Latest Posts

Topic: Rating system

einstein13
Avatar
Joined: 2013-07-29, 00:01
Posts: 1118
Ranking
One Elder of Players
Location: Poland
Posted at: 2019-07-28, 23:18

WorldSavior wrote:

Also on very imbalanced maps?

It depends on rules that will be applied to the ranking. But if it will be allowed, why not?
I expect an answer from you "that seems to be unfair to me", so I will answer like that:

If you are experienced player, you are expected to know some maps and you know that they can be imbalanced. So you should not accept (ranking) games where you don't know the map or you know that the map is imbalanced for your loss. So the rank points are covering how experienced the player is with interactions with the others.

einstein13 wrote:

Another economy tournament? face-smile.png

Kind of, but permanently. For example on Ice Wars, blue position.

I think that it would fail after some time (I would try only once, at most twice), so after let's say 2 months time, the infinite tournament would fail.

But there is a solution: why not making "challenges" every quarter? And to sum the results from all time? Then the ranking will be alive for a long time.


PS.: My previous message (very technical one) is on the previous page.

Edited: 2019-07-28, 23:22

einstein13
calculations & maps packages: http://wuatek.no-ip.org/~rak/widelands/
backup website files: http://kartezjusz.ddns.net/upload/widelands/

Top Quote
kaputtnik
Avatar
Joined: 2013-02-18, 20:48
Posts: 2433
OS: Archlinux
Version: current master
Ranking
One Elder of Players
Location: Germany
Posted at: 2019-07-29, 20:08

trimard wrote:

overall digits of 5, three on the left and two on the right of the separator. Trying to apply a value of 1500 does not work, because the left part (integers) can only have 3 digits.

THANKSSSSS, that was it. Don't know how I didn't see that!

Don't mention it. I do also struggle with every new things face-wink.png


Fight simulator for Widelands:
https://wide-fighter.netlify.app/

Top Quote
WorldSavior
Avatar
Joined: 2016-10-15, 04:10
Posts: 2091
OS: Linux
Version: Recent tournament version
Ranking
One Elder of Players
Location: Germany
Posted at: 2019-07-29, 21:23

einstein13 wrote:

WorldSavior wrote:

Also on very imbalanced maps?

It depends on rules that will be applied to the ranking. But if it will be allowed, why not?

Because playing on imbalanced maps can be equal to manipulate the ratings. So it shouldn't be allowed.

I expect an answer from you "that seems to be unfair to me", so I will answer like that:

If you are experienced player, you are expected to know some maps and you know that they can be imbalanced. So you should not accept (ranking) games where you don't know the map or you know that the map is imbalanced for your loss. So the rank points are covering how experienced the player is with interactions with the others.

Rating numbers should just express playing strength, not interaction cleverness.

einstein13 wrote:

Another economy tournament? face-smile.png

Kind of, but permanently. For example on Ice Wars, blue position.

I think that it would fail after some time (I would try only once, at most twice), so after let's say 2 months time, the infinite tournament would fail.

It wouldn't be a tournament, rather a high score list. So no failure.

But there is a solution: why not making "challenges" every quarter? And to sum the results from all time? Then the ranking will be alive for a long time.

  • Summing the results up is complicated, because different maps have different potentials for points

  • Challenges every quarter could be too much

Edited: 2019-07-29, 21:23

Wanted to save the world, then I got widetracked

Top Quote
einstein13
Avatar
Joined: 2013-07-29, 00:01
Posts: 1118
Ranking
One Elder of Players
Location: Poland
Posted at: 2019-07-29, 23:31

WorldSavior wrote:

einstein13 wrote:

WorldSavior wrote:

Also on very imbalanced maps?

It depends on rules that will be applied to the ranking. But if it will be allowed, why not?

Because playing on imbalanced maps can be equal to manipulate the ratings. So it shouldn't be allowed.

