Changes to the points and rating system

[Daggerheart]

The rating system

• The major thing I don't like about CC is the fact you loose so much more ratingpoints vs a lower ranked player than a player at your own "level".
  In one of my last games I faced an opponent with over 6000 games played and a win rate of 43%, and the guy was currently only private 1 class (!)
  I am/was ranked major and have only played about 1000 games... in that 1vs1 battle we had a even chance of winning, but I would loose a hell more points than him!
  
  I understand that CC aim to capture some of the elements of chess and rating system there... but this is a dice game...

Specifics/Details:

• I suggest a change to the way rank and rating points are calculated. I personally think that it's your win% that should be compared when calculating ratingpoints lost/won after a battle

How this will benefit the site and/or other comments:

• The change would feel more just than the current system

[TheForgivenOne]

Take into new accounts though. They win, say, 5 of 8 games, and the have a huge win percentage. And they verse someone who has played a ton more games, but only has 30%. You're saying that the new player should lose more, just because he has less games played? and has a higher win percentage?

[jrh_cardinal]

even more, take into account 1v1 (or team) players against 8 man singles players. A totally average 1v1 or team player would have a 50% win rate. A totally average 8-man player would have a 12.5% win rate. Definitely not fair to use win rate to compare those two players.


of course CC's point system has flaws, but there is no perfect system. Win rate doesn't work because people can have totally different win rates and still be the same skill level. As you say the current system isn't perfect because of the huge element of luck in the games. There is no perfect way, you're just going to have to accept it how it is.

[Daggerheart]

Yes, I frankly say that. Win % is one factor most related to your skill level and would be the best, maybe not perfect, to decide how many points you get.

Of course you could set up a condition that one needs to play, let say 20 games, before this kicks in...

The other solutions would be to have a combination of win% and title... or having a fixed number of points everybody could lose/win from each player on the board... this solutions would also be better than the current system....

In this game number of games does not relate directly to how good your are... my opinion... a good strategy player has a high win %.... if that is his/her goal...

[PoohsTuger]

I think a fixed point system whould be much better, and the most fair.
Lets say you lose 20 points if you don't win. The points x plyers in the game goes to the winner.
If we are gonna base the points on winrate, the players who enjoy 8 player games
will be "punished" for playing the setting they enjoy most.
We all have diffrent game prefrences, and the points system must be fair to all.

[TheForgivenOne]

I do not approve of everyone losing a set amount of points. Because it would make is so much easier for higher ranks to take more points from low ranks. If everyone lost 20 points per game, then this would give more incentive for higher ranks to take risks at playing lower ranks.

And how does Win percentage compare against points? There are a few players above 3000 points that have about a 30%. This would make it far easier for them to climb the board, compared to a player like King Herpes, who has an 80% win percentage. Ballenus and NUKE both have 31% win percentage, YET they are above 3500. Does this make them worse than Big Whiskey? Who has only 2500 points and a 70% win percentage? There are more holes with your suggestion then there are with this one.

[SirSebstar]

i thought this would be about the rating system instead of the point awarded system. can you change the title?

[Daggerheart]

Sorry if the title was abit off the mark, too late to change completly now... but the post affects the rating system too.

Yes, I agree, fixed points are maybe not as good as the current system...

Win rate should not be the main thing to decide the placement on the Scoreboard... but there could be a seperate scoreboard for winrate...

My point is that players that enjoy 8 players games often seem to don't care about their rating? why was my "6000 games" opponent so low?? Is it fair I should risk so much more points than him??

To get high on the scoreboards you need to win alot more than you lose... we all know that... so win% is the best way to calculate the points you get after a win... at least it should be half the factor of the points won...

Correct me if I am wrong: Many "good" players they "don't" care about their rating until they got all the medals and stuff (or they just try everything and having "wild" fun for some time), THEN they start to be very picky with their opponents and type of game so they win as much as possible against as big as possible opponents... losing as little as they can if they are unlucky and lose... thats why some of the guys/girls in the top have such low winrate...

