New Skulkblog Post
DazeGoonBoy
Join Date: 2004-03-19 Member: 27424Members
Join Date: 2004-03-19 Member: 27424Members
<div class="IPBDescription">Approaching the "stacking" problem</div> What I've written is rather large so I posted in on <a href='http://skulkblog.blogspot.com' target='_blank'>Skulkblog</a>. I was talking about the stacking issue, and have come up with a kind of solution. If any developers read this, please read the last few paragraphs to see my idea.
I would also like to hear what people think about my idea. Please look at <a href='http://skulkblog.blogspot.com' target='_blank'>Skulkblog</a>. Thanks.
I would also like to hear what people think about my idea. Please look at <a href='http://skulkblog.blogspot.com' target='_blank'>Skulkblog</a>. Thanks.
Comments
while this is not that bad of a idea, that would take a bunch of resourses to do it seems, and from what i understand, in pubs vets (or aleast me) just mess around in pubs, as to night get messed up skill wise for matchs, and what not. And games are ussly stacked one way or another, because when the game ends, its because you messed up so bad somewhere, that you have omost lost all chases of winning and the other team moves in for the finishing blow. and of course everyone has the chases to get as good as everyone ealse (improve their game play and what not) just some do it more then others and become better <!--emo&???--><img src='http://www.natural-selection.org/forums/html//emoticons/confused.gif' border='0' style='vertical-align:middle' alt='confused.gif' /><!--endemo-->
Stats are always useful for something.
As for everygame being stacked, well, the obvious answers are "Get better" or "play somewhere else". Messing up and using the excuse of stacking is quite petty, it happens, accept it, move on. I like to think that vets just mess around, but its hard to "play down" if you can kill a skulk from accross a room with a pistol, you're likely to do that every time. This is only an idea that isn't *strictly* necessary for the game, but I bet it would make NS even bigger.
I'm just thinking that a ranking system could really open up possiblities.
Any other idea?
no, not quite.... I was thinking more along the lines that an auto-balance would stop a team of generals verses a team of rookies. if there are 9v9, then the equation could be worked out. Say a general is worth ten points, major 8 etc. then the teams could be worked out to an equal standing. auto-balanced could also only come into effect when, say, one team is 33% more than the other.
As for spawn killing instead of finishing a game off, more points could be awarded for killing someone who has not died for a long while. The longer someone has been alive, ther more points No points could be awarded for someone who hasn't been playing for more than a few seconds. So to gain "skill points" to go towards your rank, you must only kill people who have lastest for more than, say, 60 seconds.
Very interesting questions. Spawn killing is an annoyance, especially in combat. I'm sure trueplayers would uphold that spawnkilling at the end of a game is worthless.
I can't see the problem. If the "general skill" on the NZ servers is high, then newbies will getting pwned anyway, what would be the problem in letting everyone know that certain players are very good. If the NZ servers are full of "Generals" and "Majors", then a rookie joining wouldn't upset the balance because in theory, they wouldn't be "that important", until they get good themselves.
I do like playing against a variety of players as well. Maybe giving the servers the option of setting a minimum and maximum skill level would only mean too many servers would set the level too high so no rookies can play. And, no rookies, no future for NS.
Okay, so just having a special auto-balance feature. Nobody should really be excluded from NS.
Well, from the sounds of it, you'd need a database to store that kind of information. You slap the database on a server so that it can be accessed. Gameplaying servers then both read (to retrieve scores) and write (to store scores) to the database [server]. That's a lot of traffic, even if you only read/write at the end of each round.
