Age :45
Group: Moderator
Location: Funland
Activity killed the cat?
Servers  /  24 May 2009, 08:54
KTX respawns #2
During years I've followed the ongoing discussion about spawn models and how they should be and what they should prevent. Since spawn model has rather large impact to the score in games and especially in duels (dm2 water/tele anyone?), how spawn model works should be considered carefully.
We already have many spawn models, but none of them are "really good". For example the normal QW respawns allows you to spawn infinitely to the same spawn spot, and the probability for that isn't all that low either, depending on the number of spawn slots. For example in end you have 3 available spawn slots, one near the upper tele, one at the center and one at the shooters. Earlier today I was testing how often you can actually respawn to the same spawn slot in a row and I found it disturbing when I was able to respawn 6 times to the center in a row. Not to mention 2-4 respawn streaks to the other spawn slots almost continuously. Ie. respawn twice to the center, then three times to the shooters, twice to the center again, etc.

Another spawn model is KT spawn safety. It means you can't telefrag anyone when respawning. This also means is that the player alive can actually block any given spawn slot and "abuse" the spawn model, forcing another spawn slot for the killed opponent. A bit too safe, and never gives the change to retaliate with respawn telefrag (some people actually wait for the right moment to spawn, trying to respawn telefrag the opponent if possible).

With KTX a new spawn model called KTX respawns was introduced. This spawn model allows respawn telefrags like normal QW respawns, but does not allow respawning to the same spawn slot twice in a row, unless all spawn slots are in use. This sounds a bit better, but the spawn model can be abused since you know where the opponent can't spawn. It means that you can walk on top of spawn slots without any fear of being respawn telefragged. You can also ignore spawn slots when you are trying to spawnfrag your opponent with luck (gl/rl shot) and choose another visible spawn slot to shoot at. All this of course, if you can actually remember where your opponent respawned the last time.

The remaining spawn model is Kombat Teams Spawns, but I won't be discussing about that in this blog.

What I'm going to tell you about, is the fresh spawn model called KTX2 respawns. I had been thinking about how to improve the current KTX respawns without making it too complicated. After a while I came up with a solution that might prove to be good or bad, but we'll never know unless we test it. This is how it works:

1) You can get two (or more) same respawns in a row.

2) If the spawn slot is different from the last spawn slot, you will respawn to that spawn slot normally.

3) If the spawn slot is the same as last spawn slot, you don't respawn there yet. Instead the server makes another random attempt to give you different spawn slot. If the spawn slot is still the same as the initial one, you will respawn there. If not, you will respawn to the other given spawn slot.

4) This applies always, so in case you are going to get third spawn in the same spawn slot, the server will redo the procedure explained above. Yes, you can respawn infinite times to the same spawn slot if you are truly unlucky.


Numbers: (map with 6 spawn slots)

Notice that the first number is the chance of getting any spawn on map, let it be the initial spawn when the game starts or just some random spawn during the game from where we start counting. The following numbers are the chances of getting it again.


Chance of getting the same slot: (individual)

Here we have a random spawn, then we check what's the possibility of getting it again. Using normal QW respawns we just choose one from the six possible spawns and spawn there, but using KTX2 respawns we do the re-check if the the spawn is trying to be the same as the last spawn was.

normal QW respawns: 1/6 -> 1/6 -> 1/6 -> 1/6 ...
KTX2 respawns: 1/6 -> 1/36 -> 1/36 -> 1/36 ...


Chance of getting the same slot in a row: (series)

Here we have random spawn again, after which we start calculating the chance for a series of respawns using the same slot. With normal QW respawns the chance of getting the same spawn is multiplied by 1/6. With KTX2 respawns the chance is multiplied by 1/6 * 1/6 because of the re-check.

normal QW respawns: 1/6 -> 1/36 -> 1/216 -> 1/1296 ...
KTX2 respawns: 1/6 -> 1/216 -> 1/7776 -> 1/279936 ...


This way you can never be sure if it's safe to walk on top of spawn slots. Also you have to take all the spawns into account when flooding luck rockets after a kill. And with some luck, you get two same spawns in a row and end up boomsticking your near-death opponent, taking his dropped rl and spawnfragging him a few times. Or then spawn to his rockets a few time


I have enabled this spawn model on my servers by default (players can change it using /spawn command of course), and the KTX revision for this change is 988. This feature will be in the upcoming summer release of KTX.
Comments
2009-05-24, 10:26
So we really want to see 2-3 frag spawn rapes at the same spot on dm4 again?
2009-05-24, 10:55
Did you actually consider about the chances of getting same spawn? How about having same spawn many times in a row?

I edited the blog for chance numbers so that people can see how much difference there will be.

