Spawn Killing

First off, I don't wanna get into a whole debate about whether it's fun or not or should be done or not...I am sometimes spawn killed and I myself spawn kill, as well, occasionally. But I do think that it would be good to give players a somewhat better way to defend against it, if the server admin chooses to enable such settings. I mostly play InstaGib CTF so I see spawn killing happen quite frequently and often many times in a row, which is why I started thinking about it. Since I don't have a deep understanding of the game at all, here are the mechanisms I know of that can help defend against spawn killing:

1) Mixing up the spawn points: I've read in a 2016 discussion on this forum that the algorithm that determines the player's spawn point was under examination; however I didn't understand whether any of the changes talked about there were actually implemented. AFAICT there does exist an option to let the player spawn furthest from the enemy, but this is actually counterproductive on Insta servers, since only 1 hit is needed for the kill and the spawn killer could even abuse this mechanic to force the player to spawn in the same spot or just spawn kill from further away to be able to view all of the spawn locations at once. Mostly I think it would be good to allow for an option that prevents the player from spawning in exactly the same spot again, since that is just too easy for the would-be spawn killer, especially with respawn times being so low the spawn killer doesn't even have to do anything except click again.

2) Spawn Shield: this does exist on most (or all?) InstaGib CTF servers, but lasts for an incredibly short amount of time; I think I only hit a player's spawn shield once or twice in hundreds of games and it certainly doesn't allow the spawner to orient himself with regards to his position before it runs out. However, I am not in favor of increasing the spawn shield duration for various reasons; I just wanted to remark on its existence and nigh-irrelevance.

3) There was some talk in 2015 about letting the player choose the spawn location; I don't know if that's been implemented but I wouldn't be in favor of that anyway.

4) Delaying respawn to offset the spawn killer's rhythm: This is not really effective on Insta servers, since 1 hit = kill and it doesn't really make it harder for the spawn killer (could be useful in regular game types, though)

5) Using a blaster jump right after spawn: I haven't really experimented with this yet but I suspect that the spawn shield doesn't last long enough till the spawning player moves his mouse down and right-clicks for the blaster shot, although I could be wrong. If the spawn shield actually covers that duration then this would be viable, though very annoying to do every time there are enemies near the friendly base (which is often an unknown, anyway).


So I have some questions:

Was the behavior of preventing consecutive spawns in the same location implemented? If so, I guess the servers I play on seem to have that option disabled.
Is it possible to force off the spawning bloom effect for every player from the server side? This effect isn't really needed and it kinda helps the spawn killer since it calls the spawning players to his attention even more.
Are there any other mechanisms I'm not aware of to defend against spawn killing?


And lastly, here is a small suggestion to add a further mechanism:
This has been talked about a bit in 2015 but as part of a much larger enhancement, but I think this small change would already be enough:

Move the camera to the spot the player will spawn at, 1 second before he spawns.

That's it, nice and simple. This would allow the player to orient himself to the location of his spawn and maybe get a glimpse of where there may be enemy spawn killers but it wouldn't allow any overly effective plan to finish off the spawn killer. It simply levels the playing field since the spawn killer will often know in what direction the new spawns will occur and can already aim that way whereas the new spawns don't know where they will spawn and where they will be facing. With this suggestion implemented, however, the spawning player would have a chance to orient himself and find out where he needs to aim or move to have a chance at killing the spawn killer. I thought of this because I have the feeling that most of the time when I get spawn killed or successfully spawn kill someone else it is because of this short time span that is needed for orienting oneself. Thus, showing the spawn point 1 second ahead of spawn would mitigate this problem at least partially.

So: is this kind of thing already possible? If not, can it be implemented? If yes, what do you guys think about it and does it have a chance of being implemented?

Honestly I think it's a problem with bad level design. I mean the most popular instagib maps have their spawns next to the flag. And if you don't spawnkill you're less likely to get the flag.

