Implement the Wesnoth Ladder

[eyerouge]

I hardly understand, never the less READ, most of it.

Of what, exactly?

However, statistics are usually a way to seperate people from other people, supporting seperation in an OPENSOURCE game.

"Separation" has negative connotations with it, but yes, statistics could be used for that. Actually any competitive framework can be described as a method of trying to "separate" players from each other. Erhm. That's the whole idea with competition - to measure different performances. Describing these performances as "separate" doesn't. in itself, explain why it's good or bad.

This is starting to sound good besides that part, a little awards would be great!

If one buys into your reasoning, then one could as well describe different awards as separating (as not all players have the same amount/same awards etc etc). But for some reason you go against your own argument here...

Opensource is here for the PUBLIC, not a few select people who think they're awesome and have played the game a lot and have acquired a new ability to own with 1 faction, on 1 map.

Yes, exactly. Open source if for the public. Not a homogeneous public where everyone is identical to the non-ladder player or to the ladder player. The game can be played casually by those who want, and competitive by those who prefer that. The first one is already true. The second is not not (in game), hence the discussion in this thread. What's the problem? I don't see any.

[Gambit]

Yes, exactly. Open source if for the public. Not a homogeneous public where everyone is identical to the non-ladder player or to the ladder player. The game can be played casually by those who want, and competitive by those who prefer that. The first one is already true. The second is not not (in game), hence the discussion in this thread. What's the problem? I don't see any.

Indeed.

Personally I could never get into competitive TBS's or RTS's (probably because I'm not good at them ), but there are loads of people who would love this feature. What my last post meant was I was asking that the required logins, the removing host power, the determining a winner immediatly upon disconnect, etc would only be on the ladder part of the proposal. Having a seperate "ranked" system built into the game would still be cool imo. And all of those restrictions make perfect sense inside of the ranked system. I just don't like the idea of putting those on the casual side of things too. He probably wasn't even thinking that direction. I was just making sure. These discussion threads are for fleshing out details like that afterall.

It is what people will make it by the way. If people start to think based soley on stats then that's how things will go. A very large part of me (my stomache ) hopes that they won't.

[eyerouge]

Gambits non-comp.:

Personally I could never get into competitive TBS's or RTS's (probably because I'm not good at them ), but there are loads of people who would love this feature.

Just as a sidenote: Try it out. I myself suck at Wesnoth compared to many of the top players (ladder or no ladder) in a 1vs1 and have played the game too little to posses the skills to be a top contender. That said, I still enjoy playing competitive. Not because it makes me feel good when I do win or bad when I lose. To me it's more a matter of keeping track of my skills/development and other stuff related to the gaming.

Sure, I'm maybe not the expected hardcore player and still consider it a game/relaxation instead of something of life-and-death importance... but does it matter to me as an individual, really, why and how others perform ẃithin a given competitive framework? In my case no, as I see no reason for it to do so. Hence, I suggest that if you ever do participate in some competitive environment, make it your own thing: Competing to just win and reaching the very top will, for 90% of the players, be a devastating experience Besides - they miss out on the obvious - participation in a formal framework can become fun in ways more adapted to the individual.

Having a seperate "ranked" system built into the game would still be cool imo. And all of those restrictions make perfect sense inside of the ranked system. I just don't like the idea of putting those on the casual side of things too.

I think all are in agreement on that one, but yeah, good to clarify: Needless to say by now, there is little to no rationale to incorporate ladder specific/required features in non-ladder games if they aren't somehow relevant or improving non-ladder game experience on their own premises.

It is what people will make it by the way. If people start to think based soley on stats then that's how things will go. A very large part of me (my stomache ) hopes that they won't.

I will try to not make this into wPolitics, but feel it is at it's place and relevant nonetheless. In the end of 2007 I realized that there would never be a ladder or any competitive resembling framework written into Wesnoth. At least not as the situation looked like then, and continues to look like within parts of the developer community. Different people had their reasons to be skeptical about turning Wesnoth into something "competitive" and thought it would spoil the game and the spirit of it in various ways.

