Ladder Site Online...

Discussion of all aspects of multiplayer development: unit balancing, map development, server development, and so forth.

Moderators: Forum Moderators, Developers

Post Reply
enclave
Posts: 936
Joined: December 15th, 2007, 8:52 am

Post by enclave » February 27th, 2008, 6:21 pm

I attached how the ranking program looks on my computer,

but it was with more options because i wanted to make separate sheet for every player when you click on his nickname.. to show his stats.. etc..
example - http://www.enclave.valsts.lv/zaraz/enclave.html

in your case program may look like second attachment..
with any other options you may need.. i can code both program and both HTML for free if you need. But not PHP.. sorry
Attachments
wesnoth ranks.jpg
wesnoth ranks.jpg (35.28 KiB) Viewed 3014 times
tebasga ranks.jpg
tebasga ranks.jpg (107.27 KiB) Viewed 3016 times

enclave
Posts: 936
Joined: December 15th, 2007, 8:52 am

Post by enclave » February 27th, 2008, 6:23 pm

if you feel like im spamming your topic, tell me, i will edit my posts to look small and empty. thanks

User avatar
eyerouge
Posts: 380
Joined: June 29th, 2007, 4:37 am
Location: wtactics.org
Contact:

Post by eyerouge » February 27th, 2008, 11:57 pm

It is not obligate to register teams.. ladder name could be "best team players" and...
when you play 2v2, when 1 team wins, both winners get + ranks
so if 4 players were "enclave" "noy" zookeeper" "noob" and noob+enclave won vs noy+zookeeper, then ranks show

enclave | 1 pt
noob | 1 pt
Noy | 0 pt (or -1pt depends on what kind of ranking you want)
Zookeeper | 0 pt

/../ if your team wins, you both get ranks.. for good team play..
I agree that your way of doing it (in the above) is viable. It certainly is different from what I imagined you requested, and, it solved the problem with the ever changing team members, but, it also brings a new problem:

What exactly does it measure? It seems to me that it measures a mix of things, and at the same time it measures nothing at all that's interesting and really says something fair about your skills as a player:

In your way of measuring it doesn't matter which players I play with in a 2vs2. If I have 10 points I could have gotten 2 with you as my ally, 3 with another guy, and the rest with my mother.

Now, when somebody looks at my amount of points, what do they tell?

One thing that's primary here is that it isn't a 2vs2 ladder. It's an individual ladder and it doesn't focus on the teams skills as much as it focuses on individual skills. (And yes, I know that the team is the sum of the individuals but there's still a difference in ranking a team and ranking individuals that can change). If that's the case, it's more or less a clone of the already existing 1vs1 ladder, but for a different format (2vs2), and also with the problem that you can never tell how good the individual is since his points also depend on all various partners he's had.

In essence, your suggestion seems to be a 1vs1 ladder but for 2vs2 games. Again, what would the point be, as that ranking is way less accurate than the 1vs1 one, and still ranks the individual in contrast to the team?
if your imagination is different from mine, i dont mind it.. it was just an idea for 2v2 ladder. You the boss, and im just trying to help.
It is different, since I reccon a 2vs2 ladder would compare teams and reward the team instead of the individuals in them.

However, that said, I'm not god and my view on the matter is as subjective as your own. That said, it's pointless having a ladder if we don't really know what we're supposed to measure with it, so the discussion is in place and interesting since it shows some of the problems with evaluating ratings for 2vs2 play.

On a siden note: I initiated the project, have done some code on it and host it, but, I wouldn't want to be seen as a boss of any kind and I've also have had help from others ;)

As a matter of fact my plans were and still are to give over the whole project to the wesnoth community and those willing to run it once it's mature enough and I feel the time is right. Only reason I initiated it was that I lacked the rating function in Wesnoth and that I knew that a) it isn't coming in a long time if ever and b) there wasn't a working/smooth/active ladder around to my knowledge.

