|
Post by Ellestar on Nov 7, 2007 1:42:02 GMT -5
Yes we looked at the TrueSkill system, but unfortunately MS only gives you have the info for free, the real details are not public intellectual property. But I am sure that we can tweak our ante system to ensure that teamers are as fairly treated as ctons and 1v1's, but we also have to ensure that while we create a system that ensures the league is competitive across the board we also don't want to create a huge rift between new players and the top players. We have enough people screaming "elitism" as it is now. If we do this system right it should create a normal bell curve of player skill values over a few month's of reports. CS Well, actually they explain it in pdf but then you need to know probability math good enough so to understand it. More importantly, they hint on the right way to count teamers using ELO-like ranking - skill gain/loss in a teamers should depend on a total team skill (sum of skills of all members). And skill gain/loss from teamers should be lower because a variance in performance of a team is significantly bigger. They say that FFA is the fastest way to roughly measure skill and teamers are the slowest (as expected, more players - slower measurement research.microsoft.com/mlp/apg/trueskill.aspx). Well, they use their system for match-making while we obviously don't. Also, with matchmaking their FFA will have players of a close skill so that makes a difference too. So IMHO for our ladder we should give a maximum weight (skill gain/loss per game) to duels, somewhat lower weight to ctons and lowest to teamers. And use ELO or some variant of it. It will not be as good as TrueSkill (it's ELO variant with a mathematical apparatus to roughly measure a new player's skill much faster), but still better than what you "invented"
|
|
|
Post by MMV on Nov 7, 2007 8:13:39 GMT -5
Since it's clear that the designers of the new ladder are trying to make it equal and balanced for both teamer players and non-teamer players, why not have an option at the beginning of the reporting process that seperates teamer stats from non-teamer stats:
example: =====================
Report Game
log in: xxxxxxxxx
Tearmer: ___ Non-Teamer: ___
(followed by the normal reporting process, etc etc etc)
Would it be that bad or that hard to show who's the seperate stats for both in the appropriate section of the new ladder? i.e. TEAMERS:.........NON-TEAMERS 1. bob..............1. jim 2. billy..............2. frank 3. john.............3. clark
|
|
|
Post by Canucksoldier on Nov 7, 2007 10:04:16 GMT -5
Since it's clear that the designers of the new ladder are trying to make it equal and balanced for both teamer players and non-teamer players, why not have an option at the beginning of the reporting process that seperates teamer stats from non-teamer stats: example: ===================== Report Game log in: xxxxxxxxx Tearmer: ___ Non-Teamer: ___ (followed by the normal reporting process, etc etc etc) Would it be that bad or that hard to show who's the seperate stats for both in the appropriate section of the new ladder? i.e. TEAMERS:.........NON-TEAMERS 1. bob..............1. jim 2. billy..............2. frank 3. john.............3. clark I don't think we need to separate game types at the front end, we will have lots of stats that can be displayed, including specific game type stats for each player and overall stats. We are going to operate under the concept that we design a system that ranks all players skill together on the same list. How we fairly rate skill from different games types, to do this, is what we are debating now. If you want to see that player X has 85% of there games played as teamers and there skill makes them the 8th ranked player, that will be available as a statistic. And to create a separate list of skill from a specific game type should not be hard, as when you report your wins, you have to enter the type of game as step 1, so all that info is in the player database. CS
|
|
|
Post by MMV on Nov 7, 2007 13:24:19 GMT -5
That's ok - I was just posting something else that Ellestar could flame
|
|
|
Post by share on Nov 12, 2007 15:15:16 GMT -5
more screnne more screnne
|
|
|
Post by Death to ALL on Nov 12, 2007 16:00:50 GMT -5
Sorry no screen shots atm. We havn't gotten much done in the last week with the CCC. We will be getting back to work on it this week and I'll post some more screen shots when we get some more done.
|
|
|
Post by deyreepher on Nov 13, 2007 1:03:25 GMT -5
Since it's clear that the designers of the new ladder are trying to make it equal and balanced for both teamer players and non-teamer players, why not have an option at the beginning of the reporting process that seperates teamer stats from non-teamer stats: example: ===================== Report Game log in: xxxxxxxxx Tearmer: ___ Non-Teamer: ___ (followed by the normal reporting process, etc etc etc) Would it be that bad or that hard to show who's the seperate stats for both in the appropriate section of the new ladder? i.e. TEAMERS:.........NON-TEAMERS 1. bob..............1. jim 2. billy..............2. frank 3. john.............3. clark I don't think we need to separate game types at the front end, we will have lots of stats that can be displayed, including specific game type stats for each player and overall stats. We are going to operate under the concept that we design a system that ranks all players skill together on the same list. How we fairly rate skill from different games types, to do this, is what we are debating now. If you want to see that player X has 85% of there games played as teamers and there skill makes them the 8th ranked player, that will be available as a statistic. And to create a separate list of skill from a specific game type should not be hard, as when you report your wins, you have to enter the type of game as step 1, so all that info is in the player database. CS It's not about what you think, it's about what we want. Put up a poll, I'm sure there are plenty of cton players who would like to have a separate rating. I don't have an opinion on the matter, yet. However, it looks like Ellestar has looked into this and would be your man in implementing this if this is what the ladder would like.