While I always understood their worries I never quite could grasp the very conservative and at times almost hostile attitude against something as basic and common as competition and the stats that are associated with it.

At the same time I will be the first one to admit that all competitive frameworks can indeed be associated with good and bad phenomenons, and, that they may affect how the game is played and also the social aspects of it's community. Nothing is perfect, and certainly not competitive systems, no matter in what sport/game you choose to study.

The big question in my mind was not why the developers didn't listen to me specifically or the part of the community that had asked for the feature(s). I believe developers in open source projects should always be free to decide whenever they work and with what. They are giving us their time and efforts for free, which is very appreciated and should hardly be questioned because they didn't happen to code every whim of the community.

The central question for me was if the nay-sayers were correct. I thought that they very well could be. I have myself witnessed very disturbing and non-productive gaming environments filled with hostile players on steroids, and I wouldn't want the Wesnoth community to turn into such an abomination. Sadly many games are non-fun because their communities suck. Now, one might ask: Do they suck because the game is competitive, or are there other factors that can explain the suckage? One obvious answer is that it would depend on the game and the player: Is the game playable in non-competitive modes or not? And what does the individual want? Let's leave that in the open for now, and head back to Wesnoth:

BfW has in it's very nature by simply allowing multiplayer games always been a vehicle/tool for competition. It doesn't matter however people define competition. In reality, and if one skips on the semantic masturbations, most common sense players would agree that any game where people play vs each other have something competitive about them and that that is perhaps the very core of competition. To try to defeat an opponent, a human being, is to compete with him/her. That's the very definition of competition stipulated by me for this discussion. Thus, Wesnoth developers has since long ago made it very possible to play competitive games using their creation.

Reply to rants about MP feat.:

Here somebody will claim that Wesnoth can be played in co-op modes, in co-work campaigns, and that the multiplayer ability doesn't have to be used to compete. Yes. So what? I never claimed it has to. Just that it can and that it was coded by the same people(?) that didn't want competition in the game. I'll even claim that most times the multiplayer feature is used is to actually compete against someone (while that also allows you to co-work with someone, i.e. in a 2vs2)

The MP-ability has been there since whenever. What lacked was 1) a formalized structure (rules etc) and 2) the technical side which enables people to keep track of the results and what not.

To me it was perfectly clear that a tool becomes what it's users & designers make of it. If you have crap users they will take a crap on it and each other using that tool. If you have a good community with a semi-decent user base they will be able to use the tool correctly (granted it works by design ofc, after all, they are restricted to adapt in that sense to the limits and possibilities of the tool given to them).

I asked myself why the Wesnoth community was more potential to become a horrific experience simply because there would be an alternative competitive body offered to the players that want to participate in it. I could honestly not think of good explanations. I had seen many strategical games, both with a worse off community and also different levels of game complexity, where competition seemed to work between those that craved it. To me it seemed very plausible that many of the nay-sayers argument, that BfW community and spirit would be pooped on by competetion, could be wrong, in the same sense it could be right. I was open minded and admitted I didn't know. I never once said it would work back then, nor did I have invaluable and unique knowledge on the topic. As a person that wanted to know PHP and also saw a void in the community of an amazing game, I took the opportunity to find out.

It has now been over 2 years since I did that. I was unknown in the BfW community, could hardly at all play the game (heck, I think I even lost against the AI at times when I setup the ladder.... ) and I wasted countless of hours on improving the code of the ladder site & learning PHP at the same time, 100% on my own. At first I was the only player registered. Later on there were 10, and very very slowly results would start coming in, and more players would sign up.

As chains mentioned (btw, he later came aboard and mastered the code, contributing plenty to it), there is now 26 000 played ladder games. Between 200 to 300 are played every week, by around currently 123 active players. All in all there are 3 300 registered accounts on the site from all over the world.

