[JASS] retrieve ip address?

[normalice]

Hi. I was wondering if there is a way to retrieve IP addresses using jass.

 

[Dr Super Good]

No there is no way to retrieve IP address using JASS without modifying WC3.

Why do you need the IP addresses? WC3 resolves the connection pretty well and if you are in game then a path to and from the host already exists.

 

[normalice]

i wanted to write a script that detects "preteamers" in ffa games. Determining if two players have the same ip address would be the most straightforward way.

 

[Daffa]

SharpCraft might have this feature, else it's impossible without modifying WC III

 

[Dr Super Good]

i wanted to write a script that detects "preteamers" in ffa games. Determining if two players have the same ip address would be the most straightforward way.

Except two players never have the same IP address. They may be behind a NAT making them appear to have the same IP address but both will still have different local IP addresses.

Also who says that same NAT people are pre-teamers? I mean it could be a LAN party so you would truly ruin your map for some people (not good design choice).

 

[normalice]

Except two players never have the same IP address. They may be behind a NAT making them appear to have the same IP address but both will still have different local IP addresses.

Also who says that same NAT people are pre-teamers? I mean it could be a LAN party so you would truly ruin your map for some people (not good design choice).

in the vast majority of cases, people with the same NAT turn out to be preteamers. It may ruin the game for a very few who are not, but that's better than ruining the much more abundant games that are ruined when there are preteamers present.

anyway, i would plan on integrating other checks, but this would give me a starting point.

but if it can't be done, then nevermind...

 

[Dr Super Good]

in the vast majority of cases, people with the same NAT turn out to be preteamers. It may ruin the game for a very few who are not, but that's better than ruining the much more abundant games that are ruined when there are preteamers present.

Then clearly your map has problems. I would advise designing the map such that pre-teamers are not such a problem. I mean if someone does strategic targeting (attacking the same target someone else is) does that make him a pre-teamer? It sounds like you are far too paranoid. In real life enemies will team up against a common cause, even if they eventually need to fight each other. This is the essence of FFA and why you have so few FFAs in tournament level (as the fights cannot be easily and reliably fixed while still appearing fair).

anyway, i would plan on integrating other checks, but this would give me a starting point.

I doubt you have the AI knowledge to come up with reliable other checks since you have little understanding of IP networks. I would advise just allowing pre-teaming and warning people that this might be the case and that they should not burn bridges with randoms and be willing to work together against teamers. I mean if it is a 12 player FFA then if two people team up it should hardly be a problem as one can quickly notice the teaming up and you can team up with some other players against them Hunger Games style.

 

[normalice]

Then clearly your map has problems. I would advise designing the map such that pre-teamers are not such a problem. I mean if someone does strategic targeting (attacking the same target someone else is) does that make him a pre-teamer? It sounds like you are far too paranoid. In real life enemies will team up against a common cause, even if they eventually need to fight each other. This is the essence of FFA and why you have so few FFAs in tournament level (as the fights cannot be easily and reliably fixed while still appearing fair).


I doubt you have the AI knowledge to come up with reliable other checks since you have little understanding of IP networks. I would advise just allowing pre-teaming and warning people that this might be the case and that they should not burn bridges with randoms and be willing to work together against teamers. I mean if it is a 12 player FFA then if two people team up it should hardly be a problem as one can quickly notice the teaming up and you can team up with some other players against them Hunger Games style.

i think perhaps I'm not explaining the severity of the problem properly...

 

[Dr Super Good]

i think perhaps I'm not explaining the severity of the problem properly...

What problem? I mean this is just some WC3 custom map, hardly the World Cup. So what if some people play it wrong, the host can always kick them next session if they do not like them (like they used to do to anyone who won the game instead of them) and they can always host their own games if they find the host being unfair.

 

[normalice]

What problem? I mean this is just some WC3 custom map, hardly the World Cup. So what if some people play it wrong, the host can always kick them next session if they do not like them (like they used to do to anyone who won the game instead of them) and they can always host their own games if they find the host being unfair.

yeah, i failed to explain it.

but thanks for answering my question.

 

[Dr Super Good]

I wonder if it would be possible to retrieve the IP address of a client using the Preload bug... Obviously this might be patched (fixed) in the future as it is a major WC3 security flaw but I wonder if it can do more stuff than just file I/O.

 

[sethmachine]

I doubt you have the AI knowledge to come up with reliable other checks since you have little understanding of IP networks.

I did not know these correlated.

Your best bet would be to simply keep your own blacklist of names/addresses and simply ban those players who offended in the past. This would only work for when you hosted (or someone who shared your list). I believe the host has every right to determine if a player can play (since they are hosting) by kicking them from the lobby. Most such games have dedicated host bots which enforce the ban lists.

A little aside but relevant I think
I once read this paper which tried to explain why practically all competitive sports and video games are either 1v1 (single players or teams) or a test of some attribute irrelevant of what your opponent does (e.g. Track: speed). The idea is that any game which involves more than two sides (i.e. an FFA) boils down to a game of politics: convincing players not to attack you and instead to fight each other, or do whatever is best for you. Kingmaking then becomes a huge issue-a player can intentionally lose but take down another and giving the game to the unaffected player, effectively removing any competitive or skill based value to the game, and hence why no such games are mainstream or done at a professional level.