|
|
|
Post by geforced on Nov 13, 2007 2:38:25 GMT -5
well ive stopped development on my ladder now, but ill tell you the main problem is these random team games, realisticly team games should be kept to clan matches like pretty much every other game.
even when somone is ranked on this basis it will mean absolutely nothing
i had the idea of a proper point system, which would end at the last night of the month, winners declared and be reset, pretty similar to the old c'n'c ladder, ive lost the formula ill try to dig it out somtime
|
|
|
Post by Canucksoldier on Nov 13, 2007 12:09:25 GMT -5
I don't think we need to separate game types at the front end, we will have lots of stats that can be displayed, including specific game type stats for each player and overall stats. We are going to operate under the concept that we design a system that ranks all players skill together on the same list. How we fairly rate skill from different games types, to do this, is what we are debating now. If you want to see that player X has 85% of there games played as teamers and there skill makes them the 8th ranked player, that will be available as a statistic. And to create a separate list of skill from a specific game type should not be hard, as when you report your wins, you have to enter the type of game as step 1, so all that info is in the player database. CS It's not about what you think, it's about what we want. Put up a poll, I'm sure there are plenty of cton players who would like to have a separate rating. I don't have an opinion on the matter, yet. However, it looks like Ellestar has looked into this and would be your man in implementing this if this is what the ladder would like. Well I'm sure that the development team will get lots of feedback when we reach the stage of a large beta test. Until then what "we think" is based on the entire Admin/TD team's experience, and the comments in this thread. And I'm sure we have a pretty complete system ready now. I'm sure that there will be lots of improvements to follow as we progress threw the testing, but it's not like there is not years of experience present in the initial development team. CS
|
|
|
Post by deyreepher on Nov 13, 2007 12:17:40 GMT -5
well ive stopped development on my ladder now, but ill tell you the main problem is these random team games, realisticly team games should be kept to clan matches like pretty much every other game. even when somone is ranked on this basis it will mean absolutely nothing i had the idea of a proper point system, which would end at the last night of the month, winners declared and be reset, pretty similar to the old c'n'c ladder, ive lost the formula ill try to dig it out somtime That is a good idea, it is good that you brought it up now. I think Geforced has hit a good point. The inclusion of clan matches could slowly lead to the implementation of league style play in Civ IV. Though pick-up teamers are nice in a pinch, playing against another established team would greatly benefit the level of play found on the ladder. The best way to promote the CCC is to have competition. The 6 week break between CCCs would be a great time to be able to ramp up the competition. Also, this problem of having ineligible players due to not knowing they are on a team, mis-spelled user names, and other various reasons could potentially be offset by having standing clan roster support with this new ladder system. The admins already have to deal with tracking user names, I'm sure it would not take that much more effort to track which clans are in existence. As this CCC, there is the issue of fake clans. However, having a pre-requisite of a set number of clan matches to qualify for the CCC would be a possible solution to this potential problem. For one, it ensures that a team has the organization to get a team together to participate in a match. Second, it introduces them to some sort of competition prior to the CCC if they are a new clan. Again, I will point out that others do see the Case's Ladder System as inadequate. Apparently, Geforced was another resource you could have drawn upon for the creation of a new ladder. Maybe you need a new ladder system advisory board and make sure that it is staffed by people who have the ability and the drive to make these things happen. Multiple heads are usually better than one.