These are by any standard of measurement very good results considering a) I got zero backup from the established community beyond a sticky-thread and later on doc's help with maprelated work. b) The project is un-official and it says so everywhere and c) it is a very lame system using a third party solution, demanding users to leave the actual game etc to report/view results/news/ratings and so on.

Of these I think c) is the most interesting. Despite it there seems to always be at least 100 active players on the ladder. That,if anything, proves one of the two points I wanted to check out by putting the ladder site online:

Point 1: Some people want a ladder. Those people use a ladder. Proven.
If 100 are prepared to use such a very impractical solution, which also has inherent flaws since it can't communicate with the real game client, how many would use an integrated system? My guess, to be honest, is somewhere in the range of 200 to 500, if it was today. One thing is for sure - the current number would only rise, as the built in ladder would be more correct and also more practical.

Next, was my concern about the wicked sinful attitude. The whole community would supposedly become corrupted. Hormone filled teenagers would grow another set of b*lls & go on a rampage, something would become very naughty in the otherwise oh-so-perfect Wesnoth community IF we associated the game with a competitive framework... at least so the prophecies said. (Yeah, I exaggerated a little, I'm allowed to do it - how else can anyone amass this amount of text? Besides, it still captures the point of they nay-camp (no pun & wordplay intended here btw))

Now we have the result. How has the community degraded? What has really become worse off? How have people that don't play the ladder become affected by it? And those that partake in it right now? I wouldn't know, since I haven't done huge studies on the subject. At least I can admit that, which is more than many others would say when they declare their firm beliefs on the topic at hand.

What I can say is this: Judging by the numbers, and by the overall attitude in the community, things are like they were before. Not much better nor worse. With the only difference that there is now an active multiplayer community that seems to have fun playing wesnoth in a formal manner, with rankings, ratings, and stats. And those that don't like it also seem to be sane enough to stop doing it or never try it.

There are very polite and friendly (non-)ladder players, and there are very idiotic (non-)ladder players. (Yes. What a suprise. I shiver... as if that couldn't be foreseen ) Overall and in general, I'd bet that if studies were conducted they would prove that the ladder has actually contributed with more positive than negative to the Wesnoth community and Wesnoth being played as a game.

Point 2: Let's face it - the scare-you-propaganda is game over.
It doesn't seem to be a true notion that Wesnoth will become the woman-that-has-sinful-adventures-for-money of Babylon just because it invites the segment of competitive player as well as the han-solo players (campaigners?) who never competes in any form or the MP:s that only co-work online. All the numbers are more or less cold facts that suggest otherwise, especially coupled with the supposed non-existent harmful attitude change in the community that I'm still waiting to become the victim of.

So to answer your worst nightmares and to get back to your post:

It is what people will make it by the way. If people start to think based soley on stats then that's how things will go. A very large part of me (my stomache ) hopes that they won't.

Some player already are total stat freaks. Some are also/or bad losers, not mature, kids, people that come and go, nasty, rude & un-polite. There is a whole plethora of users of any software, and especially a game. It doesn't seem to matter much in the case with Wesnoth - the community is more or less as (un)productive & (un)polite as ever in the 2+ years I have been around. The whims and personal characters of individuals do not build on or fall by the fact if there's a ladder or not in Wesnoth. In a worst case scenario however, a system can bring out the worst in most of us.

I argue that a competitive framework has failed to do so until this point in Wesnoth. I also believe this is the largest experiment of it's kind for this particular game & community. All in all, I'd say it has gone pretty well. I even dare say I take pride I did what I believed in - creating an opportunity for those that want to compete in Wesnoth.

Like it or not, but competitive Wesnoth is already around and people enjoy it (else they wouldn't keep using it).

Question is how this will show in the game and if the devs can learn anything at all from this. Either way, I can almost guarantee that, very much thanks to the ladder, some form of competitive functions are guaranteed to be built into BfW sooner or later. My guess is that they will be in eWes first though (and in such case there would be serious implications for BfW over time if eWes would offer everything it had equally good and on top of that a competitive framework as well, but hey - what do I know...)