So, I'm not here to rule, mainly because of my ideological background. I'm here to try give a tool to the community. Sometimes the community comes with great ideas which can be easily implemented by me even if I'm not really a PHP coder and we still lack one (like Docs for example, with the friends list) and others it comes with less good ideas or ideas that are good but that have to be put on hold due to other priorities with the ladder code (like the save game parser for example). In all cases I see a creative and productive discussion as a great addon to what we already have, so please stop undermining yourself and/or apologizing for speaking out on an open forum - after all, that's why it's here for ;) ..at least my threads in it.
2) small map and big map is like light weight and heavy weight in boxing.. other settings would be already like american boxing and english boxing.. fighting with gloves or without... if you knwo what i mean.. but whatever, im not insisting on that part of my idea.
I'm not a mapmaker or a map expert in any way, so I wouldn't know if it's you or me who has some wild thoughts on the analogies. To me it's logical to treat the size of the map as any other variable of preferences that a player has: Player x prefers to play on map size y, with settings z, and faction 1 etc etc.

To you a small and large map maybe matters much. To somebody else it doesn't (like a skilled player for example? He won't become braindead just because the map is larger). To you size matters (excuse the pun), to another player some other factor matters. You are 100% correct that game play would, if the players are somewhat logical, change depending on map sixe. However, it also changes depending on plenty of other factors. If the map size is a valid argument, so are most of the arguments concerning the other factors. It's all just subjective perceptions of what matters the most to the individual players.

In your case it's clear. I won't deny that you value the map size, but, I wouldn't see your opinion as an axiom for how players work and behave or play the game.

In my previous reply I also suggested an alternative to your idea. It's more or less what you proposed, and even to a greater extent, while at the same time keeping the Elo rating as it is: It's a rating of how good the player is in 1vs1 on whatever official map* playing with whatever and against whoever. The players can still pick their games etc and it would still show in the rating since a master of all maps is more likely to play more games and also get a better rating because of that (granted he's winning). Instead of having a million ladders for each and every player type, we'll have just one, the current, and also with the possibility to see all the kinds of stuff you asked for and then some and at the same time keeping the rating as it is today. Since you would be able to see who the best players on map x are, you'd still be able to use the ladder for the kind of "internal preferential rating" you request. Thus, case is solved :)
i can code both program and both HTML for free if you need. But not PHP.. sorry
I'm happy you want to help out but the site won't use software and will stay PHP based as it's more than sufficient granted we have somebody that works on it. Since you seem to know Visual Basic (or some other language you coded your software in) and also HTML it would be very very easy for you to learn the amount of PHP that we use on the site, which is really the core and basics that would take like 3 full days to get a grip of. As a matter of fact I used the site as an excuse to start learning PHP myself from scratch (allthough I knew VB and mIRC scripting beforehand) and have learned a great deal by working with the site. If you're interested in helping out and becoming our coder you're welcome to start learning PHP and joining - I'd give you any help you'd need if I'm able and would of course get you started up with the code.

enclave
Posts: 936
Joined: December 15th, 2007, 8:52 am

Post by enclave » February 28th, 2008, 9:12 am

thanks for reply :) was happy to read all of this :)

i was going to learn PHP long ago.. and mySQL or something like that..
but stopped as a result because it requires to have "server" on your computer..etc etc.. but i also made the server.. then somehow stopped.. maybe because wanted much and at the same time or because there was no point in using the skills further..

i would like to help and learn PHP.. yes i have skills in visual basic.. in delhi pascal, borland pascal, mirc scripts, HTML.. also javascript (dynamic html or something).. another example of my unfinished work here: http://www.enclave.valsts.lv/

where can i talk to you more privatly..? i would like to help with what i can.. maybe best way to talk would be irc.. just tell me if you are often on irc and which channel i should search for you.. or you know better way to talk?

thanks

User avatar
eyerouge
Posts: 380
Joined: June 29th, 2007, 4:37 am
Location: wtactics.org
Contact:

Post by eyerouge » February 28th, 2008, 11:30 am

I don't use irc, but can be reached through:

Mail: spam (at) eyerouge (dot) com
Icq: 462069085
Msn: subcodes (at) hotmail (dot) com
Jabber: eyerouge (at) gmail (dot) com

enclave
Posts: 936
Joined: December 15th, 2007, 8:52 am

Post by enclave » February 28th, 2008, 11:53 am

i will use msn :)

enclave (at) inbox (dot) lv -> just in case..