I play large FFAs (10-12 players) in Broken Alliances, a custom map with a large community (BrokenAlliances.com), quite often, and some simple rules we adopts are:

1. no information sharing (players can chat, but cannot say ANYTHING that gives information about their status, e.g. "I'm dying" or "Red is attacking me!" or "He is using these units! (Go counter him now that I told you." or "His base is located here!"). The only information a player can get besides observing it themselves is when a player leaves (thus they were killed off).

2. no collusion (or pre-teaming as you call it). Strictly we have no ways of detecting this or stopping it. Instead people send replays to the community and ask to ban players believed to be teaming. They are then banned from the bots.

What I am suggesting is, is that you'll need a dedicated community in order to enforce your ways of defining how the game should be played. BrokenAlliances has done this well--I very rarely face people openly teaming as the community does a good job of sticking to this code, and those that do are soon prevented from playing again.

However, I would be interested to hear a way to detect if players are in collusion by extracting information from the game state or their behavior, if you could elaborate Dr.

 

[Dr Super Good]

I did not know these correlated.

You learn about both in any decent computer science course at a University.

NATs are used mainly because IPv4 is a dumb standard that has run out of addresses. In fact some times providers such as BT use it since they do not have enough IP addresses to give their customers (usually they use provider level NAT for the cheapest services).

One could say that if people are on the same NAT then they will work together under the assumption they know each other. But maybe they will work together the same as two random public players. It is also not entirely fair to say they know each other as they might just be sharing a free wifi hotspot or an internet café (public places that use NAT). You would need to conduct some form of statistical test to see if it really is the case and how strong the correlation of same public IP and teaming up unnecessarily is. Even if the correlation is strong, this still is just one parameter your Bayesian decision network would use to come to the conclusion that they are working together.

Your best bet would be to simply keep your own blacklist of names/addresses and simply ban those players who offended in the past. This would only work for when you hosted (or someone who shared your list). I believe the host has every right to determine if a player can play (since they are hosting) by kicking them from the lobby. Most such games have dedicated host bots which enforce the ban lists.

Exactly, this is the "everyone is equal" approach. You can be your own administrator and if you do not like someone elses decision you can just host yourself. Obviously this is not at all just since most people tend to let the power to their head or make up accusations with little backing but then again is that wrong?

The idea is that any game which involves more than two sides (i.e. an FFA) boils down to a game of politics: convincing players not to attack you and instead to fight each other, or do whatever is best for you. Kingmaking then becomes a huge issue-a player can intentionally lose but take down another and giving the game to the unaffected player, effectively removing any competitive or skill based value to the game, and hence why no such games are mainstream or done at a professional level.

Exactly. One could argue that the politics itself is a sport and that is what many "diplo" maps do but the reality is that no mater how good you are you are doomed if you are playing with many players who have you as their worst enemy.

1. no information sharing (players can chat, but cannot say ANYTHING that gives information about their status, e.g. "I'm dying" or "Red is attacking me!" or "He is using these units! (Go counter him now that I told you." or "His base is located here!"). The only information a player can get besides observing it themselves is when a player leaves (thus they were killed off).

Completely arbitrary rule. It is like how the Pokémon community says X is banned or that some sport introduces a rule that you cannot kick the ball on the 11th minute of every hour for 15 seconds. It is basically trying to tame something that cannot be tamed by applying restrictions.

2. no collusion (or pre-teaming as you call it). Strictly we have no ways of detecting this or stopping it. Instead people send replays to the community and ask to ban players believed to be teaming. They are then banned from the bots.

Yeh I see this now. "I LOST DUE TO BAD LUCK! I KNOW I WILL REPORT THEM FOR TEAMING". The reality is that two players just happened to decide to attack at the same time and attack the same person because they noticed he was weak from an early scout or from a pervious attack. They then get banned despite not even knowing who each other were. Is this entirely just? I mean sure some are pretty obvious, like the blatant map hacker but not all are. Can you be entirely sure that two people pulled off a team effort or were they just unlucky and did something strategically brilliant. Or maybe two players are pulling off a team effort just they are so good at hiding it no one even suspects them. Again a pretty arbitrary rule trying to force broken game mechanics into some shape that people would like to call fair.

It might work but in reality it is trying to push a heavy weight up a hill, the weight always wins and rolls down. These rules are trying to justify some form of fairness in something that is inherently unfair. I mean the entire FFA game comes down to luck who wins as someone can lose purely that by chance everyone decides to attack them at once. Sure they could fight each other doing so but that would be tactical stupidity if you could rather kill off one competitor, unless you turn it to your advantage and use it to weakon them and kill the competitor but again that is all chance.

However, I would be interested to hear a way to detect if players are in collusion by extracting information from the game state or their behavior, if you could elaborate Dr.

One could build up strong evidence by their history. For example, if two+ people have been in sessions together a lot of times this points towards people working together. If two+ people have only been in sessions together and few not then this also points towards them working together. If two+ people happen to attack the same target or just happen to help save each other then this points towards them working together. If two+ players go complimenting builds this points towards them working together. If two+ players never engage each other in major fights or if they are noticeably underperforming when fighting each other then this points towards them working together. If two+ players APM are interrupted are around the same time in a send receive sequence (chat messages via BattleNet to each other?) this points towards them working together. If two+ players have any of the above several games in a row together then they are likely working together.