[Evelangos]

I meant I didn't read most of it.

I'm just saying that I don't want people excluding others, after all, that's why I host on this game.

I seem to be horrible at making my point (anywhere) so I'll just say: Make your ladder, but keep the stuff you want in it. (As Gambit said)

[Aethaeryn]

I'm going to reiterate my point. I'm not opposed to an integrated ladder as long as it stays for mainline maps. Nonmainline is imbalanced and will skew numbers greatly, not representing skill (try playing elves on a map with no, or too much, forest) and those not of the standard style (RPGs and survivals) require radically different skill that a ladder cannot measure. Basically, if you are going to port the ladder to the game server, port its rules and scope as well. Don't try to expand it into unproven ground or you'll turn it into something other than the success it has been. Keep it to an approved set of 1v1 (and 2v2?) mainline maps and keep it optional.

If you want stats and ranks above and beyond the current ladder, please propose it separately. Don't take a reasonable proposal most wouldn't oppose and then attach ridiculous and harmful additions (subtly and in a few sentences among many paragraphs) to it like a politician does.

This is how I can disagree with chains without disagreeing with eyerouge.

(EDIT: Please realize I speak with opinion and not authority. What is the worst I can do? Not document it in the wiki and not translate it into Latin? No, I speak here wearing my hats of player and of UMC maker.)

[Pentarctagon]

imho, 1v1 mainline maps should definitely be used and 2v2 if they can be considered balanced for that kind of play. keeping statistics for UMC 2-4 player maps would also make it easier for mapmakers to balance their maps faster. for survivals/rpgs it would be interesting to have no ranking for them, but longest playthrough/shortest playthrough/most kills/least kills/etc. names wouldn't even need to be displayed for these either.

[chains]

This is how I can disagree with chains without disagreeing with eyerouge.

I agree with Eyerouge on everything. If his words suit you better than mine, use his. My goal was simply to write down some of the many ideas we've discussed over the years so that they might be useful. Pick what you like, code what you enjoy, ignore the rest for someone else to deal with if they want to.

Since it's clear I didn't make a few points:

* I believe most of us agree Ladder implementation should be separate from casual play.
* The ladder requires registration, and casual play does not.
* The ladder requires clear winners.
* The ladder really benefits from auto-matched games.
* The ladder has a defined set of rules that could be used as a good starting place: http://ladder.subversiva.org/faq.php
* Some of the ladder features happened (or didn't) because of technology limitations such as 2v2 ladder.

Open-source is about the freedom to change the software you use to suit your needs. There are many people who want Wesnoth with a modern competitive multi-player interface. Just because someone changes a piece of software, doesn't mean you have to use their changes. It's not about separation.

When we work together, we build in open-source what proprietary games can only dream of.

[Gambit]

Time to start thinking more specifically then about how exactly you want this to work.

If I understand the process correctly, at this point you should be considering how exactly the players use this. Example:

1. The player logs in.
2. The player clicks an "auto ladder match" button.
3. The game checks the players rank and tries to find a player with a certain rank range (Default within 5 ranks).
a. If a player is found move to step 4.
b. If nobody else with a similar rank is found, ask the player if they want to increase the range. If yes, up the range by 5n where n is the number of times you've come to this step. So if this is the first time then you'd have the default 5+5(1). So you're now searching in 10 ranks from the players current rank. Next time would be 10+5(2) or 20 ranks. The next would be 20+5(3); 35 ranks. If they choose no then keep the range how it is. Either way go back to step 3 and repeat.
4. Match the players on a random 2 player default map.

A player's rank is equal to their wins divided by their losses.
0.1 = rank 1
0.2 = rank 2
...
0.9 = rank 9
1.0 = rank 10
1.1 = rank 11
...
3.1 = rank 31
...
10.2 = rank 102

Just an example. If I understand correctly, the coders will need descriptions for all aspects of the request.

[eyerouge]

Just an example. If I understand correctly, the coders will need descriptions for all aspects of the request.