Fair point. So there should be a list of maps that are "officially balanced". But how is it possible to make the list easy to fill with new maps? How should we decide which map is equal enough?


einstein13
calculations & maps packages: http://wuatek.no-ip.org/~rak/widelands/
backup website files: http://kartezjusz.ddns.net/upload/widelands/

Top Quote
trimard
Avatar
Topic Opener
Joined: 2009-03-05, 22:40
Posts: 230
Ranking
Widelands-Forum-Junkie
Location: Paris
Posted at: 2019-07-30, 18:56

Is my idea clearer now? face-smile.png Hope so face-tongue.png

Yes definitely clearer! It's a good idea! But TBH I can't exactly understand the ramifications and limits of your system until I try implementing it. So I'll comment on your whole post when relevant face-wink.png Kind of like your calculations for many players.

I know that it is completely different than yours. face-grin.png Not really. My idea is really to look for functionality first, hence why my models were different. But I intend to move closer to yours bit by bit!

he simpliest idea for rank I know is to find proportion of win/all games. And since "rating" is providing decimals, the (almost) most precise equation for rank value would be:

Huh, why not? But I don't see really the point if glicko has been proved to be effective? Maybe it could be a nice feature to add first. So I would focus on that measure first and implement the whole interface for players before working on glicko2. It seems like little work if I use your models first.

but since Widelands is a doocracy and since you are developing the tools, you make a choice face-wink.png

Then I'll see how I choose to implement it, when I'm on it. I think you have the best global design in mind. So I'll be heavily influenced by your models.

Thanks a lot anyway, you removed me a lot of bothersome problems!

But how is it possible to make the list easy to fill with new maps? How should we decide which map is equal enough?

It'll never be exact. But I don't think any existing rts game (or any game more complicated than grids) that can pretend to say differently.

BUT we can vote in the community for the map (IIRRC there is a tool for votes on the website). And we change the maps every seasons for 3 reasons:

  • it might appear maps thoughts to be balanced, after more games turn out to be very unbalanced.

  • We can add new maps

  • It allow for diversity for regular ranked game players.

why not making "challenges" every quarter? And to sum the results from all time? Then the ranking will be alive for a long time.

Why make a meta ranking? Already a ranking for every quarter is neat. And it will keep the ranks alive for longer


Top Quote
kaputtnik
Avatar
Joined: 2013-02-18, 20:48
Posts: 2433
OS: Archlinux
Version: current master
Ranking
One Elder of Players
Location: Germany
Posted at: 2019-07-30, 19:36

einstein13 wrote:

WorldSavior wrote:

einstein13 wrote:

WorldSavior wrote:

Also on very imbalanced maps?

It depends on rules that will be applied to the ranking. But if it will be allowed, why not?

Because playing on imbalanced maps can be equal to manipulate the ratings. So it shouldn't be allowed.

Fair point. So there should be a list of maps that are "officially balanced". But how is it possible to make the list easy to fill with new maps? How should we decide which map is equal enough?

I guess most players take a look at the map which is wanted to be played, and maybe some make testgames before the final game (the game which would count for ranking) starts. So i think it is unlikely that a map is chosen which will be unfair for one player.

Anyway having a tag 'officially balanced' would be nice.


Fight simulator for Widelands:
https://wide-fighter.netlify.app/

Top Quote
einstein13
Avatar
Joined: 2013-07-29, 00:01
Posts: 1118
Ranking
One Elder of Players
Location: Poland
Posted at: 2019-07-30, 20:49

trimard wrote:

he simpliest idea for rank I know is to find proportion of win/all games. And since "rating" is providing decimals, the (almost) most precise equation for rank value would be:

Huh, why not? But I don't see really the point if glicko has been proved to be effective?

The idea is to provide more statistics than only one. Just because people like to see that their positions are going up somewhere.

Another thing is that if you don't know your opponent, you can estimate his/her power by seeing many statistics. For example if a player has huge numbers in win/lose games, it means that he/she is a common opponent. But if glicko is low, it means that the player is playing with a rather small group of people.