Put that all together with correct weightings and you can get a single value, the probability someone is working together. From this you can then make a decision to do whatever (kick, ban, shout, warn, humiliate, ignore whatever you want). However will it always be accurate? I doubt it. Some teamers could fool it by specifically performing to counter the metrics (doing the opposite of their measurement). Other players will just be playing normally and due to chance be given high correlations from this. Sure one can always use it as evidence and it may mostly work, but it is impossible to always work. From it you will have probabilities of false positive and false negative.

However this will likely be the same for any manual decisions. In fact it might be more accurate than manual decisions since unlike human choices which can be made pretty arbitrary ("I do not like X so he did it"), it does justify its choices (you can query how it come to such a decision).

 

[normalice]

alright, I'll just explain it.

so I frequently help a site playffa.net with scripting and coding, when I can. I am new to Jass and very new to vJass, but so far I've managed to do most of what they've asked. They do ban and 'blacklist' preteamers, maphackers, and such - but not only are there thousands of people who participate, but there is also a new group of three players that have managed to circumvent the banning and blacklisting through vpn and such. They tend to join games and team people out for no discernible reason. The maker of playffa.net is somewhat fruitlessly trying to negotiate with them, but what else can be said? they are crap players who apparently dropped a pretty penny on some program that allows them to cheat without consequence. Of course they are going to.

Anyway, since this doesn't seem doable on the server end, I wanted to attempt a map scripting solution. The easiest way to do this would be to determine if they originate from the same location. But as that is not feasible, I guess I'll have to do this with scripting, if at all..

 

[Dr Super Good]

I still think you people are getting to serious about a game. I mean it is only WC3 for crying out loud, hardly the game to decide the fate of the world with.

Pre-teamers would not make it to high level of play as you can ban them if you feel like it. Only the low level of play would be plagued with them but in the end who cares as people will get more fun there playing how they like than having arbitrary rules forced down their throats like some form of dictatorship.

 

[sethmachine]

Completely arbitrary rule. It is like how the Pokémon community says X is banned or that some sport introduces a rule that you cannot kick the ball on the 11th minute of every hour for 15 seconds. It is basically trying to tame something that cannot be tamed by applying restrictions.

Yes this is arbitrary. Sharing information is a powerful tool to manipulate people. It's less obvious in board games, where there's no fog of war.

Yeh I see this now. "I LOST DUE TO BAD LUCK! I KNOW I WILL REPORT THEM FOR TEAMING". The reality is that two players just happened to decide to attack at the same time and attack the same person because they noticed he was weak from an early scout or from a pervious attack. They then get banned despite not even knowing who each other were. Is this entirely just? I mean sure some are pretty obvious, like the blatant map hacker but not all are. Can you be entirely sure that two people pulled off a team effort or were they just unlucky and did something strategically brilliant. Or maybe two players are pulling off a team effort just they are so good at hiding it no one even suspects them. Again a pretty arbitrary rule trying to force broken game mechanics into some shape that people would like to call fair.

It might work but in reality it is trying to push a heavy weight up a hill, the weight always wins and rolls down. These rules are trying to justify some form of fairness in something that is inherently unfair. I mean the entire FFA game comes down to luck who wins as someone can lose purely that by chance everyone decides to attack them at once...

Power corrupts of course, but by and far there are few unjust bans in the community I mentioned, and people are always given 2nd and 3rd chances if they apologize/appeal. And every ban request has to be accompanied by a replay as well.

Luck I am sure has much to do with it, but when the same people win 12 player FFAs at rates much higher than chance, I would think it does speak to some skill or knowledge.

One could build up strong evidence by their history. For example, if two+ people have been in sessions together a lot of times this points towards people working together. If two+ people have only been in sessions together and few not then this also points towards them working together. If two+ people happen to attack the same target or just happen to help save each other then this points towards them working together. If two+ players go complimenting builds this points towards them working together. If two+ players never engage each other in major fights or if they are noticeably underperforming when fighting each other then this points towards them working together. If two+ players APM are interrupted are around the same time in a send receive sequence (chat messages via BattleNet to each other?) this points towards them working together. If two+ players have any of the above several games in a row together then they are likely working together.

I'm not familiar with Bayesian network learning, but I am familiar with Naive Bayes, HMMs, CRFs, maxent, etc. I am told feature selection is a dark art. It might help to have some (labeled) training data too. But in the end because of the subjective nature of it the classifier will depend on the definition that is asserted for this kind of label.

 

[normalice]

I still think you people are getting to serious about a game. I mean it is only WC3 for crying out loud, hardly the game to decide the fate of the world with.

I think you may have mistakenly imagined my casual interest in doing this with perhaps some die-hard anger issues or something? I have no idea what else you could be talking about here...

anyway, why are you even on this site if your only role seems to be discouraging project ideas?

Pre-teamers would not make it to high level of play as you can ban them if you feel like it.

did you even read my previous post? They can't be banned.

Only the low level of play would be plagued with them but in the end who cares as people will get more fun there playing how they like than having arbitrary rules forced down their throats like some form of dictatorship.

getting pre-teamed isn't fun. and if you honestly believe that rules in a game that you willingly play is a 'dictatorship,' then you've clearly never lived under a dictatorship.

 

[Dr Super Good]

I'm not familiar with Bayesian network learning, but I am familiar with Naive Bayes, HMMs, CRFs, maxent, etc. I am told feature selection is a dark art. It might help to have some (labeled) training data too. But in the end because of the subjective nature of it the classifier will depend on the definition that is asserted for this kind of label.