It sounds a bit costly, considering that NS is funded by charity. I don't know whether it is or not, since I'm no expert - I'm just suggesting why it might be expensive is all.
they used to have a HL mod ranking system with everyones names. scores, play time etc all gathered and made into a huge table on their site
That was a LONG time ago though cause I was still in <a href='http://www.finxy.dial.pipex.com/nc.htm' target='_blank'>[NC]</a>
Ah the good old days <!--emo&:)--><img src='http://www.unknownworlds.com/forums/html//emoticons/smile.gif' border='0' style='vertical-align:middle' alt='smile.gif' /><!--endemo-->
It sounds a bit costly, considering that NS is funded by charity. I don't know whether it is or not, since I'm no expert - I'm just suggesting why it might be expensive is all. <!--QuoteEnd--> </td></tr></table><div class='postcolor'> <!--QuoteEEnd-->
Back when I was playing NS and CS in halls at Uni, one kind fella ran an internal site dedicated to ranks and abilities of the players on his servers. It was mostly for CS, but it showed how much you used/bought each weapon, number of kills for each weapon, accuracy for each weapon, number of knife kills, nade kills, headshots. It gave awards for that week's highest killer, most ****-like awp user and deagle user and greatest number of headshots and knife kills. If that kinda thing was used, surely some kind of rank system for NS should be achievable, as it has less weapons, and no headshots, etc.
Chuz
if the only problem is resources, then i says that its a good idea.
Lets say two skulks with focus lay an ambush for a slow, but fairly constant flow of marines at a chokepoint. The plan is that skulk #1 bite first and draw attention while skulk #2 attacks from behind to finish the kill. In the case where all marines are lvl1 armor or higher, skulk #2 actually gets all the kills, while skulk #1 was just as integral to the event of skulk #2 killing marine#X. Thusly, skulk #2 gets all the ranking benefits and skulk #1 gets few to none. Then there is the fact that this doesn't actually indicate personal skill at all, but ability to work effectively as a teammate with another team member. The only universal factor that you could actually judge to get a picture of all semantics of a round is how many rounds a player has won / lost. Even then, the semantics of general individual skill level on a particular server is lost out meaning that someone with a win/loss ratio at a high level of play would be considered the same as a player with a comparable win/loss ratio at a lower level of play.
Using rankings to determine balance is nigh impossible because there are too many unrecordable semantics that must be put into the picture to come up with a conclusive answer. Thusly, skill stacking cannot be resolved unless vigilant admins monitor the server actively trying to moderate this as best as possible. However, general team stacking CAN be helped as a server option. The idea here is to not auto adjust the number of players on any given team, but to auto adjust entity stats to indicate the disparities between the teams. Here is an example:
M(6) vs. A(7)
Marines are down by one man, so the idea is to make each unit count as 1 and 1/7 of a base unit as far as defence, health, and attack power is concerned. This allows the marine team have a semantic equality to the alien team. In addition, all structures have base values for defence, health and attack, for 1 unit and is multiplied by the number of units on the opposing team making it just as difficult for a 8 man team to take out a hive as a 4 man team would. The effort is to profide logistics equality so that personal skill (while still important) is downplayed somewhat and team skill is in fact promoted... lone rambos simply become useless at some point.
M(3) vs. A(6)
In this case all marine stats are doubled because the other team is twice the size. The attrition factor of the aliens is also played down as it is more difficult for them to destroy structures individually than if they only had 3 units thus forcing their efforts to be more coordinated.
Again, I would only like to see this as a server option and not a standard feature... perhaps some amx mod author could code this up to see how well it would actually work and then give the test results to the dev team to let them evaluate how worthwhile it might be to add this in (if it is not already on the board for addition that is).
Comments?
Is there any reason that a player who is part of an attack shouldn't get points? We have it already in combat, if you are close enough to a kill, you get exp. The idea for combat is that if you're close enough, you're "part" of the action, being a decoy/distraction of biting without getting the final bite. Either way, you get some exp by being part of the team play. In the example given, the second skulk who gets the kills, gets the points. The first skulk will get some points for being a good team player. I can't think of any player who doesn't "go for the kill", if two skulks are waiting in ambush, i'll bet anything both of them will go flying in, no "decoy" method, and who ever gets the final bite, fair play to them, try to get the kill next time. It's not like the opportunities are few and far between.