Recently Vampy and I made a map where we found a space efficient solution to that problem while also eliminating the need of a respawn timer and spawn shield:
A dedicated spawn room with one-way portals.
So you waste some time to get back fighting while already playing. Next thing was adding some tactical depth by letting each player choose their spawn location and direction.
You need to have reflections turned on but I made sure the portals wont impact performance by setting their render distance to less than the distance I buried them into the walls (+vis portals). So the portals should only affect performance while you're in the spawn room.


With time slider: https://imgur.com/2rekqJl

The map is called Lilium and is playable on regulars servers. (also shameless self advertisement going on here ;P - but we really sat one evening tackling exactly only this problem in a way that satisfied both of us)

EDIT: What shitty self advertisement not even providing a download link... well here you go if you're interested: http://www.mediafire.com/file/b4eb5nqbg3...4.pk3/file

The biggest problem with spawn killing has never really been those few occasional unfair situations, but rather the sessions where you get spawn killed multiple times in row systematically. The best and simplest solution to spawn killing has always been:

1. Random spawn location, makes the whole process unpredictable and harder to chain multiple frags.
2. Instant respawn, gives you less time to prepare for a spawn kill.
3. Self controlled spawn delay, in vanilla this can screw up someone's item timing.

Yeah I played that map recently, _para, it's fun! And yes, using a different level design, like the one you two implemented there, is a pretty awesome solution to the spawn killing issue; however, there are simply too many maps already out there that don't satisfy this solution and it would probably be quite tedious if not impossible to "upgrade" them all and there would also be some opposition to this, I imagine, so I don't think it's something that can solve the overall problem, though I do like the idea.

As to Smilecythe: I do agree and this is especially true for Insta maps since spawn killing is so easy there (just a well aimed click), which, however, also means that your second point doesn't really help since it's just a click and thus pretty much instant kill anyway; you now have the advantage on the side of the spawn killer that he does not become disoriented like the spawner because he doesn't have to move whereas the spawner will reappear in a somewhat unexpected position and orientation. Point number 3 only really works if the spawn killer gets bored after several seconds of no spawns and then does something else.
The random spawn location certainly helps, but does anyone know if the idea that the random option should not spawn people in the same spot twice in a row was implemented? And would it be possible to force off the "spawn glow"? I feel like those two things may help a bit by making spawn killing slightly harder.

Overall, I think that introducing the option of putting the camera into the spawn location 1 second before the actual spawn is still a good solution, because it works for all maps, no matter how bad spawn killing is there and it solves the biggest issue of spawn killing disadvantage for the spawner, which is lack of orientation due to spawning essentially without a shield in the middle of the battle while everyone else is naturally already oriented.

Is this something that could be implemented in the future?

I think for Instagib the longer spawn shield should work.

We been down that route before. People in instagib have really fast reflexes and snappy aim, making even half a second feel like an eternity. With spawn shields its just gonna be in favor of the spawner, the opponent(s) has to pretty much hide meanwhile which just slows the pace of the mode down. Otherwise it easily becomes a back and forth contest of people killing each other fresh off their spawns, with little reward to actual kills in general.

Bloodthorn: your second point doesn't really help since it's just a click and thus pretty much instant kill anyway

But you have less time to prepare for that click after a kill and when you're forced to rush you miss a lot more often.

Yeah I don't like the idea of increasing spawn shield duration either, due to what Smilecythe said but also because it just seems kinda lame to give players the option of invulnerability for getting killed as a kind of reward. I also had an annoying situation on a server today where I actually hit a spawn shield for like the 3rd time in months and this was when I was aiming at an enemy behind him that was camping there and the spawner prevented me from fragging the camper which was pretty lame as you can image since I wasn't even spawn killing.

As to the "less time to prepare for the click"-thing: I don't feel like I need time for that...if I see an enemy I aim and click as fast as I can...no matter if there is just one or many (the latter case would be equivalent to rapid respawns, which are pretty much enabled on the Regulars server anyway...I mean it's not completely instant but almost and the initial half-second the spawner needs to orient will make the response not-instant anyway, even if instant respawn was enabled).