Bayesian networks are basically decision trees of sorts. Their input is obviously the precepts from sensors and their output is actions to actuators. Unlike neural networks and other types of machine learning, Bayesian networks are much more scientific and logical. The main advantage they have is that once programed you can print a trace as to exactly why a certain decision was made (proof of correctness). The weights and operations used can be computed using probability theory during a learning phase. They are also pretty easy to compute as only nodes that have been exposed to change require updating.

did you even read my previous post? They can't be banned.

They can as high level play is all pre-arranged, there is no allowing of players who you do not know/trust in high level matches. If you are still in the "public game" phase then this kind of stuff is common and you just need to be put up with it. Heck random trolls PK you in coop maps, I do not really see this any different from just generic map trolling that happens.

getting pre-teamed isn't fun.

Neither is losing. It is fun to pre-team as it gives you that edge and is almost like cheating without cheating. It is even more fun killing a pre-team group with randoms because you noticed they teamed up, warned the others and everyone ganged on them. Now obviously with silly rules like "no talking" this cannot happen that easily.

and if you honestly believe that rules in a game that you willingly play is a 'dictatorship,' then you've clearly never lived under a dictatorship.

Some random person decides it would be cool to add some pretty random rule just because he feels bad for losing? Sounds like a dictatorship to me. Even if a group of random people mutually decide, it is still as bad as FIFA or UAFA randomly suspending/fining football players because they took offense to some arrangement of words or do not think the players are Christian enough.

In my opinion cheating is only going out of the way to modify a game such that you have an unfair advantage over others. An example is a map hack or a third party tool that exploits a bug that cannot be exploited using the game alone. The only rules I have in maps are either arbitrary ones which I decide before a session (and if they are violated I just terminate the guys connection and drop him, I am host after all and he did not have to play with me if he did not want to) or leave and make another game. Obviously I avoid PvP games as this sort of behaviour would really annoy some and how can a host be non-biased with the power to kick anyone at any time in a PvP game?

 

[sethmachine]

Dr. Super Good is right about standards in high level matches. Players who are not known are not allowed in them. Unless these trolls that evade banning are also secretly other known players, they won't matter in the actual competitions, since they have to have a known name or identity to join, and pre-teaming is not a good way at getting a reputation.

People who 'cheat' (hack or otherwise spit in the accepted rules of the game) have a habit of changing their account names frequently and never playing under the same name long. They can't because if they get known nobody will allow them to play, or at the very least, refuse to play with them. So in the big picture these guys wont really matter, and as Dr. Super Good said, it's just another form of trolling which is going to happen in any public game.

So it's hardly worth investing the time to come up with ways to detect it, and also I think it is a mistake to recognize these guys and try to negotiate with them. It gives them power, which is the last thing to ever give a troll. The best way is to not feed them at all, because what they want most is a reaction.

 

[Ceday]

Can't host bots retrieve ip address of a player?This ffa games were hosting by their bot, I never studied them but your chance seems higher than jass.

 

[normalice]

They can as high level play is all pre-arranged, there is no allowing of players who you do not know/trust in high level matches.

on this bot it is not. there is a showcase game hosted every once in a while, but it is fairly hands free any other time.

Neither is losing. It is fun to pre-team as it gives you that edge and is almost like cheating without cheating.

but it is cheating. There's a rulebook for the bot ladder that specifically classifies this as cheating.

Some random person decides it would be cool to add some pretty random rule just because he feels bad for losing?

it's a pretty longstanding and reasonable rule, actually... Again, you have no concept of what a dictatorship is. I invite you to move to north korea for a while and find out.

Obviously I avoid PvP games as this sort of behaviour would really annoy some and how can a host be non-biased with the power to kick anyone at any time in a PvP game?

it's quite simple, actually: don't behave like a child. Well, simple for me anyway... I can see how the concept of others not behaving like a child might be a bit beyond your comprehension. I mean, if your experience with dictatorships is basically limited to a game you voluntarily play having rules that define it as unfair for you to win with via means that are, in fact, unfair, your experience of this world isn't exactly vast.. It's as if you don't understand that, unlike in a real dictatorship, you are completely free to start your own games and pre-team people in them to your heart's content.

 

[Dr Super Good]

but it is cheating. There's a rulebook for the bot ladder that specifically classifies this as cheating.

Again it is enforcing an arbitrary rule with little enforcement. In my definition of cheating it is not since the game is behaving perfectly and no third party tool are being used.

on this bot it is not. there is a showcase game hosted every once in a while, but it is fairly hands free any other time.

Most people you play with you should already know as WC3 is so old only people playing it are old people who never moved on. The rest are so uncommon they might play a session or two and ruin your fun but then they will give up.

it's a pretty longstanding and reasonable rule, actually...

So is a rule that I always win. I am the host after all and am spending extra bandwidth so you can play so why should I not always be a winner? Half the fun from FFA is the silly diplomacy, why else do you think so many people read Hunger Games and other books (which I do not as they are boring).

Again, you have no concept of what a dictatorship is. I invite you to move to north korea for a while and find out.

Well if you are in North Korea I am amazed you can even write these messages without being oppressed. Clearly it must be more free than you let on.

