Tentative point values:
The point system has a twofold purpose: To provide a more accurate measure of a player's skill, and to reward some behavior that is more effective than others (traveling alone vice traveling in small packs, not camping enemy bases unless your team is doing something productive to that base, like capping it)
Points are based on the following concepts:
Points based on kills have the following properties: Killing ships in sectors where the enemy owns a base that is dockable/pod recovery is worth very little. This is to compensate for the following effects: Dragging a game out and whoring an enemy at their home will provide very little point gain. Killing an enemy where they can dock their pod means they can just launch another ship, so if they're defending a miner, you kill them, they just get another nan. Note that in the previous example if you actually kill the miner you will however get quite a lot of points.
Bomb runs distribute points by giving points to all allied ships in the sector, including nans and fighter escorts. Because turrets receive points from kills, turrets and ships that cannot mount nans get a 25% penalty on the above bonus so that good whoring gives more points, but they still average out to what everyone else receives. The way that bomb runs are scored is like any normal kill, with everyone in the sector doing equal damage (with fig escort/turrets getting 25% less).
Point distribution on kills, bomb runs: Provide points based on any kill to the following formula:
Min((%damage * K * (1+ (n-1)/(n+1))), K), where K is the kill value of an enemy, and n are the number of people who have attacked an enemy. The following scenarios describe the effect of such a formula.
Say one person kills an enemy worth 10 points, than you get Min(100% * K * (1 + (1-1)/(1+1)), K) = min(1 * 10 * (1+0), 10), or 10 points. If two people deal equal damage to an enemy, than they each get .66*10 points, or 7 points. Additionally, the Min function ensures that it's not possible to get more than K points on any given enemy, no matter how much damage you do. Net result is that working in a team is more productive point-wise than just traveling solo. Additionally, the total amount of points given out for any kill can never equal more than 2*K.
Lastly, an ideal property of base captures is that the team must control it after 1 minute. If they no longer control it, but it still exists, 0 points. If the base is destroyed, count it as a base kill. So say someone caps your base, and before they cap it, you launch in a bomber, turn around, and blow the base after it's capped. Even though you destroyed the base, you should receive no points for it. The enemy team, however, would only get credit for a base kill, not a base capture.
Lastly, then 'fun' part, the hard point values.
Heavy constructor kills: 70
Light constructor kills:40
op/tp (non giga) base kills: 30
op/tp (giga) base kills: 25
Refineries: 15
Normal Tech bases: 60
Adv Tech Bases: 90
Miner Kills: 40 + he3/5 currently on miner
Adv tech base captures: 105
Normal tech base captures: 75
Light base captures (non-refineries): 40
Giga spec ref captures: 30
Dregh He3 mine captures: 15
Kills in sectors with enemy bases: 2
Kills in sectors with no bases: 6
Kills in sectors with friendly bases: 7 (enemy base rule takes precedence)
Killing enemies which have a living miner on their assist list: 20
Killing enemies which have a living con on their assist list: 20
Killing bombers: 25
Killing stealth bombers: 18
Cap ship kills are worth their cost/1000, so 30 for a corv, 80 for a cruiser, etc.
Killing HTTs: 30
Dropping a probe that first eyes a miner: 10
Dropping a probe that first eyes a constructor: 10
Dropping a probe that first eyes a bomber: 15
Lastly: Winners get their points doubled like under the current system.
Point System Proposal, Values and Formulas
Wow, there certainly is a lot of talk about individual point systems and such around. Not a bad thing, but I would like to point out a few things that should be considered. I'm trying to be constructive here, as I have never flames on this or any board, so keep that in mind here.
Secondly, along the same lines as the old argument that "any point system can be baked", I would propose that any point system that is too specific can be baked, boiled and fried much more easily. Keeping things simple may not reflect worth with 100% accuracy, but making a system too complex with wide point differentials between such similar actions would hinder the system as a whole.
Third, there doesn't appear to be a range in which all of these points should be confined. Even with miniscule divisors, after a year, people could likely be saying that their rank is 264,196. That's quite large, no? An individual point system based on deviation from a midpoint would be much more usefule for guaging players against other players. This is the way that elo does it.
Now, as a parting note I would like to say that I am NOT a fanboy of elo. I'm not particularly fond of the system, but it is what it is, and it's better than nothing. I would like to see a system similar to what you are proposing so that we could compare it with elo and everyone could make informed decisions, but there are definitely scaling problems with an individual point system based from 0 and ending at no set point.
Hope some of this helps with your concept, as it is actually quite intriguing to me.