User avatar
eyerouge
Posts: 380
Joined: June 29th, 2007, 4:37 am
Location: wtactics.org
Contact:

Post by eyerouge » February 29th, 2008, 2:14 am

leocrotta wrote:@eyerouge: I think we touched the issue recently. Just to make it an official request/idea :wink: :
Subject: Arranging the rendez-vous. - My ideas about the contact-possibilites are the following:

case 1) PlayerA isn't in waiting list and doesn't want to be addressed for playing => no issue...
case 2) PlayerA is waiting for a game (displayed in list - overrides "Contact me to play." +/-)
=> PlayerA should be on the specified server - either the one(s) he mentioned in his profile or the one displayed right after his waiting list entry.
=> PlayerB can address him on that server. (I believe if you're waiting for a game and can get instant messages or mails then you can also idle in lobby to make it even easier)
case 3) PlayerA wants to be addressed for playing (displayed by : "Contact me to play".) but is not in the waiting list.
=> PlayerA is actually not available for playing right now but can be addressed by PlayerB via IM or Mail for arranging a rendez-vous.

Blue : todo by players.
Red : todo by coders.

(just an idea though and all under the assumption it isn't already like this, at least I didn't make the experience...) :wink:
Leo, a version of what you suggested has now been added. Hopefully it will do the trick for you. It's explained in detail at the site news.

nataS
Posts: 166
Joined: January 28th, 2008, 3:21 pm

Post by nataS » February 29th, 2008, 7:26 am

Suvorov just owned Koncre, making 6 kills against 3. Then Koncre disappeared. Win for Suvorov imo.

User avatar
eyerouge
Posts: 380
Joined: June 29th, 2007, 4:37 am
Location: wtactics.org
Contact:

Post by eyerouge » February 29th, 2008, 1:59 pm

nataS wrote:Suvorov just owned Koncre, making 6 kills against 3. Then Koncre disappeared. Win for Suvorov imo.
nataS, allthough it's not in any way overly apparent, the purpose of this thread is to inform about the ladder and/or use it as a way to discuss the ladder itself. Specific games are not discussed or commented unless really needed. (And in any case the people running this forum recently made it clear that this place is wrong for such threads)

What more:

1) the parties involved haven't complained or contacted me. If they experience any kind of trouble, they are welcome to do so and can do so if they wish. I can't interfere between x & y on z:s request even if I had the authority to do so in the first place, which I don't since I won't give it to myself.

2) the ladder is a tool. It doesn't do specific game rulings or use "wesnoth referees" since there is no need - the game is either legit or not according to the current rules when it was played. The rules are simple to understand. If a situation arises where the rules are ambiguous and/or miss out on something that could be relevant or somebody can argue that a change is needed we'd of course discuss that topic instead and perhaps revise the rules, something which happens all the time because of great minds among the players. On this point I'm guilty for not making this more apparent, hence I'll revise the FAQ and add this info.

3) I can't do any rulings on games, as it's virtually impossible to make a 100% accurate ruling, even if objective veteran players can make a qualified and good guess about the outcome of a game. The problem with rulings and the reason why I'm against them is that there are few objective and veteran players that will be around to do them, and even if there were, even they would have different opinions from time to time. It would also make the "ladder leadership" (I don't see myself as leading this project in that sens or as a boss/authority of any kind and don't want to be associated with such a role) dictatorial and authoritarian, central governed, something I believe an open source and community driven project, as I hope the ladder will be and which the game already is, should be without.

So, even if I have a personal opinion about who would win in your game, I won't speak of it, mainly because it's not a fact - it's a qualified guess,

4) This is a non-issue since you can't really "disappear" in a game of Wesnoth: You're either connected or disconnected from the game. If he disappeared by disconnecting the rules already state that he lost the game unless they agreed to re-load it. If you with disappearing mean that he stayed connected but didn't do anything during his turns it's his right to do nothing on them according to the game rules - that's why the ladder has a timer setting rule - to hinder scenarios just like that.

Even if it perhaps doesn't sound like it, it was a good thing you posted as you're the second one to indicate some kind of "ruling process" and I still hadn't added this info in the ladder FAQ. I'll fix. :)

Edit: Fix is now done.

User avatar
eyerouge
Posts: 380
Joined: June 29th, 2007, 4:37 am
Location: wtactics.org
Contact:

Post by eyerouge » March 1st, 2008, 7:20 pm

The ladder has finally been upgraded to handle inactive players.