This is a continuation of the way that Gambit paved and not a correction of what s/he wrote:

Rating System
There are really only two variations of viable working and sound rating systems. One is Elo, already in use at the current ladder and by zillions of gamers that play chess and pretty much almost any other type of game which is skill based.

The other one is Glicko2 or variations of it, which seems like an interesting option.

Beyond these two I'd be prepared to argue (but won't) that there are very few working alternatives. This of course also depends on the following question:

What-we-wanna-measure
Do we only want to keep track of stats? If so, no real rating is needed.

I think most in here have an idea that the ladder should actually try to do a rough estimation of a players skills. (I write rough and can at the same time complete that with "as accurate as possible given all the built in constraints and inherent problems with any ladder system" snce all ladder and rating system they're based on have their own issues.)

So, if that's true, only getting a rank based only on some calculation of how many wins/losses you have and the relative ration of that measures nothing relevant at all. This was an example given by Gambit, and as an example it works terrific, in how a ladder should never be done and to get us fired up with offering the details for a system.

I myself would probably vote for keeping Elo (but with slightly adjust settings of what the ladder currently uses, since it uses the wrong K-values for high-rated players, giving them less winpoints than they should have, coupled with the idiotic protection system that it still uses despite the fact that I coded a better one and the ability to shut it off completely) or going with a Glicko2 variant, or some system inbetween. I won't type down the details in here as they are widely available on the net for everyone to read.

Personally I think a ladder system should measure/keep track of everything that is measurable without making life noticeably harder on the players. Stats like how many times I played with Elfs, how much I won/Lost with them in general, and how many % of those losses were against undead, etc etc should all be there.

It's highly interesting and can help the player to find his/her weak spots and develop them further. Refining his/her gameplay even more and to new levels, by using all the statistical dumps on his/her account associated with ladder games.

This will also help you guys as developers with balance issues and what not. (I have already written more about this and in greater detail and will not repeat my self and clog this thread.)

At the very least a good system should measure most of what's on the already existing ladder (except the xp-crap which I only added to spur kids to play..) But only the sky is the limit: Data-wise storing all this info is still very easy an cheap, and it eats very little bandwidth.

In the end, focus on a ladder must stay at measuring skills, and doing it as fair and accurate as possible. Whatever is coded should have that as it's always standing primary goal, and, everything else as secondary objectives that can improve the fun and knowledge of the game and the community itself.



Ranks vs Rating
Please make a clear distinction between rating and ranking. Depending on what system one uses, this distinction is very much there. I don't know if I captured the correct wording or if it should have been vice versa when I created the ladder, but this is how it has been defined this far and by it (correct/invert if wrong):

In the case of the ladder the Ranking is your position relative to other active players on the ladder. For example, you could be 1:st, 2:nd, 7:th etc etc. Your Ranking can, with other words, change even when you don't play a game, since other players do so and can surpass you, pushing you downards. You could even climb up on the ladder if somebody else played badly and went downards, or if a player above you became inactive.

Rating is on the other hand not associated to your position to anyone anywhere in such a way. Your rating only changes once you win or lose. Nobody and nothing can take it away from you, at least not in an Elo system, unless you play. Also, while you always have a rating and get to keep it while inactive (in Elo, not in all other systems) without it decaying etc, the same can not be said about a ranking. At least not in a well coded ladder which would remove passive players from it's rankling-list.

A rating is purely built on your skills (and your opponents, those you played against) while a ranking is a cocktail of other factors as well and much less interesting or relevant for most players that actually understand the system. Thus, a player with a ranking of #1 and a rating of 1600 is probably a novice compared with a player that is inactive and lacks ranking but has a rating of 2700.

Keep this in mind, no matter what systems you'd evaluate.