Another thing that I was thinking about as simplest as possible rank, just not to add lost of work on that.

kaputtnik wrote

Anyway having a tag 'officially balanced' would be nice.

I like this idea! That would solve WorldSavior's problem and add new possibilities face-smile.png

There can be "official rank" that would be very strict in rules (f.e. balanced maps; 1 vs 1 or 2 vs 2 games only) and second rank that would contain all possible cases.


einstein13
calculations & maps packages: http://wuatek.no-ip.org/~rak/widelands/
backup website files: http://kartezjusz.ddns.net/upload/widelands/

Top Quote
WorldSavior
Avatar
Joined: 2016-10-15, 04:10
Posts: 2091
OS: Linux
Version: Recent tournament version
Ranking
One Elder of Players
Location: Germany
Posted at: 2019-07-30, 21:02

kaputtnik wrote:

einstein13 wrote:

WorldSavior wrote:

einstein13 wrote:

WorldSavior wrote:

Also on very imbalanced maps?

It depends on rules that will be applied to the ranking. But if it will be allowed, why not?

Because playing on imbalanced maps can be equal to manipulate the ratings. So it shouldn't be allowed.

Fair point. So there should be a list of maps that are "officially balanced". But how is it possible to make the list easy to fill with new maps? How should we decide which map is equal enough?

I guess most players take a look at the map which is wanted to be played, and maybe some make testgames before the final game (the game which would count for ranking) starts. So i think it is unlikely that a map is chosen which will be unfair for one player.

Yes, but this doesn't solve the problem that imbalanced maps could be abused to manipulate the rating.

Anyway having a tag 'officially balanced' would be nice.

Yes

einstein13 wrote:

WorldSavior wrote:

einstein13 wrote:

WorldSavior wrote:

Also on very imbalanced maps?

It depends on rules that will be applied to the ranking. But if it will be allowed, why not?

Because playing on imbalanced maps can be equal to manipulate the ratings. So it shouldn't be allowed.

Fair point. So there should be a list of maps that are "officially balanced". But how is it possible to make the list easy to fill with new maps? How should we decide which map is equal enough?

Good questions. First of all, periodic or symmetric maps could be balanced. But then its difficult... It's also important which colors are picked. Blue vs white on The Nile is unfair, for example, while blue vs red looks more or less fair, and white has also a counterpart which grants fairness. At the other hand, there might be "imbalanced" maps which can still be used for balanced matches. For example, a match of green vs red on Wisent Valley doesn't look very imbalanced to me.


Wanted to save the world, then I got widetracked

Top Quote
WorldSavior
Avatar
Joined: 2016-10-15, 04:10
Posts: 2091
OS: Linux
Version: Recent tournament version
Ranking
One Elder of Players
Location: Germany
Posted at: 2019-07-31, 17:31

einstein13 wrote:

The simpliest idea for rank I know is to find proportion of win/all games. And since "rating" is providing decimals, the (almost) most precise equation for rank value would be:

ranking = 1000.0 * number_of_win_games / (number_of_win_games + number_of_lost_games)

This isn't a meaningful value.


Wanted to save the world, then I got widetracked

Top Quote
einstein13
Avatar
Joined: 2013-07-29, 00:01
Posts: 1118
Ranking
One Elder of Players
Location: Poland
Posted at: 2019-07-31, 18:04

WorldSavior wrote:

einstein13 wrote:

The simpliest idea for rank I know is to find proportion of win/all games. And since "rating" is providing decimals, the (almost) most precise equation for rank value would be:

ranking = 1000.0 * number_of_win_games / (number_of_win_games + number_of_lost_games)

This isn't a meaningful value.

Yes it is.

ranking = number_of_win_games / (number_of_win_games + number_of_lost_games)

is a "fraction of winnings". Multiplying it by a constant doesn't change the meaning.


einstein13
calculations & maps packages: http://wuatek.no-ip.org/~rak/widelands/
backup website files: http://kartezjusz.ddns.net/upload/widelands/

Top Quote