North Korea is the example of what is going on. "Everyone must have my hair cut!" is just like "Everyone must not speak as it ruins my fun!". You see it is pretty arbitrary and can be made up by an individual even if a group approves.

don't behave like a child

That is surprisingly hard to do when one is 12-14 since you are... um... a child. Warcraft III is an old game. Heck politicians are still children and they are in charge of the country.

I can see how the concept of others not behaving like a child might be a bit beyond your comprehension.

Yes behaving not as a child as a child would be beyond my comprehensions seeing how I was a child. WC3 is an old game. In fact soon WC3 will be 10 years old. Judging by the fact that most people who are making WC3 maps cannot even do simple vector manipulation I am guessing most people still are children playing WC3 (just a few generations younger than me).

I mean, if your experience with dictatorships is basically limited to a game you voluntarily play having rules that define it as unfair for you to win with via means that are, in fact, unfair, your experience of this world isn't exactly vast.. It's as if you don't understand that, unlike in a real dictatorship, you are completely free to start your own games and pre-team people in them to your heart's content.

That is what I used to do. Was so much fun and the same haters would keep coming back because they were too stupid to configure their NAT to host. No bots to compete with either and banlists and map hacks were rampant. I used to love map hacking in silly maps like LoaP or when I was moderating maps for the map section to see how they were made. Fun times before WC3 became a desolate wasteland of bots no one plays, sigh.

Anyway someone raised the point about host robots. What you can do is red in the IP from the host robot (need a custom build of the robot designed to communicate this) and then do stuff in game based on this. Heck why even go in game as you could make the host robot apply the IP check in the lobby before the game starts. I actually thought you hosted the map normally, if you are using a host robot you can just make that do all your IP related stuff and even communicate stuff into the game. Whatever gives you the most fun even if it annoys people!

 

[normalice]

Again it is enforcing an arbitrary rule with little enforcement. In my definition of cheating it is not since the game is behaving perfectly and no third party tool are being used.

you don't seem to understand that your definition is of no relevance. None. At all.

Most people you play with you should already know as WC3 is so old only people playing it are old people who never moved on. The rest are so uncommon they might play a session or two and ruin your fun but then they will give up.

again, please read the post where I explained that this is exactly what isn't happening.

So is a rule that I always win.

no it isn't.

I am the host after all and am spending extra bandwidth so you can play so why should I not always be a winner?

anyway, the host is a bot, and the guy who owns the bot does not play.

Well if you are in North Korea I am amazed you can even write these messages without being oppressed. Clearly it must be more free than you let on.

i'm not in north korea...

North Korea is the example of what is going on. "Everyone must have my hair cut!" is just like "Everyone must not speak as it ruins my fun!". You see it is pretty arbitrary and can be made up by an individual even if a group approves.

no, north korea is more like "everyone must go to this barber and get this haircut, or die!" Having rules on a bot is more like "everyone must go to this barber and get this haircut, or go to a different barber, or get no haircut, or cut it themselves..." Surely you see the distinction? Yes? Yes? No?

That is surprisingly hard to do when one is 12-14 since you are... um... a child.

i'm 35

Warcraft III is an old game.

what does that have to do with anything?

Yes behaving not as a child as a child would be beyond my comprehensions seeing how I was a child.

did you ever consider growing up? It's not fun, but then it's easier on your parents when you aren't living with them at the age of 40. Though, evidently, the concept of considering the convenience of people besides yourself is foreign to you.

WC3 is an old game. In fact soon WC3 will be 10 years old.

again. what does this have to do with anything?

I am guessing most people still are children playing WC3 (just a few generations younger than me).

if any logical conclusion could be reached from this, it's the opposite of the one you came to... how many teenagers these days play the original zelda with an xbox 360 sitting around the house? The only people with any reason to really play it are those who grew up with it.

Fun times before WC3 became a desolate wasteland of bots no one plays, sigh.

i would recommend playffa.net to you, as it is pretty active and worldwide, but judging by the way this has gone you probably aren't even reading this, and even if you were you would continue to not believe it is pretty active even after seeing it for yourself.

Anyway someone raised the point about host robots. What you can do is red in the IP from the host robot (need a custom build of the robot designed to communicate this) and then do stuff in game based on this. Heck why even go in game as you could make the host robot apply the IP check in the lobby before the game starts.

Obviously, they already tried that. the host bot is third party (lunaghost) and they will not cooperate.

I actually thought you hosted the map normally,

why would you think that when i've told you otherwise like five times now?

Whatever gives you the most fun even if it annoys people!

and that's what childish is

 

[Ceday]

Obviously, they already tried that. the host bot is third party (lunaghost) and they will not cooperate.

Can you explain a bit more what you mean here?(and who is second "they" =))

 

[normalice]

Can you explain a bit more what you mean here?(and who is second "they" =))

http://playffa.net/site/

ugrilainen is the owner, if you wish to contact him... I've offered only what I know, and only as far as I understand it - which I admit isn't much. You would have to ask him for more details.

the second "they" is lunaghost.

 

[Xonok]

I'll try to bring out one of the main points that DSG made, the way I understand it.

Again it is enforcing an arbitrary rule with little enforcement. In my definition of cheating it is not since the game is behaving perfectly and no third party tool are being used.

Basically, that since there is absolutely nothing preventing this, then you can't really make a rule against it. (How do you prevent people in the same room from cooperating?)
Yes, you can ban IPs or names, but in most of the annoying cases they are both replaceable.
Thus, the point is that you shouldn't make rules that you can't enforce.

