EPTYou are only telling, that there are sooooooooo many problems. Why won't you just list them? Mca asks for that every time (btw. he has pm'ed me to ask you just minute before) and he is trying to repair it (if it's really problem with code) or make it clear, why it's not working on particular user's machine.
[D] New BW Server - Page 36
| Forum Index > BW General |
|
radley
Poland582 Posts
You are only telling, that there are sooooooooo many problems. Why won't you just list them? Mca asks for that every time (btw. he has pm'ed me to ask you just minute before) and he is trying to repair it (if it's really problem with code) or make it clear, why it's not working on particular user's machine. | ||
|
GeckoXp
Germany2016 Posts
On July 15 2013 07:25 radley wrote: What problems? You know that when you decrease mca launcher to chaos launcher functions, there shouldn't be any problems. Cause there is hardly no difference for system you are using. Mca has reasonable problems with special functions, like UPnP, streaming and injecting other users plugins. Especial with UPnP, cause this functions won't work with anyone. It depends on that, if this function is set on your router, which of course most of users don't know and blaim it on launcher. I have also seen several people claiming UPnP is not working for them. Next fake opinion and again lack of their knowledge - many people can't make because of their internet providers, which doesn't support for them port forwarding on they network infrastructure. UPnP function is only for people, who can make games after proper port forwading but are not good enough to do it manually on their router. You are only telling, that there are sooooooooo many problems. Why won't you just list them? Mca asks for that every time (btw. he has pm'ed me to ask you just minute before) and he is trying to repair it (if it's really problem with code) or make it clear, why it's not working on particular user's machine. You're talking with a brick wall here. I remember this guy coming up and asking for instructions on ICC multiple times, which he received, but did not understand, even screen shots kind of overwhelmed him. Don't expect him to list anything, but prepare to get flamed for your efforts to improve the project, he has a habit of doing that. This will most likely include something like "you are an imbecile and should die". Welcome to the foreign BW world. | ||
|
radley
Poland582 Posts
Edit: BTW, recent option added by mca -> authomatic login into iccup/fish: | ||
|
Magreidis
Lithuania146 Posts
| ||
|
bITt.mAN
Switzerland3693 Posts
But think about the consequences of installing a penalty for leaving an auto matchmaking lobby. Especially in the earlier stages there won't be many people online, and if you have your reasons for not wanting a rematch I don't like the idea that you're forced into it (or pay a penalty). What's the good reason for having such a penalty? It's like a Primary School rap on the knuckles "you should be a big boy because it's healthy for you". If I want to cheese and dodge Protoss [scum], yes I won't be a grown up l2p player, but it's my choice to make, and it should remain my choice. You can't take my rights bro, I thought this was a free country! So you could either drop the penalty idea, or if you really had to, implement some matchup filters. I've inputted my main race into Launcher (T), if someone else does so as well (P), but I haven't checked the TvP preferred matchup option, they will be a low priority opponent, who I'd rather wait 3 minutes to dodge than play instantly. Another fun bit of statistical analysis you could do is real time estimate of queues. Data data everywhere, use it! If there are only Protoss available, but I want to play Terrans (assuming these preferences are implemented), how long should I wait before you pair me up with a Protoss and tell me to get on with it? If the only other Terran online is currently in a game, from harvested statistics you could project my estimated waiting time and ask me if I want to broaden the search (ok this might start getting a bit hard to implement, but it'd be awesome if I just wanted to find TvT lovers). Generate an estimate of how long his game will be, adjust for how long ago it started, and tell me how long untill you think he'll be available, using the following statistical variables: Average game duration based on matchup Average game duration for my current skill level Average game duration modulated by map choice Decreasing chance he'll play another if he's already been logged on for a long time (I'm sure the server logs what time he logged in, you've got that data floating around somewhere) This data can be harvested for all ladder games, and the estimates will automatically get better over time. | ||
|
Birdie
New Zealand4438 Posts
We plan on having lots of statistics  the estimated time will probably be shown, and over time the search will broaden too, which is a necessity for high level players finding games. | ||
|
bITt.mAN
Switzerland3693 Posts
To a certain extent this is a continuation of my desire for TvT and disregard for TvP, into the auto-matchmaking BW era. Matchmaking system proposed improvement [Patent Pending©®™☻] ^^ See some days I wish I had chosen to study computer science, because I have no formal training in approaching this optimization problem, but I really enjoy it. So let's have a general (likely newb-level) discussion of stuff you programmers already have the answer to, but us other guys would like to add their two cents. The million-ladder-point-question is: "Match now with imperfect opponent, or wait for a better option?". In the ideal world I would be happy with whoever is immediately available to play. But actually y'know what, I'd rather not play someone too different in skill to me, I'd rather not play on FS (implement veto map list!), [at this point the matchmaking capability is about equivalent to SC2's] and I'd certainly not rather play against