A ranking system is VERY possible, but i will admit that it won't be absolutely perfect because of the many different aspect of NS (which in my opinion, makes NS a brilliant game). The Players are pretty much all different and good at separate things. A ranking system based on kills alone wouldn't be a good judge, for there are player like me who like going gorge and defending or skulks and be a constant annoyance. This is why I say the scoring system should be secret, so not to encourage any one particular style of play. It's not a crazy idea for someone, who plays 5+ hours a day, can kill anything using any method, can build quickly, can really turn a game around etc. and call them "General". In the opposite way, someonewho does nothing,plays rarely, gets killed a lot, and F4 every game should be called "private". The semantics are not that recordable. Lets lists the "point scoring" events: Kills, building, being close to a kill, healing, etc. In combat,these are recorded anyway, as exp. How about the way the points are earned are slightly tweaked (vanilla skulks killing HA being worth more than Onos killing LM. Building hive/tf worth more than building OC/turret) and they could be saved.
This idea is still very possible (I admit it, I just want to see an NS league) and it could stop stacking, but I'm not a programmer, and I can't do the maths to make it work fairly.
Thanks for the very good comments guys <!--emo&:)--><img src='http://www.unknownworlds.com/forums/html//emoticons/smile.gif' border='0' style='vertical-align:middle' alt='smile.gif' /><!--endemo-->
Don't get me wrong... I'd love to see an NS Ranked League as well, but you can really only do that with predefined teams (ie. clans) and measuring team success (ie. win/loss). This also assumes that the gameplay is adequately balanced for such type of competitive gameplay... which it still isn't for the most part. If there were some game mode other than combat that made 1v1 conflicts viable then you could actually go into some sort of League system on a per player level where such statistics such as K/D and some other measurements actually make sense. However, NS isn't (and likely never will be) a 1v1 oriented game... it isn't really even oriented for small teams(4v4). Good matches can be seen between mid-sized games (6v6, 7v7, maybe even 8v8 though not likely), but then they get disproportionate again with large teams (8v8 and up).
One thing that CAN be implemented at this point in time is Commander rankings where a server eliminates the ability to eject a commander and for the commander to get out of chair, and the commander is then rated on wins and losses. This is still somewhat subjective as there really isn't an equivalent factor on the opposite side, but it is still doable.
Leagues aside though, my previously mentioned solution would address the problem of team stacking and even skill stacking (albeit indirectly and not entirely effective). The only change I'd like to make to the suggestion is where I mentioned the multiplier of structure stats being the number of people on the opposite team. This needs to be changed to the number of people on the larger team (as we are adjusting player stats to be the equivalent of the larger team).
One flaw i have seen in my own plan however is the effects of res acquisition and distribution on disparate teams. In order to address this I suggest making the change that a resource nozzle outputs the number of resources equal to the number of people on the larger team per tick. This would also make for much quicker games... so I'm not sure that I entirely like the idea unless we greatly reduce the tick speed. Another change would have to be making res a floating point number as 6 aliens equally receiving 7 points per tick would have to recieve 1 and 1/7 each tick... either that or only distribute points to aliens when there is enough to distribute a point to everyone from a single pool of "carryover" res. In that case on the 6th tick, each alien would then recieve 2 res instead of 1... much better solution now that i think about it.
Again... comments? I understand that the suggestions I am making probably need to be placed in the S&I forum, but they are related to this thread as well... I'll likely copy them over when I get the time. I also understand that the suggestions imply major changes how some of the game mechanics work and would require a good bit of coding to change it all. I just see this as a means of leveling the playing field somewhat in times where personal playing skill seems to still be emphasized over teamwork in a team oriented game.
With all this talk of Flaya, XP-Cagey, Joev have something big planned. And Cagey being a website guru (says nem0) I can but wish....
Team stacking is a simple problem. Your solution is too elaborate.
Marines are down by one man, so the idea is to make each unit count as 1 and 1/7 of a base unit as far as defence, health, and attack power is concerned. This allows the marine team have a semantic equality to the alien team. In addition, all structures have base values for defence, health and attack, for 1 unit and is multiplied by the number of units on the opposing team making it just as difficult for a 8 man team to take out a hive as a 4 man team would. The effort is to profide logistics equality so that personal skill (while still important) is downplayed somewhat and team skill is in fact promoted... lone rambos simply become useless at some point.
