I looked it up and It's g_spawnshieldtime. Doesn't that solve the problem?

No one likes to be in the receiving end of spawn killing, but spawn killing is a traditional and important game dynamic in a competitive arena shooter. The point of the spawn system was to make spawns fair, but not a leisure trip that you can just happily walk out of. Dying is supposed to hurt your game and progress, not to reward with an opportunity to come back even in score. I love to spawn kill, but not when it's pointlessly easy.

You only see spawn shields used in casual servers which mostly are all instagib or overkill.

"spawn killing is a traditional and important game dynamic in a competitive arena shooter" is an argument by tradition. It's not a good way to argue cause then "war is a traditional way to solve conflicts between people". You sound really experienced in shooters... but remember you're only experienced in the ones you've played. You're a newb when it comes to ones that haven't been made yet and you cannot use your experience to argue against them. In fact the more experienced you get with something the more you've committed to it and at that point you might become opposed to changes because of that commitment, which in itself doesn't contain an argument.

I'm saying this cause I'm worried you're not truly open-minded. Maybe you can convince me otherwise. Or well it doesn't change anything if you do. Then you should be okay with my post.

Let's explore this idea (again):
-you're let to choose the spawn point yourself within the next 5 or whatever seconds
-if you don't do anything the game has some default spawn system that spawns you somewhere
-this means you can't easily be spawn killed without at least some effort by the other player, who has to guess where you spawned, which in turn will be affected by how good they think you are. This leads to a recursive mind-game which by itself wouldn't IMO be valuable.
-BUT at the same time you have to have an idea of where you actually want to spawn. So because neither player can 100% certainly guess where the other one will spawn next instead of even just knowing the 4 alternatives, all you're left with is estimating the relative worth of each alternative. This you CAN get good at and it makes it at least somewhat less guessing and more experience and quick thinking.
-you may also choose a more "quiet" and safe spawn but you might not be able to do anything about your opponent's continued control
-if your opponent was very low, the situation may be slightly different and they may actually want to try to AVOID you off the spawn. In this case you can try to predict which way they went.

I'm sure a solution can be arrived at as for how exactly to choose the spawn point. Maybe sending the camera to a certain set angle with wall textures disabled (or however that works) so you can see most of the map well. I think the discussion about how exactly to implement this should be completely separate from whether or not it's a valid idea.

He may not have experience in games that haven't been created yet, but neither does anyone else! Smilecythe is a long-time contributor to these forums and an avid (high-skilled) dueler in this game. This being the case I value his opinion. End of story.

What I'm saying is why can't you just test various options without having to commit to one? You know that's a way of gaining the experience...

If you're good at the game, does it mean you're good for the game?

But well I don't know anyone well enough, I was just a bit worried. Also thanks for ignoring everything I wrote.