[Sauron]

All that is truly needed is interface/protocol for communication with HTTP server hosting ladder - the game client might just send AJAX-like requests; the matter of linking in-game userids with ladder userids is more complicated as many players use multiple instances and server does not support multiple clients per nick - thus one could store another node in config - like ladder-userid.
One must though think out what needs to be done - not only outcome reporting is needed, but also the interface for declining/confirming particular battle result, listing of pending unconfirmed battle results and many things I haven't realized yet.

What is the result of such a solution? The fork team gets dependent on ladder team - nothing bad about it, but some organisational matters arise - someone should take care of the cooperation;

Long ago I suggested in my thread's mod there should be "less random" ladder for some agreed settings - maybe this could be linked with the old idea of mine.

[eyerouge]

What is the result of such a solution? The fork team gets dependent on ladder team - nothing bad about it, but some organisational matters arise - someone should take care of the cooperation

In general, it's a bad idea for teams to depend on each other unless the teams are stable and there is continuity in their work.

I wouldn't recommend any team anywhere to depend on a second party if they consider coding an in-game ladder & have the knowledge. Using a site is really only a solution for games which don't plan on or don't currently have a ladder.

Reasons are about control over one's own product, work(flow) and end result, and mainly and maybe even more importantly in this particular case - the ladder has one semi-active coder working on it (Tessa) and none which are lined up to carry the torch once he starts doing something else on his spare time.

Because of that I'd argue that the eWes-team should not depend on any external crew for their development of a ladder, unless of course, they can strike some kind of agreement with that crew and get the stability of it's work and existence re-assured.

[Sauron]

Because of that I'd argue that the eWes-team should not depend on any external crew for their development of a ladder, unless of course, they can strike some kind of agreement with that crew and get the stability of it's work and existence re-assured.

Reimplementing ladder for purposes of this fork, especially as C++ code is an unnecessary reinventing the wheel. Much better - delegate someone to cooperate with ladder team - as co-developer able to pick up from where it ended. Well defined interface plus access to source code of existing ladder is enough to move ladder to any host situation requires.

On the other hand - I think game might allow various ladder servers depending on user preference - e.g. 'non-random' ladder, 'default-wesnoth' ladder and so on. The server developed here should be only advanced enough to provide basic testing platform for interface - i.e. fully functional, but w-o effort put into "look and feel" - kind of "base class" extendable by users at will. Just my suggestion.

[chains]

While this is an idealistic idea Sauron, the main feature of a ladder currently does not exist for Wesnoth. That main feature to have is ladder matches being randomly matched between players. Reporting of the matches is solved by the current solution. Wesnoth communicating data to the ladder would only cause things to get more complicated. Reporting is taken care of more expertly by the ladder players than an automated API solution could match. However, matching ladder games is where things are still very exploitable.

For ELO to work effectively, players have to be randomly matched. Once you can choose who to play, your score can be manipulated. Also it's important to remember that the ladder doesn't have a team, doesn't often do backups, is sitting on my personal machine, and really only has administrators on a haphazard basis.

[Yogibear]

I would like to start working on this, but I think there are two questions we have to get clear first:

1.
Should this be mandatory (namely the random matching of players)? I don't have much of an opinion on this one. Maybe this should also be discussed within the ladder forum thread.

2.
Should the game use the existing ladder website functionality or should there be a completely new implementation, with its own database and reporting mechanism?

I tend to say no, at least not in the first step. On the one hand, i don't like to rebuild existing functionality unless it badly needs it (and i don't think it does here, the website makes a good impression). On the other hand, it most likely means to modify the wesnoth server. This is something i would like to avoid if possible as it would involve to set up a new server and completely lose multiplayer compatibility with current mainline wesnoth.

But - I don't feel in the position to decide this for the whole ladder community, so please follow your own thoughts here.

[Gambit]

I would like to start working on this, but I think there are two questions we have to get clear first:
1.
Should this be mandatory (namely the random matching of players)? I don't have much of an opinion on this one. Maybe this should also be discussed within the ladder forum thread.

They decided no.

2.
Should the game use the existing ladder website functionality or should there be a completely new implementation, with its own database and reporting mechanism?

This hadn't come up.