A way to fix this is to use the winrate from the last 25% of your games when calculating points or something like that... like to hear opinions on this...

[TheForgivenOne]

Not with the 2 people i mentioned. They are both Generals, and i have a higher Win Percentage then them. Heck, the Conqueror doesn't even have a 50% win percentage. On the first page alone, there are well over 60 people that are below 50% win percentages. I've seen a few that are at 25%. Win Rate does not mean you are better. There are even players who are at Sargeant first class, who used to be high Generals.

[Dako]

Yeah, consider my old suggestion about points :)

viewtopic.php?f=4&t=120108

[Woodruff]

Yes, I frankly say that. Win % is one factor most related to your skill level and would be the best, maybe not perfect, to decide how many points you get.

Of course you could set up a condition that one needs to play, let say 20 games, before this kicks in...

The other solutions would be to have a combination of win% and title... or having a fixed number of points everybody could lose/win from each player on the board... this solutions would also be better than the current system....

In this game number of games does not relate directly to how good your are... my opinion... a good strategy player has a high win %.... if that is his/her goal...

As has already been pointed out, this doesn't even logically follow given the differences between a 1-vs-1 player and an 8-man-game player.

[jrh_cardinal]

JOHNNYROCKET24 1646 19323 13104 (68%)
Emmdizzle 5103 1114 521 (47%)


this may not be the very best example since anyone who knows JR knows he's really good (plus used to be really high ranked), but the point stands. Emmdizzle is the conqueror, JR is a Lieutenant, JR's win percent is 20% higher, does that mean JR should be risking more points than Emmdizzle? NO. There are different game types. Each one has a different win probability.

1 Emmdizzle 5103 1114 521 (47%) Conqueror Conqueror 25 4.8 United States United States
2 Blitzaholic 5053 12364 8877 (72%) Field Marshal Field Marshal 101 4.8 United States United States
3 rabbiton 4817 1462 982 (67%) Field Marshal Field Marshal 19 4.7 United States United States
4 mhennigan 4667 421 201 (48%) Field Marshal Field Marshal 8 4.8 United Kingdom United Kingdom
5 Velvecarrots 4420 1241 821 (66%) General General 32 5.0 United States United States
6 David Hoekstra 4126 1319 1030 (78%) General General 19 4.7 United States United States
7 chephren 4040 852 481 (56%) General General 19 4.9 Belgium Belgium
8 laddida 4019 2718 1834 (67%) General General 21 4.9 United States United States
9 Kiron 4012 329 155 (47%) General General 7 4.6 Canada Canada
10 thebest712 3956 3460 2255 (65%) General General 28 4.7 Belgium Belgium


these are the top 10 ranked players on CC right now. Three of them are below 50% win percentage simply because of the game types they play.Obviously it has nothing to do with skill since they are in the top 10 out of basically 20,000 people, but they just play different types of games normally. You can not determine someone's skill level by looking at win percent without knowing the type of games they normally play.

[Arama86n]

a quick reply; win ratio would be very unfair. example:

If I played only 1v1 I'd have a 55% win ratio
If I played only 8player standard games I'd be lucky to have a 15% win ratio.

get my drift?

[Daggerheart]

I will repeat this:

In one of my last games I faced an opponent with over 6000 games played and a win rate of 43%, and the guy was currently only private 1 class (!)
I am/was ranked major and have only played about 1000 games (and have about the same win%)... in that 1vs1 battle we had a even chance of winning, but I would loose a hell more ratingpoints than him!

So how could this "problem" be solved then? It's unfair as hell...

People are making games with only invites now to bypass this unfairness... we all know it...

At least it should be possible to set a restriction (optional) for the game you make, so only a players at your own current level or higher could join it.

[jrh_cardinal]

I will repeat this:

A perfectly average player will win 50% of 1v1's
A perfectly average player will win 12.5% of 8-player games

therefore, if you have two players, we'll say one's a major, one's a private first class, and the major plays almost all 8 man games, he would have a win% of maybe 20%. The private first class plays almost all 1v1's,s o he would have a win% of 45%.