So anyway...I still think that this 1-second "preview" of the spawn location is a really good way to mitigate the negative sides of spawn killing that doesn't have any real drawback and since I've done some Java programming before (haha) maybe I can take a look at the code if no one else wants to take a gander but I sure would appreciate some pointers of where to look first!

IMO with a map you generally want to aim to make sure that:
1. The spawn locations generally aren't out in the open.
2. There's some distance between them (ie: not clustered). Preferably you want them spread out a bit and even not in LOS of each other, especially if team based.
3. That things like some armour and weapons are close-by (if applicable to game type).

That has to be balanced as part of the map and its aim, but they're the sort of things I look for in a good map.

So anyway...I still think that this 1-second "preview" of the spawn location is a really good way to mitigate the negative sides of spawn killing that doesn't have any real drawback

The drawback is that now the one who spawns holds the advantage, because he can potentially see his enemy and prepare in advance to hit them. My prediction is that it'll have virtually the same effect as with the spawn shield: game becomes a contest of reverse spawn frags instead of frags themselves.

Doombringer has the spawn system that you're talking about, you'll see some instances of it used in the link. I can only imagine it being abused more distinctively in instagib. Even if it's not abused effectively, it's certain to slow the pace of the game down.

EDIT: I find this feature actually more applicable for vanilla than instagib, because in instagib you have all you need to get back into the fight right off the spawn.  In vanilla you need to find a weapon and armor before you can really do anything, if you don't do that and just rush back into the fight.. you might as well be spawn killed.

So anyway...I still think that this 1-second "preview" of the spawn location is a really good way to mitigate the negative sides of spawn killing that doesn't have any real drawback

The drawback is that now the one who spawns holds the advantage, because he can potentially see his enemy and prepare in advance to hit them. My prediction is that it'll have virtually the same effect as with the spawn shield: game becomes a contest of reverse spawn frags instead of frags themselves.

Doombringer has the spawn system that you're talking about, you'll see some instances of it used in the link. I can only imagine it being abused more distinctively in instagib. Even if it's not abused effectively, it's certain to slow the pace of the game down.

EDIT: I find this feature actually more applicable for vanilla than instagib, because in instagib you have all you need to get back into the fight right off the spawn.  In vanilla you need to find a weapon and armor before you can really do anything, if you don't do that and just rush back into the fight.. you might as well be spawn killed.

I checked out the link and they use significantly different mechanics there from what I had in mind: in Doombringer you have to wait a second and then the camera goes to the spawn point and you have the option to wait an arbitrary amount of time before actually spawning (at least that's how it seemed to be to me).
This is not what I have in mind as a feature, though; my idea was that as soon as you press your respawn button, the game will check whether you can respawn in exactly 1 second or less (if not then nothing happens until this is the case) and if so it will move the camera to the spawn point and automatically respawn you in exactly 1 second. This means you cannot "wait and check" on your spawn point first to see what's going on there and spawn at will but rather that once you see your spawn location (and this only for 1 second) you will spawn automatically in 1 second instead of when you want. This is to simply get rid of the disorientation of suddenly being in a somewhat unpredictable and unexpected location, not to allow you to  plan or prepare anything.
So I don't think it gives the spawner any great advantage at all, but rather removes his disadvantage of being completely disoriented at spawn, unlike the spawn killer, who knows where he is on the map (he is oriented) and also knows where the possible spawn locations are (and usually he will try to keep them all in his field of view). Now, one could argue that this means a spawn killer, who likes to actually camp and remain completely still for maximum aim while spawn killing will be more likely to get wasted, but this isn't really a bad thing because that kind of spawn killing is just super lame, especially in InstaGib; if you're gonna spawn kill, you should at least have to wobble around a bit and occasionally change locations...it's not supposed to be like taking candy from a baby.
Edit: just to further clarify: this mechanic would mean that in some cases the spawner may see the location of the spawn killer ahead of spawn but he will still have to aim his mouse there after actually spawning (your camera movement before spawn should be reset upon spawn) and he won't be able to choose a good moment for spawning or "prepare" for spawning (since it's only 1 second); meanwhile, the spawn killer knows that the spawner will spawn in his field of view, which means he will also simply have to move his mouse to the spawner and click; one may argue that he has the disadvantage of not knowing exactly where the spawner will appear, but if he chose a decent location for spawn killing, it will be in his field of view and will require less mouse movement than the spawner would require (if the spawner sees him at all, which he often won't if the spawn killer has a good location that he changes frequently). So it would require slightly more finesse on the part of the spawn killer but he certainly wouldn't be at a disadvantage, he simply has different advantages from the spawner; and, btw, I would have no problem with completely disabling the pretty much useless spawn shield completely if this was implemented.