Alrighty, firstly, I think it would be unwise to make the game "think" any more than it has to. It has been stated before that the game CAN trak damage dealt by a single player, and even the amount of damage repaired by a single nanite but having it TRACK that would cause some srious lag because it would have to be tracked and recorded by an external app running on the server and parsing the game. (wish I had the link to the post, but I don't. It was quite a while back)Gappy wrote:QUOTE (Gappy @ Nov 22 2006, 09:55 PM) Heavy constructor kills: 70
Light constructor kills:40
op/tp (non giga) base kills: 30
op/tp (giga) base kills: 25
Refineries: 15
Normal Tech bases: 60
Adv Tech Bases: 90
Miner Kills: 40 + he3/5 currently on miner
Adv tech base captures: 105
Normal tech base captures: 75
Light base captures (non-refineries): 40
Giga spec ref captures: 30
Dregh He3 mine captures: 15
Kills in sectors with enemy bases: 2
Kills in sectors with no bases: 6
Kills in sectors with friendly bases: 7 (enemy base rule takes precedence)
Killing enemies which have a living miner on their assist list: 20
Killing enemies which have a living con on their assist list: 20
Killing bombers: 25
Killing stealth bombers: 18
Killing HTTs: 30
Dropping a probe that first eyes a miner: 10
Dropping a probe that first eyes a constructor: 10
Dropping a probe that first eyes a bomber: 15
Keep in mind that it's possible for a team to get (up to) 2x the any number listed above divided amongst everyone who worked to accomplish that goal.
Lastly: Winners get their points doubled like under the current system.
Secondly, along the same lines as the old argument that "any point system can be baked", I would propose that any point system that is too specific can be baked, boiled and fried much more easily. Keeping things simple may not reflect worth with 100% accuracy, but making a system too complex with wide point differentials between such similar actions would hinder the system as a whole.
Third, there doesn't appear to be a range in which all of these points should be confined. Even with miniscule divisors, after a year, people could likely be saying that their rank is 264,196. That's quite large, no? An individual point system based on deviation from a midpoint would be much more usefule for guaging players against other players. This is the way that elo does it.
Now, as a parting note I would like to say that I am NOT a fanboy of elo. I'm not particularly fond of the system, but it is what it is, and it's better than nothing. I would like to see a system similar to what you are proposing so that we could compare it with elo and everyone could make informed decisions, but there are definitely scaling problems with an individual point system based from 0 and ending at no set point.
Hope some of this helps with your concept, as it is actually quite intriguing to me.
Why do campers taking on a vital role in many Exp games get punished compared to a scout podding a newbie in open space? A big problem with a points system is that it will never be able to tell whether an action was useful to the team or not (eg. Camping is useful, pointless whoring isn't, no stats system will ever be able to tell the difference).Gappy wrote:QUOTE (Gappy @ Nov 23 2006, 02:55 AM) Kills in sectors with enemy bases: 2
Kills in sectors with no bases: 6
Kills in sectors with friendly bases: 7 (enemy base rule takes precedence)
I would also worry about the increased server load that your system would involve. Tracking all these events in a 50 v 50 game is going to be a big task.
How do people lose points? As I read your system you accumulate points with each game you play. Surely then someone who plays a lot is going to accumulate a lot of points even if they basically suck whereas an uber vet whore who plays 1 SG a week will be ranked very low.
Finally a few suggestions;
Shouldn't IC heavy bases gain more points like Giga Lt ones gain less?
Shouldn't killing Caps give points like the other base killing ships? Of course, you'll need to go through the shipyard structure for each and every core and try to balance the various different SY implementations against each other.
Maybe probes that give subsequent eye to bbrs, cons etc should get points too (half the first spotting value maybe?) to encourage good probing across the map rather than probes around the enemy home.
TP1/2 drops should probably give points.
Should Enh/Ad Miners be worth more points for a kill given that they're harder to kill? Same goes for IC and TF miners, both of which are harder to kill than normal miners (except on PC2, so you'll need an exception there too).
How about carriers? Why aren't they worth extra points?
Hope that's enough to be going on with, I'll chime in again if I think of more things that could have point values.
Spidey: Can't think of a reason I'd need to know anything
Lot of good points above. I like the half-cost spotting, but only once per sector, to prevent a bomber from just disappearing and reappearing repeatedly to give a confederate points.Raveen wrote:QUOTE (Raveen @ Nov 23 2006, 01:24 AM) Why do campers taking on a vital role in many Exp games get punished compared to a scout podding a newbie in open space? A big problem with a points system is that it will never be able to tell whether an action was useful to the team or not (eg. Camping is useful, pointless whoring isn't, no stats system will ever be able to tell the difference).
I would also worry about the increased server load that your system would involve. Tracking all these events in a 50 v 50 game is going to be a big task.
How do people lose points? As I read your system you accumulate points with each game you play. Surely then someone who plays a lot is going to accumulate a lot of points even if they basically suck whereas an uber vet whore who plays 1 SG a week will be ranked very low.
Finally a few suggestions;
Shouldn't IC heavy bases gain more points like Giga Lt ones gain less?
Shouldn't killing Caps give points like the other base killing ships? Of course, you'll need to go through the shipyard structure for each and every core and try to balance the various different SY implementations against each other.
Maybe probes that give subsequent eye to bbrs, cons etc should get points too (half the first spotting value maybe?) to encourage good probing across the map rather than probes around the enemy home.
TP1/2 drops should probably give points.
Should Enh/Ad Miners be worth more points for a kill given that they're harder to kill? Same goes for IC and TF miners, both of which are harder to kill than normal miners (except on PC2, so you'll need an exception there too).
