Implement the Wesnoth Ladder

[eyerouge]

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.

Elo works best when pairing is random or near random within a certain skill class. I would make it a) first try to match within a similar skill class and if there is nobody around within 30 seconds I would make it re-match randomly or add the Elo span to the similar skill class so it's similar skill class plus/minus x... rinse and repeat until you get a hit.

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

Unless somebody guarantees to code on the ladder and to work with you it seems as a foolish move to rely on the project. As far as I know Tesafilmschen is the only active coder on it. You would have to consult with him and his intentions to maintain it.

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

To be honest, 95% of the ladder community has done nothing for Wesnoth as a project, nor the ladder itself, except for playing wesnoth and partially doing so on the ladder itself. What I want to say with this is that you shouldn't expect any elaborate or serious answer from the ladder community since it is pretty much the same as the average joe within the Wesnoth community - passive, inactive and clueless about what would be good or bad implementation of a ladder system. What people think is also totally irrelevant in contrast to what they can prove or argue for. Sadly the BfW forums, as most forums on the internet, are not really a workshop for ideas as they are for opinions and personal feelings and experiences.

I'm making this bold and perhaps unpopular statement that could easily be interpreted the wrong way based on the fact that virtually nobody that has ever voiced an opinion about how the ladder should work has the slightest knowledge about how Elo works and what the ladder (doesn't) measure(s) and what the results of it mean. Because of that it's mostly uninteresting and not relevant what the ladder community happens to think since it's often their own whims, and they're seldom based on anything that has any rationale behind them. For that reason alone, combined with the notorious lack of input as it usually is in a vast majority of all open source projects, it is quite often perfectly fine to proceed and code anyways.

That said, I still believe this is the way to do things: Going open with it and being welcoming as you guys do it has my full support. (Not that it's needed or that I would be in a position where it mattered or not, but just for the record and not to be mislabeled in this discussion.) Just don't wait forever for the input, because it may very well never come. It would also make sense to post this request on the ladder, as a news item, as most ladder players probably don't lurk in this exact forum.

[Yogibear]

Sadly the BfW forums, as most forums on the internet, are not really a workshop for ideas as they are for opinions and personal feelings and experiences.

I happen to disagree, at least for this subforum .

Because of that it's mostly uninteresting and not relevant what the ladder community happens to think since it's often their own whims, and they're seldom based on anything that has any rationale behind them. For that reason alone, combined with the notorious lack of input as it usually is in a vast majority of all open source projects, it is quite often perfectly fine to proceed and code anyways.

I sense some bitterness here but i can understand where it comes from. Anyway, i wasn't opting for a democratic decision of the whole ladder community, rather for some qualified input of "those who know best", namely you, chains, Rigor, Tesafilmchen and maybe a few other interested ladder players.

That said, I still believe this is the way to do things: Going open with it and being welcoming as you guys do it has my full support.

Indeed, i am looking for a pragmatic efficient way to push things forward here. I just have the feeling that i don't have enough information yet to set up a solid technical basis.

[nelson]

Has there been any forward movement on implementing the ladder within Wesnoth? I'd like to point out that the 3rd party nature of the ladder inconveniences not only ladder players, but non-ladder players who are trying to join a game.

Let's say that I am only interested in playing in a ladder game right now, and I create a game in the lobby that I intend to be a ladder game. No matter what I say in the title of the game, even if I password protect the game with the password "ladder", people will join the game who do not know what the ladder is, and immediately become uninterested upon learning they would have to register on some external website. This wastes their time and annoys them, and has the same effect on me. If there were a native ladder implemented within Wesnoth, then I would only see other ladder players when I am trying to play a ladder game, and non-ladder players wouldn't keep wandering into my game and would no longer be frustrated b/c I don't want to play with them.

Please, for the benefit of all players, implement a native ladder.

[Gambit]

Unfortunately, all I can tell you is IIRWIIR. I do imagine it would go faster with some programmers.