M(3) vs. A(6)
In this case all marine stats are doubled because the other team is twice the size. The attrition factor of the aliens is also played down as it is more difficult for them to destroy structures individually than if they only had 3 units thus forcing their efforts to be more coordinated.
Again, I would only like to see this as a server option and not a standard feature... perhaps some amx mod author could code this up to see how well it would actually work and then give the test results to the dev team to let them evaluate how worthwhile it might be to add this in (if it is not already on the board for addition that is).
Comments? <!--QuoteEnd--> </td></tr></table><div class='postcolor'> <!--QuoteEEnd-->
Here is just one example to show the problem with that system. Let's say in a 6 vs. 6 game, a skulk runs down a corridor at a marine, skillfully running on the walls and ceiling to aviod being hit, whatever, and the skulk is at about 40 health when he gets there. He now has to connect with two bites (assuming level 0 armor) to kill the marine. At the end of the fight, the marine is dead and the skulk has 10 health. Now, let's take that same scenario and look at it in a 3 vs. 6 game (3 marines). If the marine focuses on one skulk now, he will probably kill it before it gets to him, because he is doing double the damage. The second skulk will get there uninjured. However, it will now take that one skulk 4 bites to kill the marine as opposed to 2, and this one skulk at full health will take about the same number of shots to kill at this point as the injured skulk in the first scenario did when he got to the marine. So, it is too advantageous for the smaller team. Also, with this happening, the K/D ratio would favor the smaller team, meaning that the larger team would need an increased spawn rate to compensate.
The biggest point is that it does nothing to address the case where both teams have 6 players, but the teams are completely unbalanced due to the players, and I think that's what the original suggestion was meant to address.
The biggest point is that it does nothing to address the case where both teams have 6 players, but the teams are completely unbalanced due to the players, and I think that's what the original suggestion was meant to address. <!--QuoteEnd--> </td></tr></table><div class='postcolor'> <!--QuoteEEnd-->
It's SUPPOSED to make the smaller team that much stronger. In the case of M(3) vs A(6), each marine is the logical equivalent of 2 marines. How likely is it for a single skulk to kill 2 marines compared to 1? It means that you're likely have to treat that one player as if it were 2 and attack it with 2 or more skulks. The point is to encourage teamwork in general, while making it feasible to be somewhat equivalent in power when your team is offset by a player or two. Rarely do we see cases like M(3) vs A(6)... that's pretty unrealistic... my examples were simply to illustrate the points more comprehensively. In many cases (mid to large sized games) the bonus to one team will be fractional if at all.
Now I will concede that in the extreme cases (M(3) vs. A(6) for example) that the smaller team's units might actually be a little TOO powerful. For instance, if that lone marine is at 1 hp... he's still packing the attack power of two marines when realistically he should at that point only be packing the power of one. But then as I said before, it isn't really meant to balance the extreme case.... though it offers a pretty nice edge for those last few minutes on the losing team after half your team f4's... sorta like an "adrenaline survival rush".
I'll also concede that it doesn't affect skill stacking either... but then I've already said that the only viable solution for that is vigilent admins as opposed to any system you could put in the game.
However, I could see that if there was in fact some sort of ranking system, servers could limit access to the server based on ranges of skill to try and get the most balanced games possible for a given server. That'd especially help out on clan pub servers that are really only looking for decent competition and not too interested in having noobs visit them. Obviously the servers would have to come default with that option off to allow noobs to visit most pubs in general, but its a novel idea at that.
A better way to fix this problem is just to have your server make others join randomly NO MATTER WHAT and unable to f4 in the first 3-5 minutes of the game. This would split teams and randomize it so there would be no stacking, and if you feel bad because you cannot play with your friends in this system then tough luck.
Ps: If your about to rebuke my suggestion by saying this is a team game and that you should be aloud to play with your friends, then tell that to clan members that are not aloud to play on the same team.