It made no sense to have a ladder cluttered up with people that aren't active for the time being. It also seems to lack function to list those individuals since it's the active players who are actually present and use the ladder.

Because of this I have now implemented the passive mode: If a player hasn't played a game the recent 30 days he isn't included in the ladder table.

I sum it up in details in the also updated FAQ/Rules:
What does passive rating mode mean?

In order to keep the ladder meaningfull, up to date and to encourage activity we have a rule that says that you must have played at least one game within 30 days if you wish to be listed as a valid contender in the ladder. If you don't play a game for 31 days your account will be set in passive rating mode. While beeing in a passive rating mode you won't get listed as competing in the ladder and you'll temporarily lose your place in it until you play a new game. You'll then be considered to be an active player again and you will be automatically removed from the passive rating mode. You will of course also regain your proper rating, taking the new game into account, as usual. And to answer the question everybody fears: No, you won't lose rating while beeing in passive rating mode. If you had 1500 when you were put in the passive mode, was passive for 3 months and then play a game where you win 10 points, you would then be an active player again, but now with 1510 points.
Add to this the following, also available in the FAQ/Rules:
1. To be rated you just need to play a game. The more games you play against different opponents, the more accurate your rating becomes. All players that have played one or more games always have a rating.

2. To be ranked you must have played at least 1) 3 games and 2) have played at least one game the recent 30 days. The rank orders players by comparing their rating and also the amount of games they've played. The more, the better, in both cases.

3. To be listed in the ladder table you must 1) have a rating that's 1400 or above, 2) have played at least one game the recent 30 days and 3) played at least 3 games in total.
Lastly, while you are logged in you can now also see the amount of days until you are put in passive mode in your profile.

Post your thoughts or questions on these topics I personally thought that 30 days seemed to be a reasonable amount of time for a player that's registered on a ladder and also that the coming-back-to-the-ladder-part is easy and straight forward, non-punishing for those who have been away from Wesnoth for whatever reasons.

User avatar
eyerouge
Posts: 380
Joined: June 29th, 2007, 4:37 am
Location: wtactics.org
Contact:

Post by eyerouge » March 3rd, 2008, 11:22 pm

RSS support finally added. Now you can get the result of the latest games anywhere you want and/or the names of the newest players.

User avatar
eyerouge
Posts: 380
Joined: June 29th, 2007, 4:37 am
Location: wtactics.org
Contact:

Multiplayer Divisons?

Post by eyerouge » March 10th, 2008, 1:59 pm

I have been thinking about dividing the ladder into 2 divisions, one major where the moe skilled players would be, and one minor where us mortals would be found, and I'm wondering if it sounds like a system that would make sense for Wesnoth?

The way it would work is that a division is defined by a ceratin range of Elo rating. As an example, all players that has 1500+ would be in the major/pro division, while the others who have below would be in the minor.

That way you'd see 2 different ladders when viewing the ladder table. The major division would have it's own ranks, and currently OWlface would the leader. The minor would have it's own ranking, and currently I would be the leader of it. At the same time, players would still be free to play against anyone who is part of the ladder, independent of which division they appear in.

The point being? Well, I guess it's mostly psychological: At times I wonder why on earth somebody would be interested in the ladder or following it's results if he's on the 50:th place, as an example. It makes more sense and seems more intuitive to keep the better players in a separate list and the not-so pro on another. An argument for it is that the not so pro players would be ranked compared with other players in their own skill class. And that seems as a fairer way to rank them.

So, even if it's true that I suck compared with all players, it's not as true that I suck compared with players that have about the same skills. If so, maybe it isn't really all that fair or interesting that all players appear in the ladder right now, since they are of very different skill classes. I don't know, what do you guys think? To me it seems like if all players except the top 20 will be discouraged from participating and playing since they'll think they aren't doing that well, while the truth is that the maybe are but that the comparison of the ladder, which includes all the top players as well on it, isn't forgiving ;)

The system I suggest would work in a way that allowed every player to play against every other that wants to, just like it already is. Nothing would change on the players side or how the games are reported. However, you'd appear in different divisions (major / minor.. or maybe we should call them senior / junior?) depending on your rating. Whenever you qualify (get high enough rating) the site would move you into the correct division. So, one day I could be number 2 in the minor division, and after 1 week and many victories, I get transfered into the major division, where I could end up as number 21.