Now, is win % a good representation of skill? You are stuck on this one thing, and if you listen to anyone else who has posted, you will realize that your viewpoint on win % is inaccurate. Yes, the point system is unfair, especially in 1v1's, but there is no way to correct it.

Let's look at the options, shall we:
1 keep it how it is. Okay, if a major is playing a private in a 1v1, he needs to win about 9 out of 10 of them in order to maintain points. That is unrealistic and unfair, but people manage to get well past major (maybe by not playing 1v1's )

2 change it to win% or include win%. As previously stated, this is not plausible with the multitude of game types and number of players/teams in each game. Win% can vary greatly depending on preferred game type much more than ability level. People should not be punished just because they like 1v1 or team games, so that won't work.

3 make it a flat point system (get 20 points for a win, 20 for a loss). Think about it. If a player has 5000 points today, and beats 75% of his opponents, someone who joins the site today will have 0 chance of catching up to him given that they play the same amount of games. It should make sense if you think about it for a bit.
Therefore, a flat point system won't work as it is not fair to punish people for not joining in '06.

What other options are there?

[Daggerheart]

You should have read my last post... I gave up on the win% thing...

[jrh_cardinal]

I will repeat this:

In one of my last games I faced an opponent with over 6000 games played and a win rate of 43%, and the guy was currently only private 1 class (!)
I am/was ranked major and have only played about 1000 games (and have about the same win%)... in that 1vs1 battle we had a even chance of winning, but I would loose a hell more ratingpoints than him!

So how could this "problem" be solved then? It's unfair as hell...

People are making games with only invites now to bypass this unfairness... we all know it...

At least it should be possible to set a restriction (optional) for the game you make, so only a players at your own current level or higher could join it.

ok, I guess you did. I gave up after the first line assuming you were saying the exact same thing again

I will now direct you to this suggestion, but I promise that one will not go through either, the idea has been rejected multiple times before, TheForgivenOne explained it well in the last post on that thread (so far)

[Dako]

Ok, since you ignored my link, I will quote myself here.

Hello there.

Got 2 thoughts on the current CC point system. To be honest, I don't see a problem with current system, but here is 2 suggestions for you to discuss. Please open a ranking table in another window so it will be easier for you to understand what I mean here.

Number 1. Do not award points for the win when the difference between players rank is more than a certain number.
a. for singles it can be 5 levels (so brig will get points only from Sergeant 1st Class and higher)
b. for team games it can be a size of team * 5 (so for dubs the difference between sum of ranks cannot be more than 10).

This will prevent farming because you will get 0 points for the win and will loose 0 as well.

Number 2.
Change the award system based on new criteria - level of the challenge (similar to MMORPG games like WoW).
Level 1 - difference in rank is 1-3.
Level 2 - difference is 4-5
Level 3 - 6
Level 4 - 7
Level 5 - 8
Level 6 - 9+

And when the game finishes you transfer points according to the level of the challenge.
Level 1 = 20 points
Level 2 = 16 points
Level 3 = 12 points
Level 4 = 8 points
Level 5 = 4 points
Level 6 = 0 points

That means a cook will not get a ton for winning a general because it doesn't make sense. Everyone can win due to luck. But if you are lucky against general - doesn't mean you need to be awarded with points (thou advancing you in rank you haven't deserved). And for general - it makes no sense to win over cooks again and again - that will not grant you new skills or anything.

And to advance in ranks you need to play opponents of your level all the time (or be very smart and selective in farming).

Of course, all the numbers can be changed to suit CC needs - this numbers are just an example. Also, it will be easier to "teach" new members - no one will get points for the win and you can be safe about keeping your score.

Both concepts will nearly eliminate low-ranked players farming or make it extremely hard to control.

So what do you think about that? Please vote and comment.
Thanks.