So the only potential drawback would be that it disallows instant respawn but firstly I've not seen a server so far that has instant respawn, anyway, and secondly, if someone thinks that instant respawn is super important they can always choose (as the server admin) to disable this feature; but again, even on the Insta servers there's usually at least a 1 second timer for respawning so all I'm proposing is to create an optional feature that  makes that 1 second "mandatory" (meaning it can't be skipped if the player wanted to observe his death location for a bit, first, for example) and puts the camera at the spawn location 1 second before spawn.

I can't quite imagine how your idea turns out in practice so I'll just wait and see for a test run. Perhaps it has different application in CTF, I've only really played 1v1 actively and imo there was nothing wrong with the spawn system as it were.

In fact any flaws with it have always been related to map layout rather than the system itself. Here's couple video examples: 1, 2.

even on the Insta servers there's usually at least a 1 second timer for respawning

You must be playing CTF almost exclusively then, because instant spawns were and I think still are fairly common in 1v1. The longest forced respawn I can remember on an instagib server was 0.7sec in the Prophets servers. Maybe things are different today and it indeed depends on the admin like you say.

An alternative to the glowing spawn shield could be Red Eclipse's approach with temporary (nearly complete) invisibility. The player would be vulnerable, but hard to see, the biggest giveaway being the spawning effect.

Regarding spawn point effects; spawn camping was a problem in Nexuiz at times too, despite it lacking such effects. Players quickly learn the spawn locations, especially on the popular maps, the effects just provide visual aid to speed up that learning process.

Regarding spawn point effects; spawn camping was a problem in Nexuiz at times too, despite it lacking such effects. Players quickly learn the spawn locations, especially on the popular maps, the effects just provide visual aid to speed up that learning process.

Yep totally agree these spawner effects are stupid. But I still think mappers should put more effort into their spawns and maybe make cluster-spawns for randomness.
One other solution could be a ray-traced spawn location jitter.
Or better a info_spawnarea_[team1,team2,deathmath,etc]. So you define not a point but an area where players could spawn randomly.

But I still think mappers should put more effort into their spawns and maybe make cluster-spawns for randomness.

Part of the problem is that one side is expecting good results for instagib modes and one side for vanilla. The features and balance decisions you do affect the two sides differently. There aren't many duel/FFA maps that take instagib spawn killing into consideration - simply because it's not as big of a problem in vanilla. On another side, if you overdo the measures against spawn killing, view blocking every spawn point or such you end up with an abundance of redundant clutter for Vanilla. CTF is prolly the exception, because it's pretty much the only mode that has it's own instagib exclusive mapping scene.

But I still think mappers should put more effort into their spawns and maybe make cluster-spawns for randomness.
One other solution could be a ray-traced spawn location jitter.
Or better a info_spawnarea_[team1,team2,deathmath,etc]. So you define not a point but an area where players could spawn randomly.

Overkill uses teammates as spawn points, it tests 6 places behind and to the sides of the player. It makes sure it's not above or in front of void. The code could be adapted (or as we in Xonotic like to do - copy pasted) to generate additional spawn points. Insta could use the original spawn point as the start and place new ones while fanning out the search from there. Vanilla could use items as the start and put additional spawn points near megas or weapons.

You must be playing CTF almost exclusively then

I guess so; I've tried most of the other game modes but I guess for now I just like Insta CTF the best by far.

An alternative to the glowing spawn shield could be Red Eclipse's approach with temporary (nearly complete) invisibility. The player would be vulnerable, but hard to see, the biggest giveaway being the spawning effect.