Is divisions something I should implement or not? Drawbacks?

grrr
Posts: 252
Joined: May 25th, 2007, 9:49 pm

Re: Multiplayer Divisons?

Post by grrr » March 10th, 2008, 3:06 pm

Division only make sense if you want to restrict players by it:
Players can only play a scored ladder game with players of the same division. Since you are against that, you basically answered the question yourself already =) Keep in mind though that the ladder ranking would become more expressive if skilled players are restricted to get their points from other skilled players only. Currently, your ELO system has only a gap of 300 points between tops and flops (so it seems), so you still gain points if you win vs. unskilled players.

User avatar
eyerouge
Posts: 380
Joined: June 29th, 2007, 4:37 am
Location: wtactics.org
Contact:

Re: Multiplayer Divisons?

Post by eyerouge » March 11th, 2008, 9:35 am

grrr wrote:Division only make sense if you want to restrict players by it:
Players can only play a scored ladder game with players of the same division. Since you are against that, you basically answered the question yourself already =)
Well, I tried to show that there are some reasons for why it could be worth using division even if the players are not restricted to play with people from their own division, since the wouldn't be because we use one Elo rating and the Elo rating is in return used to place the people in the correct division. The reasons I gave were mainly psychological and motivational, and there would be some practical as well as it seems to make less sense to list every player, independent of skill class, in the same list and assign him/her a position (like it's done now).

The problem with only allowing players to play others from the same division would the that the number of games played would become lower since there are less players to play with then. Hence, it's not really an option at this point, even if it would be ideal at a later one.

Keep in mind though that the ladder ranking would become more expressive if skilled players are restricted to get their points from other skilled players only. Currently, your ELO system has only a gap of 300 points between tops and flops (so it seems), so you still gain points if you win vs. unskilled players.
I recently added the win p/loss p/total points statistics into every players profile. By looking at it you can see and get to know how a player usually acquires his Elo points: If he usually wins about 12 p when he wins, then he usually playes equally skilled players. If he wins more than that then the system suggests that he usually plays better rated players.. and vice versa.

Furthermore, there is a difference in ranking and in rating: With rating I mean Elo rating. Ranking on the other hand would be your position in the ladder. The two do of course correspond, but except for your Elo rating the ranking also takes into account the number of games played. If we both have 1500 and I have played 1 game more than you, I'd have a better ranking even though we have the same rating.

There is also an issue that is related to the Elo system it self: It wasn't designed to disallow to good players playing against bad. That fact is even included in it's design: If a good player wins over a bad, he would gain a low amount of Elo points added to his rating. At the same time if he were to lose against a bad player, he would lose a great deal of points. So, it's fair to say that this is already covered and that players are more or less, indirectly by the Elo rating, encouraged to play with people from the same skill class. Also, if the difference is 650 (if I remember it correctly), the better player would get 0 points for a win.

I also believe that divisions would make the ladder ranking more expressive, as you wish to have it. Imagine this: Currently the 10:th ranked player must look x amount of good since he has 40 players beneath him, with lower ranking. If we were to divide the ladder into 2 divisions, putting all players that have above a certain Elo rating in one and the rest in the other, and the player at 10:th place still remains in the majoir (better, pro, senior) division at 10:th place, wouldn't it then look as he was y good instead of x good? After all what makes a players amount of "good" isn't just how well he fares with the people ahead of him in the ranking - it's also how many that come after him in the same.

If all this is true, there's yet another argument for divisions. Or am I wrong?

User avatar
governor
Posts: 267
Joined: December 8th, 2006, 12:32 am

Re: Multiplayer Divisons?

Post by governor » March 16th, 2008, 2:45 am

ELO already ranks players causing divisions (i.e 1400's, 1500's, etc.). There is no need to create divisions, finding opponents can be challenging enough already.

Anyway, I vote against it.

Some quick points:
- New players will improve faster when they take on more experienced players.
- Do not expect any ranking system to normalize for a long time.
- Divisions will guarantee that beginners learn bad tactics and continually deploy them because they become successful when they should not be.

Post Reply