|
|
|
Post by Canucksoldier on Nov 13, 2007 12:23:05 GMT -5
well ive stopped development on my ladder now, but ill tell you the main problem is these random team games, realisticly team games should be kept to clan matches like pretty much every other game. even when somone is ranked on this basis it will mean absolutely nothing i had the idea of a proper point system, which would end at the last night of the month, winners declared and be reset, pretty similar to the old c'n'c ladder, ive lost the formula ill try to dig it out somtime Well it is going to be problematic finding the best solution to wieght team games with cton/ffa's and 1v1 games. But in the end we have to provide a single skill rating to rank every player on the ladder. In addition we could rank players separately on a 1v1, cton and teamer list, and even use different weighting for any of these as required. We have a good starting system now with two differnent systems in place in parallel to do testing on how they will work with live test data. From that we will be able to better make a determination on what system we will use in the long run. CS
|
|
|
Post by deyreepher on Nov 13, 2007 12:27:00 GMT -5
It's not about what you think, it's about what we want. Put up a poll, I'm sure there are plenty of cton players who would like to have a separate rating. I don't have an opinion on the matter, yet. However, it looks like Ellestar has looked into this and would be your man in implementing this if this is what the ladder would like. Well I'm sure that the development team will get lots of feedback when we reach the stage of a large beta test. Until then what "we think" is based on the entire Admin/TD team's experience, and the comments in this thread. And I'm sure we have a pretty complete system ready now. I'm sure that there will be lots of improvements to follow as we progress threw the testing, but it's not like there is not years of experience present in the initial development team. CS You have qualified individuals who can support this project like Ellestar and Geforced. Why would you keep the the circle that makes the decisions on the development of this ladder to the programmer, the admins, and the TDs? What qualifies this closed group from making all the decisions? For one, you said it yourself, you and the admins were happy to continue using the Case's Ladder system, which most of us find grossly inadequate. Second, it seems like Ellestar and, to some extent, Geforced have been researching the algorithms needed to create a system that will encompass the various styles of play that are found within the ladder. The admins have not demonstrated this level of dedication or research into what would help better service our varying demographic of players. Again, I will state that it seems that your team has a limited vision of what would help benefit the ladder. If you are not going to include the before mentioned individuals in the process of developing this new ladder, I do not know what to think other than you are very controlling of this community. It is not like I am asking for you to open the doors up to myself or the players that I play with to making decisions. I am asking you to include players that I have vague relationships with to be included in the developmental process of this new ladder. You stated your team does not have the experience or the know-how to create a system from scratch. That obviously means you could use a few members. Two members have stated they may have sort of skill set to help you out. Why exclude them?
|
|
|
Post by Canucksoldier on Nov 13, 2007 12:57:10 GMT -5
well ive stopped development on my ladder now, but ill tell you the main problem is these random team games, realisticly team games should be kept to clan matches like pretty much every other game. even when somone is ranked on this basis it will mean absolutely nothing i had the idea of a proper point system, which would end at the last night of the month, winners declared and be reset, pretty similar to the old c'n'c ladder, ive lost the formula ill try to dig it out somtime That is a good idea, it is good that you brought it up now. I think Geforced has hit a good point. The inclusion of clan matches could slowly lead to the implementation of league style play in Civ IV. Though pick-up teamers are nice in a pinch, playing against another established team would greatly benefit the level of play found on the ladder. The best way to promote the CCC is to have competition. The 6 week break between CCCs would be a great time to be able to ramp up the competition. Also, this problem of having ineligible players due to not knowing they are on a team, mis-spelled user names, and other various reasons could potentially be offset by having standing clan roster support with this new ladder system. The admins already have to deal with tracking user names, I'm sure it would not take that much more effort to track which clans are in existence. As this CCC, there is the issue of fake clans. However, having a pre-requisite of a set number of clan matches to qualify for the CCC would be a possible solution to this potential problem. For one, it ensures that a team has the organization to get a team together to participate in a match. Second, it introduces them to some sort of competition prior to the CCC if they are a new clan. Again, I will point out that others do see the Case's Ladder System as inadequate. Apparently, Geforced was another resource you could have drawn upon for the creation of a new ladder. Maybe you need a new ladder system advisory board and make sure that it is staffed by people who have the ability and the drive to make these things happen. Multiple heads are usually better than one. We have stated that clan matches will be part of this system, but we have not developed that system very far until we get the basic code for the league system more developed. So we will be taking note of idea's like Geforced, and yours, when we start the clan system. The good thing about all this is that since we have complete control over the code, if we don't like something the first time we can just change it. And I foresee this happening as no one really knows what the best way to do a clan system will be or what features we might want in the future. CS
|
|
|
Post by geforced on Nov 13, 2007 13:57:52 GMT -5
Well it is going to be problematic finding the best solution to wieght team games with cton/ffa's and 1v1 games. CS just lol
|
|
|
Post by Canucksoldier on Nov 13, 2007 13:57:58 GMT -5