Regarding spawn point effects; spawn camping was a problem in Nexuiz at times too, despite it lacking such effects. Players quickly learn the spawn locations, especially on the popular maps, the effects just provide visual aid to speed up that learning process.

This wouldn't really work for Insta, though, because 1 click = dead so the spawning effect is enough for that....most of the time when I spawn kill I will have a good chance of dying if I don't kill people right after spawning (and vice versa when people try spawn killing me).

Is there a way to include some server option that forces spawn effects off for all players?

But I still think mappers should put more effort into their spawns and maybe make cluster-spawns for randomness.
One other solution could be a ray-traced spawn location jitter.
Or better a info_spawnarea_[team1,team2,deathmath,etc]. So you define not a point but an area where players could spawn randomly.

Overkill uses teammates as spawn points, it tests 6 places behind and to the sides of the player. It makes sure it's not above or in front of void. The code could be adapted (or as we in Xonotic like to do - copy pasted) to generate additional spawn points. Insta could use the original spawn point as the start and place new ones while fanning out the search from there. Vanilla could use items as the start and put additional spawn points near megas or weapons.

While I still prefer my idea, because it takes care of the root of the problem (at least for Insta), which is disorientation in the first second after spawning, I understand that it might be tedious to implement. The idea you describe with the random spawn would probably help a bit since there is, of course, a benefit to knowing the spawn locations that gives an advantage when spawn killing and this certainly sounds like it could lessen that advantage.

The spawning effect can be turned off server-side with g_spawn_alloweffects 0.

The spawning effect can be turned off server-side with g_spawn_alloweffects 0.

Thanks, gonna try to hunt down the server admin then...I think it's Slava for the Regulars Insta CTF server.

Anything I can do to help with implementing the copy-paste-type fix you talked about or perhaps even my idea? I may even be able to help with code but to be honest I'm pretty much an amateur who has mostly played around with Java thus far...but if you point me in the right direction I can check if it's something I can maybe get into.

Just a little update:
I was able to contact Slava and got him to disallow spawn effects server side on the "The Regulars" Insta CTF servers, which may help a bit with spawn killing in some instances.

In discussing this with him he had some good ideas about helping further with spawn killing:
It's possible to detect spawn killers by logging the time a player is alive (spawn until death) and running a routine upon each player's death if this value is below say 1000 milliseconds (1 second). This routine could look at who fired the killing shot and thus identify the spawn killer for possible reprimands of sorts or react in some other way.

What Slava had in mind specifically, as I understand it, was to simply increase the time the spawn shield is active for the spawn killed player once he is spawn killed more than once in a row. This shouldn't bother too many people, I think, since there is no real penalty for spawn killers but instead this just makes it harder for them once they spawn kill for a bit. As soon as the player dies in more than 1 second this mechanism resets for that player. Slava didn't mention by how much he wants to increase spawn shield time but I think 500 milliseconds or so should be enough. This is for public servers running Insta CTF, though, as I doubt it would make much sense to implement this in vanilla game modes or on tournament servers for various reasons.

As I understand it, Slava may implement this on the "The Regulars" servers soon and I'm excited to see if this will work.

Actually variant with dynamic spawn schield was suggested by malice. I will respond here when it will be ready for testing. Also, I have few other options on my mind.

3) There was some talk in 2015 about letting the player choose the spawn location; I don't know if that's been implemented but I wouldn't be in favor of that anyway.

Actually this was implemented by TimePath, if I recall that correctly, and there was even few duel servers that had such feature (it was votable). But this feature aren't become popular, cause it slowed game and it wasn't merged in xon master branch.

The spawn selection was actually implemented by me, just as a proof of concept mod using existing functions from the onslaught gamemode. The issue wasn't slowing the game (it takes up the same respawn time), but rather giving the spawning player an advantage over the survivor(s).