Edited by Renzo on 24 May 09 @ 12:08CET
2009-05-24, 11:29
I might be misreading it, but it seems like your numbers, or your phrasing, is misleading. The chances of spawning to the same place twice in a row on this theoretical 6-spawn map is 1/36, as opposed to 1/6 in normal qw respawns.

That's still pretty low, and we're unlikely to see the horrendous double 9-spawnfrag runs made famous by that UL/Reppie dm2 game.

What I'd like to see is a spawn model that attempts to even out the amount of spawning that you do, ie. every time you spawn somewhere, it becomes less likely that you will spawn there in the future. It would do a lot to even out the fabled RA spawns of dm4, it would make that lolspawnfrag run in that dm2 game I mentioned pretty much impossible without compromising the gameplay.
2009-05-24, 11:43
Quote:
I might be misreading it, but it seems like your numbers, or your phrasing, is misleading. The chances of spawning to the same place twice in a row on this theoretical 6-spawn map is 1/36, as opposed to 1/6 in normal qw respawns.

That's what it actually says, doesn't it?

I gave the individual changes for each spawn (last spawn ignored) and also the chances of getting that same place in a row.

EDIT: Well ok, now I see it. The first number on those series is always "random spawn", and the second one is the FIRST possibility of getting that same spawn. Others note that too.

Edited by Renzo on 24 May 09 @ 12:45CET
2009-05-24, 11:45
My apologies. It was just a wording thing. I see what you were saying now. Do you have any thoughts on my suggested model?
2009-05-24, 11:52
Edited my previous reply, but let's your suggestion. It looks a bit complicated code, even though the idea is rather ok, with exception. Did you consider the fact that if you get lots of spawns in the first 5mins to a certain area, let's say dm2 water/tele, you are bound to spawn more to the low area towards the end of the round? If your opponent(s) can anticipate this, then he's probably going to try ignoring the water area and rush towards low after each frag.

The problem is always that the spawns shouldn't be all that predictable and you shouldn't be able to change your way of playing because of them. Of course I took only the worst case scenario into consideration with your spawn model but that can happen. Like that spawnfragging with KTX2 spawns, but even so.
2009-05-24, 12:01
That's a very good point.

Perhaps the odds decrease could be relatively low (maybe a compounding 10 or 15% decrease) so the chances of it affecting dm2 would be insignificant, but a 30 or 40 frag dm4 game would note a significant improvement in spawn fairness.

It sounds like it might be a coding minefield, but think of the benefits to dm4 alone if implemented successfully!
2009-05-24, 12:33
regardless of spawnmodel, i think it should be kept simple.

I like the "completly random, but not the same spawn twice in a row"- best.
2009-05-24, 14:38
The certainity of safely moving through ya-entrance spawn point after spawn fragging an enemy on it has been removed.
2009-05-24, 14:40
Yeah, that was always a huge thing for me while playing dm4. The knowledge that certain spawns were "safe" meant I could focus my attention elsewhere safely.
2009-05-24, 17:23
Suppose average dm4 duel game gives total 40 frags = 40 spawns, chance of doubled respawn = 1/36, chance for not doubled respawn = 35/36; chance for no double respawn in whole match: (35/36)^40 = 32.4%, chance of AT LEAST ONE double spawn = 67.5%.
In 2 out of 3 dm4 games there will be at least one "double respawn".
2009-05-24, 17:55
In average TP game (400 frags - 400 spawns) there will be 6 doubled respawns.
Chances for "at least 3" is 95%, chances for "more than 15" is 0.2%
2009-05-24, 18:30
I like this update, hope it will become standard.
2009-05-25, 19:38
I can make it default in KTX and the server package release.
2009-06-18, 11:15
I really don't like this change. When changing the game rules there are two things you have to balance, 1) fixing something, 2) not making players have to change the way they play too much.

Because of the probabilities this change doesn't really fix anything at all. But it still manages to force changes on player's routine in various spots: Need to fire that extra spam rocket or strafe past that spawn, when last 5 years you have learned to run straight through it to save time.

The 110 direct hit rocket damage change would be nice example of something that fixes stuff and doesn't force people to change their game. Anyway, I wouldn't be opposed to new spawn system, if it actually makes the game better rather than just little bit different.

Also would really appreciate stuff like this being in the news or something, since I for one always miss these blogs.
2009-06-20, 18:16
Medar wrote:

...

1) you can always toggle the old spawnmodel, which is considerably worse since you just know when it's safe or when you won't get a frag from a spawn with flood rocket. Do not expect any player to agree with you with the spawn model change though.