How about carriers? Why aren't they worth extra points?
Hope that's enough to be going on with, I'll chime in again if I think of more things that could have point values.
No special points for tp2 droppers because they receive points from the end base kill (as a nan in the same sector).
On the camping issue: If you're camping a base, it's for a good reason, and usually to kill or cap it. You're not getting many points for the camping because you'll be getting points for the end result of capping/destroying the base. If you're camping the base but aren't doing anything with it, than why should you be rewarded with points?
The rationale behind not differentiating between enh and adv miners, etc is because it's a results based system, and the economic damage is the same regardless of whether it's a normal, enh, or adv miner. IC Ops you may have a point, but largely I don't really see them as really *that* much worse than other bases. Giga has less of a value because you can spam them so easily, and they can die easily against a coordinated team with Dis2 or killer1.
Caps and carriers I'm adding to the table, thanks for pointing them out.
Well lets say you do 30% damage to a 10 point target, and someone else does 70% damage to that target. The total point value from that target would be 10 * (1+(2-1)/(2+1)), or 10 * 4/3 = 13. You would get .3 * 13 = 4 points, while the person who did 70% of the damage would get 9 points, hope that answers it.papsmear wrote:What percentage should be awarded for assists?
Lastly, on the point issue and points 'stacking up'. The points aren't meant to be a direct replacement for ELO. Rather, they're to measure how good a player was in any game, and then derive an overall ranking from that value. So lets say that you start off as a noob, and average very few points in any game, much less than the average player. Because you're below the average, your overall ranking is going to go down. As you get better and better, you'll perform higher on the point scale at the end of games, your overall score at the end of games will get higher and higher in relation to other players (unless you're augz), and thus your rank will improve.
Don't look so much at the absolute value of the points, but rather look at how they relate to each other. For instance, killing a techbase is worth 30x dogfighting a fig in their sector. The way that I see rankings being derived from these point values is such that a standard deviation is calculated between all the scores, and your distance from that standard deviation affects to what degree you can push your rank to. To rise to expert 9, so to speak, you would have to *really* perform in nearly every game you join. The degree to which any game affects your overall ranking is affected linearly according to time spent in any game, and the duration of the game. A game 2x as long will have 2x the impact as a game that was only half as long.
We've upped our standards. Up yours.
Just want to add on as a comment, but the main issue with points is scaling. The more people there are in a game, generally the less valuable killing an enemy is. In a small game, chasing someone down a sector just to pod them could be quite valuable if they only have 5 people on their team.
One possible way to remedy this is to have individual kills worth a variable amount of points based on how many people there currently are on the opposing team. So make kills of an enemy in a sector where they have a base worth, say, 3 for games < 5, 1 for games < 10, and .5 otherwise.
While that may not seem like much, say you've got a bomb run of 10 players Killing an op with 10 people is going to award about 70 points, or 7 points per person. Similarly, killing an adv techbase with 10 people would be worth about 21 points.
The primary goal should be, I think, providing the most points to behavior that results in successfully ending the game. If the point system helps motivate people to do the best thing possible for their team, that's just beneficial side effect.
One possible way to remedy this is to have individual kills worth a variable amount of points based on how many people there currently are on the opposing team. So make kills of an enemy in a sector where they have a base worth, say, 3 for games < 5, 1 for games < 10, and .5 otherwise.
While that may not seem like much, say you've got a bomb run of 10 players Killing an op with 10 people is going to award about 70 points, or 7 points per person. Similarly, killing an adv techbase with 10 people would be worth about 21 points.
The primary goal should be, I think, providing the most points to behavior that results in successfully ending the game. If the point system helps motivate people to do the best thing possible for their team, that's just beneficial side effect.
We've upped our standards. Up yours.
Now that's a tricky question. Following the same lines as bombers, one solution is to do distribute points sector-wide when a constructor starts its build animation. A number off the top of my head would be a static 7 points to any nan in the same sector at build time, 7 being what I figure is about the average that a fig escorting the constructor would get in kills.smapdy wrote:QUOTE (smapdy @ Nov 24 2006, 09:10 PM) no points allocated for nans on cons. what about adding a point value to planting a friendly con? distribute points the same way as bomber run, where 25% (or maybe higher) penalty applied to ships not capable of mounting nanite.
One thing that's clear about the changes I've proposed is that there have to be some code modifications, and I certainly know this and knew it when I proposed it. However, pretty much everything should have a data structure already in the game (such as assist lists). Some major parts missing are conversion of assist lists, and hooks into ICE so that core developers can modify the point values (some of which are currently available). Since we're going to an XML format for core definitions from the previous .igc format, implementation would best wait until after that's complete so that code doesn't have to be fixed in ICE in order to make this all workable too._SRM_PsycHosis wrote:QUOTE (_SRM_PsycHosis @ Nov 24 2006, 11:37 PM) i love your system, it sounds like a good idea. but the drain on the servers.. would be deathdealing in large games.
remember how the aniversary games brought down servers? im seeing that in 20v20 games using this system... is no good unless it can be SAFELY implimented.
We've upped our standards. Up yours.