Well I'm sure that the development team will get lots of feedback when we reach the stage of a large beta test. Until then what "we think" is based on the entire Admin/TD team's experience, and the comments in this thread. And I'm sure we have a pretty complete system ready now. I'm sure that there will be lots of improvements to follow as we progress threw the testing, but it's not like there is not years of experience present in the initial development team. CS You have qualified individuals who can support this project like Ellestar and Geforced. Why would you keep the the circle that makes the decisions on the development of this ladder to the programmer, the admins, and the TDs? What qualifies this closed group from making all the decisions? For one, you said it yourself, you and the admins were happy to continue using the Case's Ladder system, which most of us find grossly inadequate. Second, it seems like Ellestar and, to some extent, Geforced have been researching the algorithms needed to create a system that will encompass the various styles of play that are found within the ladder. The admins have not demonstrated this level of dedication or research into what would help better service our varying demographic of players. Again, I will state that it seems that your team has a limited vision of what would help benefit the ladder. If you are not going to include the before mentioned individuals in the process of developing this new ladder, I do not know what to think other than you are very controlling of this community. It is not like I am asking for you to open the doors up to myself or the players that I play with to making decisions. I am asking you to include players that I have vague relationships with to be included in the developmental process of this new ladder. You stated your team does not have the experience or the know-how to create a system from scratch. That obviously means you could use a few members. Two members have stated they may have sort of skill set to help you out. Why exclude them? Dey, you continue to put words into mouth that I have not said. I have never said I was "happy" with everything about cases, only that it was the best free option available, and although far from perfect, it gave us the basic functionality that we needed to operate. No one has said that people like Geforced or Ellstar can not be invited, if they so desire. To date I don't beleive that they have indicated that they want to assist us, although it seems that they are working on something with MGT, although I'm not aware of the status of that project other than what Geforced has stated in this thread. But if they want to contribute there expertise it would certainly be welcome. But as to your statement that you believe that we somehow have not researched league systems, I have no idea how you came to that conclusion, as I assure you have have looked at every availabe online system that has been developed recently. Right now we are looking at testing both the ante system we originally designed and a system based on the Glicko system. We will assess both, and more if required in order to make the best system for the league. Were did I state that we don't have the skills to design this ladder? I'm not saying that we don't need or won't take everyones help, but we are far from being uneducated in this area ourselves. You seem to have the impression that we don't want help, that is not the case, we are just not at the point were we should bring in a larger group, we have invited the TD's because we want there feed back on the tournament system, to get it right the first time. If there are members of the community that feel they have something to offer this effort they are free to contact myself or DTA. CS
|
|
|
Post by DrShot on Nov 13, 2007 18:59:49 GMT -5
This discussion reminds me of the wedding in Monty Pythons quest for the Holy Grail: "Lets not argue over who killed who, this is a happy occasion..." For me and a few others that I know, the ladder's best purpose is having created a gaming environment that suits many needs. Priority goes to a common playing field governed by basic rules. A venue that (mostly) assures consistency in gameplay and the fore knowledge that a match will be played out (sans quitters) to the rules set prior to the start of said match. Ranking systems and measures of skills are nice. Obviously this takes precident over every other conceiveable factor with a few players and thats ok. A pissing contest is their greatest form of satisfication, good for them I say. Likely it would seem most players know who is at what level in the community. Breaking down games to a linear scale to denote rank or skill is fruitless. How far should this progress? Single matches: 1v1 cton and the like... How do these compare to Team games. How much emphisis should be designated to a player in a team game? Does a teammember that was a true standout in one particular match deserve an adjusted credit score? Does a team member(winning team lets say) that dies deserve a lessor credit or portion of the 'prize' credits assigned. How about different Era's? Should another Era be worth more than the next? Streaks? Does one receive a progressive amount of skill or rank credits for a continuing streak or conversly for a losing run? These are just a few examples. Where will this end or should it end? Should we split hairs or are broad based actions only required(desired)? Keep it simple.
|
|
|
Post by Canucksoldier on Nov 14, 2007 0:24:28 GMT -5
You are right Shot, we are putting just as much effort into the rest of the details of a good league. But in the end I would prefer to do both, satify the average player with good overall services and interface, and make the ultra-competitive players that want the best rating system we can give them. With a little luck and effort we will make everyone happy.
CS
|
|
|
Post by ordoabchao on Nov 14, 2007 0:27:08 GMT -5
if you are going to have multipe areas to become #1 in (which is good idea) why can't u just have different algorithums for each class of game??? have duels/ctons/ffas/teamers each different in the way they calculate and change your skill or whatever. 4 way weight could be somethin like
35%(duel)/25%(Cton)/25%(FFA)/15%(Teamer) or
30/25/25/20%.
|
|
|
Post by Canucksoldier on Nov 14, 2007 0:31:57 GMT -5
if you are going to have multipe areas to become #1 in (which is good idea) why can't u just have different algorithums for each class of game??? have duels/ctons/ffas/teamers each different in the way they calculate and change your skill or whatever. That is a possibility, but we don't want to make things more complicated than need be until we are sure that it is the best course of action. There is a lot to be said for keeping things simple too. CS
|
|
|
Post by ordoabchao on Nov 14, 2007 0:34:50 GMT -5
shouldn't be too much more...u should already be weighing the various alternatives so u should have the algorithims or whatever...only extra work is making them work with the report screen...thats if its still a common skill pool anyway just with different weights and calculations for the 4 main types of games.
|
|