[nelson]

Do you mean that there is a programmer slowly working on it somewhere who could use some help? Or do you mean that there are literally no programmers working on it right now?

I'm a programming noob who could potentially help someone more experienced, but there is no chance I could take point on a project like that myself.

[Yogibear]

Do you mean that there is a programmer slowly working on it somewhere who could use some help? Or do you mean that there are literally no programmers working on it right now?

Unfortunately the latter, actually. As this is a bigger change request, we should maybe try to identify some smaller subtasks, so that it is easier to start moving. On the other hand i am pretty busy with real life atm, and there are also a few other wesnoth related things (like for example making Wesnoth Experimental compatible with 1.8.3/1.8.4), so yes, this is heavily lacking resources (has it ever been different ).

Maybe (just maybe) i lately found a ladder player who is willing to actively participate in development. If you are interested, i will see what i can do to give you (or others) a smooth start. For that it's probably best to meet on irc, i will try to be on the irc channel more regularly again.

[Rigor]

very interesting thread, actually more informing than the 50++ pages ladder site online. i especially like the direct approach where we talk money already.

i wanted to point out that there is some serious effort coming from just_end_turn (jet) and quetzalcoatl who are both willing to work on the ladder, but not as it is, but "coding it from scratch" if i understood both of them correctly.

what i cannot estimate is how novel it will really be, and how much more work, but they claim they have a better system to deal with certain technical things i dont understand. some things i find noteworthy to tell u at this point is that they r also trying to find a fair rating system, and although they have not been all to specific, i would heavily emphasize GLICKO2 at this point for any other changes on the ELO. and no more incentive for noob players.

and Q is much into statistics, eyerouge u should really talk with him i think he would like ur other thread and didnt know about it yet: http://forums.wesnoth.org/viewtopic.php?f=36&t=29733

talking about something different: could i have a short (plz really short) update on what we all have agreed so far? including information like who will do what kind of work, which language used, which features do we want to have implemented (what is the latest info about implementation in mainline?). ive been reading some of eyerouges marathon posts which were really really great but for giving me a short overview too much.


more than doing nothing, reading and writing, i prefer some action, and i want to see some results anytime soon too (before i have a family and no more time for this life project - thx for understanding )

my understanding is:

now jet and Q are working on ladder stuff, and tesafilmchen hasnt been online for so long that i dont know if he can help at all. i will post in the ladder news that we have gained momentum and that we might need assistance form some coders who might be able to help. sounds good?

[Yogibear]

now jet and Q are working on ladder stuff, and tesafilmchen hasnt been online for so long that i dont know if he can help at all. i will post in the ladder news that we have gained momentum and that we might need assistance form some coders who might be able to help. sounds good?

It would be excellent if you could provide us with some sort of priorities for the ladder support suggestions.

I reached for a low hanging fruit (which turned out to be not that low ) by implementing to set up games from a predefined profile: http://www.wesnoth.org/forum/viewtopic.php?f=36&t=29735. This is already working, i just need to provide a new release for you, which i hope i will be able to soon.

But of course i don't know if that is the feature which benefits you most atm. So if you could make up a list of what is most wanted, that would be very nice.

[Rigor]

good news from the rating system front: http://www.wesnoth.org/forum/viewtopic. ... 14#p465114 but most of u prolly know about it already plz answer something this guy is really nice to us, we can have the code of his rating program for free.

[sidzej]

I'm kind of scares of this proposition so I will add my subjective opinion into this topic.

First of all, the question is WHY to add any ranking system to the wesnoth lobby information about players? There were arguments here that it shold be done becouse:

• 1. To not needing to go to external website to check someone.

• 2. To not waste time with people that are not ladder players that joined the game.

• 3. To keep record about players to not have all those "no-names" all over everywhere.

• 4. To being able to know wich players wants to play what kind of games, so people can easier get together.

• 5. To separate noobs from veterans.