Hi and sorry for delay. I implemented new dynamic spawnshield and it's currently available on regulars votable server. Tomorrow it will become available on main regulars server and probably after this it will disappear on votable.
This change introduce 2 features: spawnkill counting and spawnshield time that dynamically adjusted depending on how many times you was spawn killed in a row.  If player was killed in sv_spawnkill_time (1.6 on regulars) seconds after spawn, this kill is considered to be spawnkill. You can see counters for spawnkills in scoreboard (if you reconfigure scoreboard) or with teamstatus command. First one counter (spawnkilled) shows how many times player was spawnkilled and second one (spawnkills) shows how many times player spawnkilled someone. Spawnshield will work according to this logic.
Here's an example of how it's working:

• regular spawn (0 spawnkills): g_spawnshieldtime (0.15 seconds on regulars)
  
• first spawnkill in a row: g_spawnshieldtime + g_spawnshielddynamic (0.15 + 0.6)
  
• second spawnkill in a row: g_spawnshieldtime + g_spawnshielddynamic + g_spawnshielddynamic / 2 (0.15 + 0.6 + 0.3)
  
• third spawnkill in a row: g_spawnshieldtime + g_spawnshielddynamic + g_spawnshielddynamic / 2 + g_spawnshielddynamic / 4 (0.15 + 0.6 + 0.3 + 0.15)
  

Maximal possible value for spawnshield is g_spawnshieldtime + g_spawnshielddynamic * 2 (1.35 on regulars).

Please leave some feedback.

spawnshield time that dynamically adjusted depending on how many times you was spawn killed in a row.

That's like how Quake Champions did it and having tested it - all I can say that frankly it's cancer. The problem with it was that it's impossible to tell who's going to be spawn shielded for how long. You might run into someone, who's been killed many times in row by someone else and they now have an unfair edge over you. Even if you have a visual or a clue that someone is shielded, the varying duration is what makes the experience inconsistent.

Also, why so long spawn kill time? 1.6 sec is enough to hook-rush quarter/half way of many of the space floater maps.

Few people have asked how to show spawnkills in scoreboard. You can use this command: scoreboard_columns_set spawnkills spawnkilled. To return back to default use: scoreboard_columns_set default. Here's how it looks in scoreboard.

Also, why so long spawn kill time? 1.6 sec is enough to hook-rush quarter/half way of many of the space floater maps.

I pick that value because it was easy to test with, it's not hardcoded and it's easy to change it. I just played few games with this thing against real people and my feelings that it really helps against players who deliberately camp near spawn point in order to make spawnkill, especially when you spawned in same spot few times. Also in comparison with old settings (previously server had 0.45 spawn shield) I hit spawn shield less times.  But I played only few games, so I might be wrong. 

Also, when I tested these settings against bots, I found that it might really gave unfair advantage in some cases. For example on small map, like oh_crap 1.1 seconds is enough to make capture. But in order to accumulate this spawnshiled time player should be spawnkilled 3 times (not just killed). Also, I'm not going to use these settings for competitive servers (that's why I going to disable it on votable after testing).

I haven't had much time to test it since I've been pouring most of the free time I can spare (which isn't much) into the Defrag WC, but I surely appreciate that you made this effort, Sl@va!
FWIW, in the few games I've played the spawn shield change didn't mess anything up; in fact, I even intentionally spawn killed a bit to see how it would work and I even had the feeling that the timings could be set a little higher. In any case, I think this change is quite sensible for non-competitive servers as those servers are for everyone; if you wanna go spawn kill, go to votable; if you wanna go take candy from a baby, however, well too f*king bad, you're gonna have to deal with the changes.
I'll report back after the WC and more testing though.

Edit: after many games of watching the system in action I must say that it does what it should, which is making spawn killing harder. It is still possible to spawn kill since the spawner still has to hit the spawn killer with the first shot before the shield wears off, which may not be easy because of the necessary reorientation on spawn, but it gives the spawner a fighting chance. In many hundreds of kills with this system only one was denied to me because of the extended spawn shield without me spawn killing (this was on warehouse and the spawner spawned outside so I could hit him when I was flying through the air toward their base) and even then I think it's fair enough that this happened since the guy must have gotten spawn killed in order to get that longer shield.