Protoss. I'd be prepared to wait a bit to play versus a different race, so please Mr. Server, crunch some numbers and line up my ideal opponent for me! Matchup Priority Preferences Select Preferred Opponent Race [Simplest Implementation] 4 Checkboxes (plus 4 already for my default race) I'd check T, Z, R, and leave P blank. Then the system would know to place Protoss as lowest priority when searching for a new opponent. It's 8 checkboxes that let me put in a bit more data, that would really help land people with their preferred matchups (or if they specifically want to train a MU e.g. to get more comfortable I may pick only P). In this way you could make a better choice for the decision. You could stop here. Scale of 0 to 1 For Each MU [Has Drawbacks] A series of checkboxes with all 16 (including R) matchups. e.g. TvT 1, TvZ 1, ZvZ 1, TvP 0, all else=0 In this way the auto matchmaking would preferentially allocate me to a T or Z opponent with equal weighting, and only if I say "FINE I'LL DO IT" I'd take it in the Protoss. Scale of 0 to X For Each MU [Best] It'd be nicer if you could rate them on a scale of 0-2 or 0-4. 0 being NO WAY IN HELL (probability = 0) which is less ambiguous (thus more international) than a 1-5 scale. Implement by "Matchup" and a number box next to each, limited to the set {0,1,2,3,4} or if you wanted to be fancy, a little drop-down menu, which would make it more intuitive, thus easier for weak-English users. So if I select TvT 4, ZvZ 3, TvZ 2, TvP 1, TvR 1, and the rest (e.g. RvX) to 0, you'd be able to make an even more informed decision! This is the best as it allows for offrace MUs. With the others I'd be at most be landed 50/50 ZvZ or TvZ (with the first only TvX), but with more info I can be placed with the guys who'd be cool with ZvZ rather than those who wanna avoid it (who are the majority). You would also empower the tiny subset of the least popular race who actually want to TvT to hook up! mmmmmh | ||
|
bITt.mAN
Switzerland3693 Posts
If you implement MU requests in auto-matchmaking it'd just pair you up with a suitable partner. But it'd make it even better if your respective races were automatically selected (but not fixed), so that you don't have to dick around trying to negotiate which MU to play. You could not even bother making a time-consuming (pretty) GUI MU-preference interface, just input the details through command prompt parameters, or even by using certain syntax markers in your B.Net profile details, if accessing that can be automated e.g. + Show Spoiler + Type things in this format somewhere in your profile description: <insert whatever message you want>or <insert whatever message you want> Define the "/vs" command as a request for an auto-matchmade game. Make it work by default, but let more advanced users add some parameters to customise the search to their liking. This way, you don't even need to bother making a time-consuming (pretty) GUI MU preference interface. I have no idea how you'd do the auto matchmaking, and you're probably way ahead on this than me, but please make the command convenient and quick to type (so not "RankedMatch" or "GetMatch" but "vs" or something short). So now, syntax (again lol who am I to talk about this I'm no programmer). There are three options A. Define your own race, then define your desired opponent race e.g. "/setRace T" "/vs vT=4, vZ=2, vR=1, P=0" //numbers are weightings, you're prepared to wait longer the bigger the number is. 0 should be "absolutely never", if you're sick of it but would still play a certain MU, put a 1 instead. This would work if we all promised to stick to our set main race, cuz you wouldn't have to input preferences for each and every possible MU. It'd get problems if say, I wanted to prioritize ZvZ over TvZ but my main is T. B. Define individual MU preferences. "/vs TvT=4, ZvZ=3, TvZ=2, else=1" Would pair me with someone who typed anything that included vT, or vZ as long as they don't put ZvZ=0. If none of them are available the "else" would send me to anyone left. C. Combination of Both Define your own race, but still leave the option if you've got a pet MU you'd want to prioritize e.g. "/setRace Z" "/vs TvT=4, ZvZ=3, ZvT=2, else=1" Would grab any TvT candidate, but otherwise tell people that I'm a Zerg. So in this scenario, if I input "/setRace T" "/vs TvT=4, ZvZ=3, TvZ=2, TvP=0, else=1" A Protoss is immediately available, but I don't want him. Then a Zerg becomes available, but he only wants ZvP, he wouldn't go with me (unless I wait long enough to roll PvZ). Finally a Terran comes, but he's like a normal foreigner Terran and skips the matchup, breaking my heart (his preferences would exclude TvT), so I'd be put against him in a ZvT. Without this system, I'd join Protoss, waste his time by leaving, then join Zerg, try to argue him into playing a MU he doesn't want to play (probably insulting him and yelling l2p), and finally join a Terran and have him switch to an offrace or flat out leave. | ||
|
bITt.mAN
Switzerland3693 Posts
If my goal is deciding if I should wait for a potential TvT to pop up, useful statistics (which are easy to do) could be accessed by "/MUstats TvT 2" where 2 is how far back in time I want the statistics to roll, in hours. It'd give me the number of TvTs played, #TvTs as % of total games, average duration, most used maps, rank distribution of who plays TvT, and maybe even list of people currently available to play TvT. You could also allow parameters for me to limit the statistics to people in my skill-range. Statistics on average game duration based on map, MU, skill level. Also distribution of Races used, MUs played, and races set as 'main'. All sorts of fun stuff that's somewhat available on iCCup, but you really have to dig for it, and it has no way to define your sample period. You could even have things about "average gameplay time of PvT where Z wins" which would be different to that of P winning the same matchup. Finally, if everyone inputted their