[Jatekos]

That is a great idea, Dako.
I think option 2 along with changing what win % means (should be based on players defeated instead of games won) would be a great step towards a more skill-based approach and calculating system.

Here is the link regarding the suggestion about changing the win %:
http://www.conquerclub.com/forum/viewtopic.php?f=4&t=118277

[pmchugh]

I don't see why anyone didn't just say "true win percentage". I can't remember the specifics of how to work it out, i think obx made a thread about it a while back, meh.

Anyway I don't think that any of the suggestions here are better than what we currently have, though there is definate room for improvement as far as the scoreboard goes.

[Daggerheart]

Dako, I agree that the rating system and points awards MUST be revised, and your toughts are interesting.

The way the system works now is the greatest flaw of CC, and keeps a good site from getting fantastic...

[Woodruff]

Ok, since you ignored my link, I will quote myself here.

Hello there.

Got 2 thoughts on the current CC point system. To be honest, I don't see a problem with current system, but here is 2 suggestions for you to discuss. Please open a ranking table in another window so it will be easier for you to understand what I mean here.

Number 1. Do not award points for the win when the difference between players rank is more than a certain number.
a. for singles it can be 5 levels (so brig will get points only from Sergeant 1st Class and higher)
b. for team games it can be a size of team * 5 (so for dubs the difference between sum of ranks cannot be more than 10).

This will prevent farming because you will get 0 points for the win and will loose 0 as well.

Number 2.
Change the award system based on new criteria - level of the challenge (similar to MMORPG games like WoW).
Level 1 - difference in rank is 1-3.
Level 2 - difference is 4-5
Level 3 - 6
Level 4 - 7
Level 5 - 8
Level 6 - 9+

And when the game finishes you transfer points according to the level of the challenge.
Level 1 = 20 points
Level 2 = 16 points
Level 3 = 12 points
Level 4 = 8 points
Level 5 = 4 points
Level 6 = 0 points

That means a cook will not get a ton for winning a general because it doesn't make sense. Everyone can win due to luck. But if you are lucky against general - doesn't mean you need to be awarded with points (thou advancing you in rank you haven't deserved). And for general - it makes no sense to win over cooks again and again - that will not grant you new skills or anything.

And to advance in ranks you need to play opponents of your level all the time (or be very smart and selective in farming).

Of course, all the numbers can be changed to suit CC needs - this numbers are just an example. Also, it will be easier to "teach" new members - no one will get points for the win and you can be safe about keeping your score.

Both concepts will nearly eliminate low-ranked players farming or make it extremely hard to control.

So what do you think about that? Please vote and comment.
Thanks.

I don't care for option #2, but I LOVE option #1.

[jrh_cardinal]

I don't care for option #2, but I LOVE option #1.

sp you think people should be punished for being good?

Option #1: SuperColonel plays NewbieCook in a 1v1.
SuperColonel victory: 0 points
NewbieCook victory: 100 points

fair? didn't think so




Revised Option #1: same situation
SuperColonel victory: 0 points
NewbieCook victory: 0 points

SuperColonel: "Why am I wasting my time playing some newb when I can't improve my rank (since that is all I care about). I'm going to start playing only private games and 2000+ tourneys"

---THEN----

SuperNewbieNewRecruit joins CC: "Hey, this rating system is cool. I want to play against good players so I can learn how to play well, then I can get good and have an awesome crown like them."
...
...
...
"Why are there no awesome people with crowns in games?"

Oh, that's right. Crown people want points so no longer play in public games at all







The "revised" option also punishes people like mhennigan, Thai Robert, Kiron, and rabbiton. Super-high ranked freemies who are hard-pressed to play in private games with people close to their rank.

[Woodruff]

I don't care for option #2, but I LOVE option #1.

sp you think people should be punished for being good?

No...if you go to his original suggestion thread, you'll see that I have amended it to read as I thought it did (I mis-read it). Which I love.

[pie12345]

i like it