I think that the 4th argument is really cool and I like it. I can imagine that there can be some sort of "players wanting to play a game" window in wich players could add themselves and add a coment aobut what game they would want to play. After couple of weeks a special subculture language would be aslo naturally created with like "L 1300-1600, all" or "2 vs 2, Loris" so it would be really clear to read, easy to use and I think also very easy to implement since it would only need add, delete, edit for a single text for each player that want to use it and some refreshing actions to be able to wiev it by all others. I'm sure everyone would love this feature.

About 1 and 2 they look tempting and nice at first impression, but the problem is that they can't be done without 3. But will refer to 2 for a while. I think that it's good for ladder comunity that new players that don't know what (L) means in game titles can't play it. I know from my own experience when I started my MP carrier that I totally didn't know what it is and thought there are Legia Warszawa fans all over this game everywhere, but after getting some experience, becoming beter player, asking some people about it I finally discovered what it is and when I discovered it I was a player that had enough skills to at least get some 15 turns strugle againts ladder veterans. I don't think that ladder players so much desires ladder points that they want people who are new to game and don't know what that external site is to wasting their time to playing with them. It's good that people need to get familiar with MP first before even having a chance to participate in ladder, with a build in ranking system one would just "ohhh, there is a ranking here, lets play it!" and that would be bad for both sides (experienced player would have an uniteresting game for him/her and new player may feel like "omg, not game for me, I'm out from this game"). Couple words about 1 - it's would be good, but it can't be done witohut 3, and I'm totally against 3.

But before 3 I will talk about 5. I think that existing all palyers from new ones to dauntless ones in a single lobby space is a great thing that should not be destoryed under any circumstance. It makes being wesnoth community one wesnoth community, not slums wesnoth community, middle-class and noble community that do not have any contact with eachothers. It may be good if you are in the noble class, but if you are in the slums zone it will just make you becoming an evil person and you will mentally stay there forever, so it's bad idea in sense of new players - how can they become veterans when they are only delaing with other noobs all the time and don't have contact with all those soft, fresh, charming aristocratic babes?

But lets get to the main thing - the propably most evil thing of all this proposition - point no. 3. First it will really separates players, if everyone will know eachother ranking and full game record it will just psychologicly create and "I'm beter than you noob" atmoshere and will affect in fe. creating fake account to win against them to get beter record. And if some people will start doing it than those who don't do it will feal that's unfair, that they are working hard for their record and the others just get it without any problems. I know it's a bit of an extreme example, but I'm sure that there will be some crazy people that will do it and it will affect us all by craeting atmosphere that something is wrong here and unfair. External ladder site (despite I think is to strict in the map/eras rules and to much try to "measure the skill" as correctly as it's possible) is a very good thing becouse only people who wants to participate in it have contact with it and it do not affect all and is not realated with wesnoth client at any stage. The other problem about 3 is addopting keeping players record to the disconection mechanism in the game. Whne someone disconect you dont know he/she disconects becouse he/she left or is it a disconection. You can try to check if the return to the game will be made in an time range like 10 minutes for example, but then what if the return will be at difrent nick (that nick will have own record track and the old will just lost the game) or if it was coused by for example some unpredicted real life situation, and what if during that 10 minutes the one who not got disconected must also disconects?. You can also get and option whne somoeone quits "threat as defeat", that way everyone somoene will disconect the opponent can just win the game by using it. You also can make a mechanism that game must be reloaded and finished in 1 day - that's can be good idea, but how to done it? Doing hash from game status when someone disconected and checking if the hosted game match this hash so the server will be sure it's the same game? You can do that, but what if the movement bug occured and you need to rehost the game from previous save - you would need to keep milion of hashes of this game every action players will made to make it work properly and fair to match players with game to make sure all will be ok. You also can ingore disconection and just threat all disconetion as drops, not defeat - sounds great it everyone will stay to the end... but since you can just quit to not destroy your rnaking and change lost into drop than why not doing it. And at the end - what if server wil crush? Does players will need to host this game in 10 minutes after it will get restarded to not get double defeat for example? I'm not know the wesnoth project from the kitchen, but I'm aboslutlly sure making any ranking system that will keep record of the players that will be buil-in into current MP mechanism will be ultra buggy, very complicated and hard to make, unaccurate as hell and just bad at the end - external site rankign system is the best system for keeping track of players that can be done in this game, others will just get nailed and spiked by the real life gameplay, since you need to be really really lucky to have a game without a single disconection and there are just to many variants to implement it in a way to make it work without any user interaction (like reporting the game on an external site). In the 3. point I consider also achievements. I think achievements like are to much adictive mechanism, since whne you see an acheivement you want to achieve it becouse you have some purpose in life that need to be done. I consider it rather an MMO mechanism that do not well suits into the turn based strategy spirit, but in case of achievements I'm rather neutral so I don't think they are evil, but also I don't they are good, can be funny to have an "Die Hard Hero" achievement fe. (you won 20 games in wich your leader had 1hp at at least one moment), but having achievements like "You won 50 games" I think is not cool and it creates the same problems - how to check if game was won? So in conclusion achievements should rather be related with things that can be easilly to measure, like "Horse Slayer" (you killed 1000 mounted units in your MP carrer) or something similar - but in this case also appears the problem "what if the game will be rehosted from save? will the old kills will be deleted? and if yes than how the hell to implement that to do it corectly?!" and they also will be unacurate if the answer will be no.

And now couple of own thoughts about keeping record that I think can be done witohut problems and can bring good things while not bringing the bad. I would like to have a keeping track about players game history that would only contains two things: number of started games (Games rate) and number of players that think the player is a jerk (JERK rate). It would work in a simple way: when a player start a game that contains more than one players (had to be difrent humans not the same player) the Games rate of all players goes 1 up. By the game start I mean that game reached stage when there is turn 1 after the host clicked the start game button. Any disconection any actions that will be made during game will not couse any problems with this rate, it tells how many game that person started - easy to do, easy to control, almost bug impossible. The JERK rate however will be telling how many diffrent players threat this person as a jerk and thinks he is not worth to play with. To add a JERK rate someone would need to use the same window that is made for banning and click "Point a jerk" or somethign more diplomatic and the player would get a JERK rate increased by 1. It would work in a way that only registered players who registered at least 1 or more months ago can point a jerk and can point a jerk only once (you can't point a jerk multiple times, you always can do it only once per jerk). The JERK rate would made players that behave badly and quiting games to start all over from diffrent nicks since their JERK rate will fastly increase and noone will want t paly with them and would encourage everyone to behave correctly since you will have problems when you will be a jerk. It will also help no jerks to get information about who is a bad wesnoth experience and should be avoided. As I know from trying to host a 2 vs 2 it's really hard to host a game witohut quiters on maps other than Isars and that sort of info - a players with low percentage of started games / jerk rating ratio would be nice to make sure that that person is reliable enough and I can host a game without fearing the game will be stupidly destroy by "sorry, I have to go" at turn 5 or "faster, faster, faster! to slow, bye". This sort of keeping player record I think have a chance to eliminate bad people and avoid them while not having all no. 3 drawback.

PS. About thinks that need to be done in wesnoth I consider personally movement bug after disconections as a primary ultra urgent thing so before even trying to think about making something like built-in ranking system mechanisms you should concentrate on it and fix it somehow since it's the thing that really mess up the games and is extreamlly annoying.

[Gambit]

If you read the stickies in this forum you will note that ideas proposed here are not for mainline, but for an experimental branch of Wesnoth whose main goal is to implement features denied based on developer preference rather than merit or how much the players like them. Ideas such as this one.

If you have found a bug in the core game you should go to http://bugs.wesnoth.org to see if it has already been reported, and report it if it hasn't been.

[Greenace]

Buzzed though this thread, so I might have missed someone mentioning this but...

If there are 2v2 games you should be able to pick your ally, eh?

I have some people I met outside of wesnoth and we do games on here, we would be more willing to use the ladder competitive games if we could join and be on the same side. This lacks the negatives of letting you pick your enemy. Though frankly I have no idea how it would be done.

~~~~~~~~~~And~on~another~tack~~~~~~~~
I have played some other RTS' and in my opinion the ladder not only was a competitive feature but aided in finding people who were on your same level.

Several times in wesnoth I have put put up against people who were waaay better than me and died a fast death or people waaay worse and then felt bad .

So a display of your ladder status next to your name when in the non-ladder area would help you get matched with people who are on the same level.

[Scatha]

As I understand it, it would be helpful if there were concrete suggestions for how this would be implemented. Reading through the thread, it seems to be a bit light on concrete suggestions. Therefore I'm going to make one, and try to describe how I think the behaviour might most sensibly work. Even if other people think my suggestion is imperfect, hopefully having something precise will allow them to say exactly how they think it should function differently.

On logging onto the server, players would be presented with the 'casual room'. There would be a button to switch to the 'competitive room'.

Casual Room

This would essentially be the same as the current multiplayer lobby.
- [Optional]: It could be split into subrooms, e.g. '1v1', '2v2', 'Survivals', 'RPGs', to make it easier to find the type of game you're interested in. In this case it would be entirely down to the players to create games in the appropriate rooms.
- [Optional]: As per Greenace's description above, it could display a 'ladder level' next to players. This should probably not be complete information, but rather a rough indication of how well they are rated by the ladder system, including some kind of fading in for confidence (if you've played no competitive games, it doesn't show a rating; if you've played a few the system has a guess as to your ability and displays your rating faded; if you've played a lot the system has a better idea and displays your rating).

Competitive Room

The lobby for this room would display current competitive games available for kibitzing. It would have three (or possibly more -- see below) buttons for setting up games, as well as a *preferences* button.
The three buttons would be: 1v1, 2v2 assigned partner, and 2v2 chosen partner.

Clicking the 1v1 button would begin a search for other players looking for a compatible game. What is compatible would depend on your ELO rating (or other rating system if so chosen -- someone who understands ELO better than I do should have some say over the way the numbers are used here), as well as the preferences you have selected via the preferences button. The game would initially look for someone within 20 (??) ELO points, and if it didn't find anyone within a certain amount of time (20 seconds?) widen this to 50, then 100, then 200, then 300 points, in intervals of the same length.

Preferences

The "preferences" button would allow you to select:
- which of the default 1v1 maps you were happy to play
- which factions you were happy to play, or to choose a leader
- how to balance finding a match close in skill versus finding a match quickly. This would affect the '20 seconds?' above.
- restrictions on who you're happy to play against. e.g.:
- only play against opponents who have more than one map they are happy with
- only play against opponents who randomise between all factions
- only play against opponents who randomise between at least [n] factions
- only play against opponents who choose their leader

In Game, 1v1

Upon launching you into a game, the game shows you the leader the opponent chose, if any, or a list of the factions they randomised from. Games use default map settings, and have a timer (set to a standardised value ... if the player base becomes large enough there could be a 'no timer' option in the preferences).

In game, there should be an 'offer concession' button (with a confirmation box! You don't want to do this accidentally). [Optional: allow the other player to decline the concession] The remaining player should be able, if they so desire, to continue playing against the AI, but this should not be default option.

In the event of a player dropping, in the first instance they should be allowed 2 minutes to reconnect. If they do, great. If not, the other player should be given the option of reporting the game as a win, loss, or inconclusive (draw). The remaining player can just wait without selecting anything; if the other player rejoins the option box goes away.

Optional and desirable -- have a 'save to continue later' option. This prompts the other player to agree; if they don't, nothing happens. If saved the game can be continued by a mechanic like the 'challenge', below.

Upon resolution of the game, the players' ratings are adjusted.

Optional: Like the current ladder, there could be an opportunity to rate your opponent's sportsmanship. This is desirable, in that it provides extra incentives for people to behave nicely to one another. There could be an option in game preferences to only accept opponents with a sufficiently high sportsmanship rating. There may be some worry that this could be divisive; personally I think it would be of more help than harm.

2v2 Games

For automated 2v2 games, the procedure is much like 1v1. You can choose which maps you are happy to play, with which factions, etc. The matching algorithm tries to make the two sides as even as possible. If one player drops or concedes, their teammate gets control of their side.

To play with a teammate of your choice, you issue an invitation by typing their username. They get a confirmation box saying asking them to accept. If they accept, you are placed in the queue looking for a game as a pair (with map settings based on the asker's preferences). You can be matched against two people who were assigned together rather than chose teammates.

Challenges

[Optional]
To allow players to set up their own 1v1 ranked games, I propose a procedure similar to allowing a 2v2 game with a partner of your choice: you issue a challenge by typing the name of the person you wish to play; they then get a confirmation box.

Apologies if I misunderstood and that wasn't what was wanted at all ...

[chains]

In the hopes of sparking someone's interest. Here's one thing that both Wesnoth and the ladder need:

A game result screen.

When a game is over. The winner should be determined and a summary should display that shows who won. It would be neat if the screen had some cool art on it. It could even tell you how long the match took or how many units you produced, killed, and lost.

Some questions that need answers for this summary to be implemented

1. What happens when all players, but one disconnect. Can the last player end the game or must they play out the game vs AI? I suggest that the last team in a game be able to end the game if they choose to.

2. But, what if a player is only momentarily disconnected? There should be a window of time for a player to get disconnected and come back. Say, after 2 minutes, the player still in the game has the OPTION to press end game, if they choose.

This feature would increase the *YES!* factor when you win a game. It also happens to pick off one of the needed features for the ladder. (If this currently exists in game, GREAT. I haven't played video games in a while.)

[Bast]

Giving such a description of the wesnoth ladder, like such a success, without explaining the bad things the ladder brings to wesnoth is in my opinion not enough to take a good decision.

Besides, knowing the positives & the negatives points brings me to the same conclusion.
I think that we lost in wesnoth, with the ladder, an huge social part of the game.

Before the ladder, in the old days, it was necessary to speak with others to find good games. Random opponents don't allow it.
It was automatic for the players to join & obs games where they had friends or with the experienced players. At least that's what i did, and that's how i've begun to know friendly & good players.
When i was obs a game, i can meet other observers & arrange another game, or just wait for the one i am observing to finish & take part in the game that would follow which could become a 1v1, a 2v2, a 3v3 or even a 4v4 if people had the time to play it.

After the ladder, most experienced players begun to launch a ladder game when entering wesnoth. Some of us did refused to suscribe, others did, well you know the story. With the time, we finally all suscribed to ladder, even being forced to if we wanted to had the chance to play a good game.
The ladder did bring to the game another level for the 1v1 games, but it did killed an huge amount of arranged games, which means killing most 2v2 & 3v3 good games in the lobby.
This is actually why i stopped playing wesnoth for a while: because it became impossible, or really hard, to play good games that are not ladder. But ladder is just about 1v1 and wesnoth is more than just 1v1.

I'm not saying this as a complain against the ladder system. It is for me some facts that i wanted to share.
And on the other hand, i'm thinking that the only way for me to find as before, every day, good 2v2 games, maybe to implement a rating in game that would allow to create 2v2 (or 3v3) only for players that are beyond a rating X. Or an ingame 2v2 ladder (got some hope from ladder 8 to that).
Saying this, i remember the old debate on this forum against the ladder system, against players ratings, etc...
I do play for pleasure, i'm not seeking any rating, but i do want to play interesting games with players that are not too far from my level.

I don't actually have a real solution for this problem. But i'm pretty sure the problem isn't the lack of motivations of experienced players to play togethers things that are not 1v1 ladder...
So any solution is much needed.

My 2 cents.