Pre-teaming is not allowed.

This is a hard, if not impossible rule to enforce.

If a person is already breaking the rules of your game (arbitrary or not), then it gives them an advantage. Thus, some extra rules (no talking) can actually make it worse, since it takes away a way for players to counteract to such things.

One way to counteract this would be to build the map in a way that strongly discourages pre-teaming, but not teaming.
For instance, you could make it so that the ability of 2 players to cooperate is highly limited unless their positioning helps. That way pre-teamers would actually be penalized in most cases for trying to hold up some kind of loyalty to eachother's goals.

tl;dr - Don't make a FFA map if you want everyone to fight 1v1 all the time. Teams will be made and it should be encouraged by gameplay mechanics to happen spontaneously. By making rules against something you can't prevent you simply give the advantage to those who break them.

 

[Ceday]

I am getting stupid or there are more than 1 types of bot around W3.

Isn't there a host bot called ghost++ which is used for hosting w3 games, it can give hosted games extra network control, saving/storing game data etc.And it is a open source C program which can be modified by anyone.

If I am correct about these, why they rejected to use?

 

[normalice]

Basically, that since there is absolutely nothing preventing this, then you can't really make a rule against it.

well, that's why i'm here. to find a way to prevent it... was that not made clear in the original posts? Or is it routine for you to answer "how do I do this?" with "you aren't doing it doesn't make sense to do anything based on it."

Thus, the point is that you shouldn't make rules that you can't enforce.

it's enforced easily for most pre-teamers, but one group of three people seemed to find some program that circumvents playffa.net's blacklist criteria.

 

[normalice]

I am getting stupid or there are more than 1 types of bot around W3.

Isn't there a host bot called ghost++ which is used for hosting w3 games, it can give hosted games extra network control, saving/storing game data etc.And it is a open source C program which can be modified by anyone.

If I am correct about these, why they rejected to use?

again, I don't know the specifics. All I know is that this is handled on some level fairly well, but one group of guys from Norway managed to figure out a way around it. There is apparently a way to block them anyway, but lunaghost will not implement it.

this is only as I understand it. You would have to talk to ugrilainen for more accurate information.

 

[Dr Super Good]

i'm 35

I was trying to bring across how old WC3 was that I was 12 when I was playing it the most. That is nearly 10 years ago as I am now 21. People act like kids in games usually because they are?

what does that have to do with anything?

There are few people left playing WC3. When I used to log in to battlenet there were over 100,000 people playing WC3 alone, and 300,000 on battlenet and that was per region (both EU and US had similar numbers). Now it is struggling to reach anywhere near that number.

The players who are left are die hard fans of the game. There are very few players who are new. As such the people playing your map regularly will mostly be familiar and you can know to either trust or not trust them. The odd new person who tries out the map will either slowly integrate into your community or will stop playing after the first few sessions.

People in the community will mostly do things like silly rules and stick to them. Only people outside the community or who have turned "bad eggs" will not follow them. At high end play it should only be with people from the community so you should have no problem with teaming and other things you are making out as a problem.

did you ever consider growing up? It's not fun, but then it's easier on your parents when you aren't living with them at the age of 40. Though, evidently, the concept of considering the convenience of people besides yourself is foreign to you.

What does that have to do with anything? My age is on my profile and I am classed as an adult, have a University degree and am even employed. Amazing how much happens in 10 years.

no, north korea is more like "everyone must go to this barber and get this haircut, or die!" Having rules on a bot is more like "everyone must go to this barber and get this haircut, or go to a different barber, or get no haircut, or cut it themselves..." Surely you see the distinction? Yes? Yes? No?

Seeing how few players there are left playing WC3 it is near impossible to even fill up 1 bot let alone two. In fact I have not managed to find a full house game that remained full house for more than 5 minutes from start in the last year.

The only people with any reason to really play it are those who grew up with it.

Actually most of the players never even brought the game. Entire third party VPN networks are flooded by players from developing countries which are playing pirated versions of the game. They are playing it rather than SC2 (which has free arcade now) because they lack computers able to run anything more modern. Again WC3 is nearly 10 years old and even a phone could run it if not for the architectural differences.

i would recommend playffa.net to you, as it is pretty active and worldwide, but judging by the way this has gone you probably aren't even reading this, and even if you were you would continue to not believe it is pretty active even after seeing it for yourself.

I seldom play WC3 and most certainly not PvP maps. After modding the game as long as I have any fun is ripped out of PvP as one sees flawed and unbalanced mechanics so easily, even if they are completely invisible to most.

Obviously, they already tried that. the host bot is third party (lunaghost) and they will not cooperate.

Well do as you said and host your own host bot. Do not like the service from one robot, setup your own!

well, that's why i'm here. to find a way to prevent it... was that not made clear in the original posts? Or is it routine for you to answer "how do I do this?" with "you aren't doing it doesn't make sense to do anything based on it."

If you can solve such a problem as this then it will be only a mater of time before the government starts taxing you for breathing air. Same kind of problem.

it's enforced easily for most pre-teamers, but one group of three people seemed to find some program that circumvents playffa.net's blacklist criteria.

Well depending what it is, it could be anything from joining the game with a fake name to using proxies if they are past a NAT to connecting to the bot via different VPNs which give them unrelated IP addresses. Which unfortunately again shows that IP sharing detection is not really a viable way of solving your problem. The risk of false positive and the ease of generating a false negative shows that such a system cannot be used to determine if two people will work together. It may be statistically related to it but the factor alone is too indeterminate.

 

[normalice]

The players who are left are die hard fans of the game. There are very few players who are new. As such the people playing your map regularly will mostly be familiar and you can know to either trust or not trust them. The odd new person who tries out the map will either slowly integrate into your community or will stop playing after the first few sessions.

as mentioned ad nauseam, this is completely false... There are thousands of users, many of which are unknown to the guy who maintains the bot.

At high end play it should only be with people from the community so you should have no problem with teaming and other things you are making out as a problem.

the games matches are organized by 'first to join.' playffa.net does not control or attempt to control who can join which games, except of course the blacklisted players.

Seeing how few players there are left playing WC3 it is near impossible to even fill up 1 bot let alone two. In fact I have not managed to find a full house game that remained full house for more than 5 minutes from start in the last year.

that might have less to do with world-wide WC3 activity and more to do with this:

I seldom play WC3 and most certainly not PvP maps.

Actually most of the players never even brought the game. Entire third party VPN networks are flooded by players from developing countries which are playing pirated versions of the game. They are playing it rather than SC2 (which has free arcade now) because they lack computers able to run anything more modern. Again WC3 is nearly 10 years old and even a phone could run it if not for the architectural differences.

there does seem to be some truth to that. Maybe 15% or so of the players I play with seem to be from the U.S.

Well do as you said and host your own host bot. Do not like the service from one robot, setup your own!

um, i don't run the bot. I'm merely trying to attempt a relatively off-handed script to help improve it.

If you can solve such a problem as this then it will be only a mater of time before the government starts taxing you for breathing air. Same kind of problem.

it really isn't. this more of your "rules i don't like = dictatorship" logic?

Well depending what it is, it could be anything from joining the game with a fake name to using proxies if they are past a NAT to connecting to the bot via different VPNs which give them unrelated IP addresses. Which unfortunately again shows that IP sharing detection is not really a viable way of solving your problem. The risk of false positive and the ease of generating a false negative shows that such a system cannot be used to determine if two people will work together. It may be statistically related to it but the factor alone is too indeterminate.

it isn't an urgent matter. I'd be quite happy to run some scripts that collect information for review to see if there is anything I can use to better refine who is labeled a pre-teamer. However, at the moment I am strictly limited to data derived from game activity. If there was some way to extract player orgin or something, that might be useful..

 

[Dr Super Good]

the games matches are organized by 'first to join.' playffa.net does not control or attempt to control who can join which games, except of course the blacklisted players.

This is the main problem. You are playing with unknown players. Some of them may not even be native English speakers and not understand any of the rules you try and communicate to them. Others will be trolls who are there to ruin everyone's fun. Some people may have decided to join as a team unaware what the map even is or any of the rules you are trying to enforce. This brings back to the enforcement viability of those rules and the fact that they just cannot be enforced in public games.

Use the host robot to open up a private game (yes host robots can have many games running at once as long as only the number of keys attached to it are in the lobby at any given time) and invite only trusted people to it who you know follow your rules.

This is how SWAT Aftermath used to be played with public games being filled with low skill, new or uninterested players that could not manage more than Hard/Insane. For Nightmare games private games were hosted where players where let in by invite only from the map based chatroom. Only people who were known to have the necessary skills or were close to having them were invited. This prevented people from ruining the sessions as the game required a lot of skill and knowledge to play well and even 1 person failing could cost the game at high difficulties.

that might have less to do with world-wide WC3 activity and more to do with this:

No I was just going on the lobby stats that you are shown when you enter BattleNet chat mentioning how many people are on the service and how many are playing WC3.

it isn't an urgent matter. I'd be quite happy to run some scripts that collect information for review to see if there is anything I can use to better refine who is labeled a pre-teamer. However, at the moment I am strictly limited to data derived from game activity. If there was some way to extract player orgin or something, that might be useful..

You are overcomplicating things. Instead of getting a PhD from analysing game behaviour to detect possible violations of some arbitrary rule by writing an intelligent agent capable of detecting reliably detecting a violation from various game percepts you could just not play with them or even resort to manual moderator. Anyone can be made a moderator of a WC3 host robot and any moderator can be allowed to ban whoever he wants (except if they to are moderators or the administrator).

Do not like what someone is doing? Bam he is gone from the game in no time. Even works during gameplay (it will disconnect kicked/banned player). Want to justify your bans? Even though you do not have to you can add triggers to the map to non-bias monitor player activity, such as who is attacking who. Suspect someone is violating your rule? Just look at the data and decide for yourself if he gets the banhammer. Do not want to break your own rules using the data? Simply obfuscated it and only query specifics if you notice a problem in the obfuscated data.

The main metric to measure would be large fights. These are not a few scouts or odd lost unit entering combat but large battles with considerable number of units (or if it is a hero based game, involving hero on hero for considerable quantity of damage). By analysing when, where and with who these large fights occur should allow you to pickup cheating. Specifically you want to look for strong correlations in targets or sequences.

If two players engage in large fights against the same target near the same time there is some chance they are working together. If this pattern is persistent it is more likely.

If two players share common victims of large fights there is a some chance they are working together.

If two players are noticeably avoiding each other (few or no large fights against each other) then there is a chance they are working together.

If a player involved in a large fight has their assaulter engaged in a large fight by another player (either to rescue them or as a Carthage attack if the game has bases) then there is a chance they are working together.

If player appears to be engaging in large battles with the victim of a just ended large battle then there is a chance they are map hacking or sharing intel. The more cases of this the more likely.

Put metrics like this together and you can start to build up a case against a player for violating your rule. It will never be 100% accurate but it should easily pick up on obvious ones which would score usually high in some of the metrics.

 

[normalice]

This is the main problem. You are playing with unknown players. Some of them may not even be native English speakers and not understand any of the rules you try and communicate to them. Others will be trolls who are there to ruin everyone's fun. Some people may have decided to join as a team unaware what the map even is or any of the rules you are trying to enforce. This brings back to the enforcement viability of those rules and the fact that they just cannot be enforced in public games.

i'm getting a little tired of repeating myself. they are enforced. In all cases but one, the pre-teamers and map-hackers are caught and subsequently banned shortly after the game is over. As such, pre-teaming and maphacking is minimal on this bot. However, one small group of players has found a way to circumvent the banning process.

Use the host robot to open up a private game (yes host robots can have many games running at once as long as only the number of keys attached to it are in the lobby at any given time) and invite only trusted people to it who you know follow your rules.

Read.

No I was just going on the lobby stats that you are shown when you enter BattleNet chat mentioning how many people are on the service and how many are playing WC3.

the bot is not hosted by battle.net. most bots are not. indeed, many of those who join do so via w3arena and such.

You are overcomplicating things. Instead of getting a PhD from analysing game behaviour to detect possible violations of some arbitrary rule by writing an intelligent agent capable of detecting reliably detecting a violation from various game percepts

coincidentally, that's exactly the sort of thing I'm getting a PhD in.

you could just not play with them or even resort to manual moderator.

nope. i know that sounds feasible to someone who hasn't read anything I've written about activity, but try to understand: there are hundreds of games per day, many of them running in parallel, and only three moderators who do not play very often, if at all.

 

[Dr Super Good]

the bot is not hosted by battle.net. most bots are not. indeed, many of those who join do so via w3arena and such.

I never said it was. I said that on BattleNet you can enter the chat and it will print you statistics of how many people are currently on line. The numbers reported are under 1/50 of the size they were 7 years ago.

Most robots are listed on BattleNet as all you need is 1 CDKey per listing and it works in all regions. This is how you can be in a game with people from the EU, NAE and NAW all at the same time. It can even list it on VPNs at the same time and everyone can play together as BattleNet only acts as a coordination server and actual hosting is all peer to peer.

However, one small group of players has found a way to circumvent the banning process.

I would guess they changed their names and IP addresses. Both of which are trivial to do.

Read.

You are not reading what I said. You make a chat room and only invite people you trust to the game. You know a private game?! That way you can be 100% sure no one will even try and cheat. That is how every serious community in both WC3 and SC2 does it when they do not want public "trash" ruining a session.

coincidentally, that's exactly the sort of thing I'm getting a PhD in.

Then why you asking here? If you are already a post graduate computer scientist you would know everything we have said already and would know that any sort of IP detection is totally useless alone.

there are hundreds of games per day, many of them running in parallel, and only three moderators who do not play very often, if at all.

Hundreds of games per day sounds like a fantasy to me... Anyway so what if some people "cheat" playing a public? No public custom map that needs axillary "word of mouth" rules can have any sort of ranking. Only rank private and trusted games. Heck why rank anyway as I could swear people played FFA to have fun and not because they wanted to gamble wins and losses.

 

[normalice]

I never said it was. I said that on BattleNet you can enter the chat and it will print you statistics of how many people are currently on line. The numbers reported are under 1/50 of the size they were 7 years ago.

i'm pretty sure I wasn't disputing that.

You are not reading what I said. You make a chat room and only invite people you trust to the game. You know a private game?! That way you can be 100% sure no one will even try and cheat. That is how every serious community in both WC3 and SC2 does it when they do not want public "trash" ruining a session.

... you speak as if you completely failed to comprehend that this is not my bot. Also, you failed to comprehend the fact that the whole point of playffa.net is the hosting of ffa games to a broad audience. Private games would undermine that.

Then why you asking here? If you are already a post graduate computer scientist you would know everything we have said already and would know that any sort of IP detection is totally useless alone.

IP detection can help. Do you understand that? Any good classifier is going to use multiple features. Gameplay data is easy. I can do that. I can't do IP/NAT/origin detection, which is why I came here. A simple "no" would have sufficed, by the way. A long discussion on the philosophy of the rules and motives of someone else's bot isn't what I had in mind. It was entertaining for a while, but I'm getting a little tired of repeating myself only to have the repetitions continue to not be read...

Hundreds of games per day sounds like a fantasy to me... Anyway so what if some people "cheat" playing a public? No public custom map that needs axillary "word of mouth" rules can have any sort of ranking. Only rank private and trusted games. Heck why rank anyway as I could swear people played FFA to have fun and not because they wanted to gamble wins and losses.

fine. you can consider it to be fantasy. But you're welcome to check the link I've posted several times now to prove your theory wrong. Or just type playffa.net into your address bar and push 'enter'.

 

[Tal'Rarity]

There is a possibility with preload hack and I/O now that I think about it however I seriously doubt it. Never know though.

Sorry I can only suggest this, I can't provide much insight to this matter.