MU availability, it'd make for great real-time statistics. How many people are there online who would be willing to play TvT, limit it to within my skill range, how many are available right now, how many do you think will be available in 5 minutes. Combining user inputted MU preferences and statistics can let you answer the question: if no one is currently availabe, how long untill at least one eligible partner is expected to win or lose? How about at least 3, 5 etc. This would be so cool. Now I know most people don't care that much about MU picking, but they do care about sparse population in the higher levels. So, apply the same type of analysis to answer questions about: tell me how many people there are online within my level/an arbitrairy range, what are their races, how long till they're available. Of course you'd auto-display win-loss records as well as points when the game starts, but if you wanted you could eaisly do win-loss per MU, as well as %MUs played (top 3). If you didn't want to have to automate so much, while in a game, you could type a command to tell the server your preferences for the next map (e.g. I'm in a TvP and I want to stick to that MU, I tell the server something to do with TvP, and it flags me as available for a PvT opponent), that way you could arrange your next match real-time and cut out wait time altogether. So all in all, that's me, I'm done, maybe some more ideas later but I'm spent for now! EDIT: in case you're curious, that's 2200+ words I should have spent writing my work/project instead <; | ||
|
hp.Shell
United States2527 Posts
| ||
|
thezanursic
5500 Posts
If you do so please make a new thread for it so that it could get featured and get more attention! Keep it up! | ||
|
9-BiT
United States1089 Posts
| ||
|
Birdie
New Zealand4438 Posts
On July 24 2013 15:35 9-BiT wrote: Once you "release" the server, how hard is it to implement changes? Because lots of people are talking about nice features, but I would rather just have a good server first, and then add features as you can. Well, if we release it as it is now then it's functional but so featureless that there'd be no reason for people to switch other than novelty. If we release with a bunch of good features, there's a higher chance people will switch over and stay. | ||
|
thezanursic
5500 Posts
On July 24 2013 15:35 9-BiT wrote: Once you "release" the server, how hard is it to implement changes? Because lots of people are talking about nice features, but I would rather just have a good server first, and then add features as you can. If it has a lot of really interesting changes the server would be markatable. I'm pretty sure that a lot of inactive BW players would try out a server with a well made automatch making system and similar features. If they release it before it is ready. People will just flock to the server, see that it's not well populated yet and has no interesting features and leave and not give it another chance. | ||
|
Magreidis
Lithuania146 Posts
 , but still i wish you best of luck! One more thing, maybe you could make showcase of the server? That would be awesome, and some people would say if this server is worth it or its just another "crap". | ||
|
TehRei
Sweden261 Posts
On July 24 2013 21:58 Magreidis wrote: Birdie, do you have hosting server that you would like to have? Because if people will flood it, and the server will be slow and stuff? What then, GG? And you need to get a good support for server that might be a problem , but still i wish you best of luck! One more thing, maybe you could make showcase of the server? That would be awesome, and some people would say if this server is worth it or its just another "crap". Shouldn't be a huge problem. All games in BW are played peer-to-peer, with little to no interaction with the server. My understanding is that the server only has to handle chatrooms and such features, which, logically speaking, shouldn't be bandwidth-intensive at all. I don't know how matchmaking is handled, but I guess that might place some additional load on the server. I highly doubt that it'd be enough to warrant worrying about it though  Also, when the server is slightly closer to launching, I sincerely hope that they will showcase features and have a bunch of hype around it. Maybe have the first "official" game played after launch be a showmatch between two awesome and well-known players? Or something along those lines... But until the server is closer to launch I think it would be best for them to focus only on implementing features, hunting down bugs and getting everything up and running. | ||
|
YangJia
United States54 Posts
Just my observations, ironic such passionate BW players are going to expedite the death of their game in the foreign scene. | ||
|
Sum41
Chile345 Posts
| ||
|
Birdie
New Zealand4438 Posts
On July 24 2013 21:58 Magreidis wrote: Birdie, do you have hosting server that you would like to have? Because if people will flood it, and the server will be slow and stuff? What then, GG? And you need to get a good support for server that might be a problem , but still i wish you best of luck! One more thing, maybe you could make showcase of the server? That would be awesome, and some people would say if this server is worth it or its just another "crap". We have a hosting server, that's not a problem. On July 25 2013 04:02 Sum41 wrote: this idea of a new server sounds awesome, when will it be available? Soon(TM). Somewhere between September this year and December...next year I'm inclined to be optimistic and say well before December this year but given that we are technically amateurs working in our part time instead of full time professionals, it isn't as simple as 1 2 3 done. But our progress is good enough that this year is entirely realistic, between September and December. | ||
|
Atlasy
Hungary229 Posts
Is there any chance of the Koreans switching over? I mean if the server could be populated with them then that would be great. | ||
| ||