2) People have stated (Fucu and Milton are the ones worth mentioning) that they didn't like the old KTX spawn model, which made me to re-think the spawn model, and this KTX2 spawnmodel is the one that gets rid of 100% certainty while the possibility on a 6 spawn map to respawn twice the same spot in a row is 2,8%. The more spawns, the harder it becomes.

3) If I want to have new spawn model on KTX, it will be done. If I want to enable it on my servers, it will be done. If someone else doesn't want to enable such spawnmodel, whatever they do is fine by me, don't enable it then.

The bottom line is, feel free to dislike the new spawn model. Other people might dislike it too, some people might love it, some people won't probably care since they won't die 70 times in a duel round.

Medar wrote:
Anyway, I wouldn't be opposed to new spawn system, if it actually makes the game better rather than just little bit different.

..

Also would really appreciate stuff like this being in the news or something, since I for one always miss these blogs.

There has been a lot of spawnmodels in the history of QW, the KTX spawn model has been around for less time than the older spawn models. Nothing ever changes if noone is doing it, just look at QW map pool how "versatile" it is.

I'm not going to make news about every damn change we have on KTX, only the ones that are the most visibile and/or affect the gameplay the most. Any and all changes will be on the KTX release notes once we get new release this summer.
2009-06-25, 12:54
1) I much much rather have only this spawn model than ability to toggle between different ones. Having a single set of rules for the game is really important and I hate the attitude "we really don't know what's best, so every server admin/player can just configure it to what they like". Not that it's a problem in QW anymore too much.

2) I don't think the old KTX spawnmodel is the best one either. But I also think that your change helps very very little. You keep saying that it doesn't matter that much if you don't die 70 times in a duel, so how it's supposed to fix anything? Rather than such a small change, I would look into different models that are not solely based on your last spawn. You can also look at dying location, enemy location, time before your last spawn etc. It doesn't make too much sense to have special treatment for your last spawn spot if it was 3 minutes ago.

3) There is absolutely no need to freak out like this. I know this is exactly the situation I would like the least. I want every server have the same game rules. I want there to be one QW, not QW the renzo way and QW the <another server admin> way. I'm sure people can be reasoned with, and you don't have pressure them with your position as a developer and server admin.
2009-06-25, 16:38
Medar wrote:
...and you don't have pressure them with your position as a developer and server admin.

You must be confused.

It is the league admins that make the rules for their leagues, not me. They can choose whatever spawnmodel for their league if they wish so, however in 4on4 spawnmodel means nothing, much like in 2on2, it very seldom decides the outcome of the games. Duel is different of course, but even so the figure 70 I gave above was exaggeration, it's really 1/36 on a map with 6 spawns.

Developer and server admin? You forgot to mention as player, even though I haven't played that much lately. Also I notice the word "pressure". Now, if ANY server admin or whoever thinks I've told them to use this particular spawn model, raise hand. Noone? I thought so. And that's because I haven't bothered my time wasting messaging server admins to change their spawnmodel or update their servers for this particular build (or newer) since I've got other things to do than that. I... hmm, we (as in KTX team) just give the possibility to use the spawnmodel, and even endorse it (at least I do) since it's somewhat better than KTX spawns. And the settings I use/default on the servers I run, well, that's my business, noone else's. Not happy people can toggle the settings when they play, if required.
2009-06-26, 12:38
Like it or not, in practice you are making the decisions about these rule changes for the entire QW community. When you include something as a default setting an the servers, and the new KTX release, changes are very high it will be used in tourneys/leagues too. Sure if it's something major, and people really hate it they will reject it, but minor changes like these will go through.

Most people just want to play the game anyway, not stress about some rules and stuff. So they will accept whatever happens to be used as a default.

Just hope that you do understand how much power you have in these decisions, and try to consider other people's opinions too when making changes.

Plus I'm not saying it's bad thing you are making these kind of changes. QW needs a driving force to push it forward, but trying to include people and ask for their opinions before the changes happen would be nice.
2009-06-28, 18:23
Medar wrote:
Just hope that you do understand how much power you have in these decisions, and try to consider other people's opinions too when making changes.

We (the KTX team) always give chance to use different settings, but we like to default certain stuff on because we think it's the best possible setting for our mod. Big changes always require more attention from players, also I think I mentioned this in one of the previous replies, but some people wanted to have another spawn model. The spawnmodel I came up with changes little, so that it can't be considered something that changes gameplay drastically. Also I see people liking this spawn model in these blog replies, so at least someone likes it.
2020-10-29, 20:24
2020-10-29, 20:31
2020-10-29, 20:34
Tried to delete posts apologies. If admin sees this message and two previous posts can delete them ty.
You have to be logged in to be able to post a comment.
Username:
Password: