Reyhan

I got lucky with trinitrophenol. I didn't pick easy wins; most of the 8 games I played were against 1400-1700 players. I thought they were fair matches but many of them resigned at the first sight of my army...

Precisely Many players above a certain rating are bored or they fear imbalanced games.

... hmmm My smurfing started in July when I tried to test rated games in A25 svn. I made accounts and played rated games between each other to see if the rating bot still works. Then I logged into the real game (A24) with those accounts and people's reactions were funny, so I went along with it and uncovered many other smurfs. Then it became a funny thing to do in late A24 because everyone was bored by the OP archers and turtling meta, smurfing became a source of entertainment, and many people found it funny. This is when the family smurfs appeared In A25 as cavalry rush became OP, slightly higher rated players became the target. This is when smurf accounts became necessary. Another thing is, I still find it entertaining to create new accounts and see what people have to say.

He has been unmuted. Problem solved.

Why was he muted? Unmute him please, he said nothing wrong. people get muted or banned randomly.

Smurfing doesn't always work. Sometimes I get spotted just from my playstyle: unusually fast boom, no rush, mostly women in early game. Sometimes, people place their suspicion on FeldFeld or Metafondations before they think about me. So to smurf better, I purposely slow down my boom and make soldiers or cav in early stages, then go to 200 pop before clicking P3 (to imposter Rauls and Hamdich). I used to compomise techs to maximise unit production but now I research techs just too make myself less obvious. Impostering other players often work as well. I created this fake identity Reyhan to protect my friend from abuse in the lobby. I just claim that 'Reyhanna is my sister, don't touch her or you are dead' everywhere so that she(he) can play in peace without being targeted as a threat or being trolled. When I create a smurf account, I have learnt to be flexible and random to confuse everyone. The only bottom line is not abusing unsuspecting cosmic noobs.

I know you are a very nice person, Stan, but this is an extremely controversial topic and involves gross violation of multiple rules. Of course. We can just add more features and fix the any bad game mechanics.

It is almost completely useless right now. Too many people are unrated, too many are underrated, too many are overrated. This promotes kicking new players, which is the problem. But @Dakara has made some good points here, and has given me the following inspirations: There is already a player profile lookup function, and that allows you to document the number of games played. Now, we can add in more information, including: Date of account creation Total number of games played, including unrated and TG, and their outcomes Total time spent playing. This allows you to tell smurfs apart from genuine new players: smurfs are likely to have 100% win rate whereas a genuine new player would have a much lower win rate and more number of games before they reach the same rating. Here are examples of a real account and a smurf account: Real account: This player clearly has reached 1429 after accumulation of experience over many games, some were won some were lost. This looks like the natural progress of someone from nub to OP. Smurf account: This player has only played 8 games and won all of them. On average they won 25 points per game, which implies they have to beat very strong players (1400-1600) players every time to progress this quickly. No new player can reach 1501 without a single defeat, therefore this must be a pro player's second account, who wipes out any unsuspecting challengers in the lobby. Hints such as this can tell you who is smurf and who is genuine. If we include the team games as well then we would get an even better picture.

They did help. But enemy cav come and go faster. It is economcally unwise to keep 10 cavalry idle in my base just to protect me. If you rush the enemy, then I am totally dead. In fact, one of them did rush but a single team of 10 cav can never beat 30 cav on the move. Furthermore, they must prevent being outboomed by the boomers on the opposite team. You have a point. However, you must also remember that a TG is not a 1v1: if you rush, as long as you didn't kill everyone on the enemy team, one of them will boom up and ram your CC down at minute 14 with a full army. If all 4 of you rush, if just one enemy is ahead of you then your team is dead. It is a very risky strategy to go all rushing. The most optimal situation is 2 rushers and 2 boomers, or 1 rusher 3 boomers. My teammates were no noobs; all of them were 1400+, and they did help me by diverting some enemy cavalry away. This is not plausible if you are 1600+. Just how many 1600+ players are there, and how many of them can be online at the same time? Now, even a 2000 player can make a mistake, so there is no guarantee that having all 1600+ players in a TG will prevent you from being placed in a 3v1 if one of them tells everyone that you are an easy target. Yes, bullying is encouraged by the meta and it is very effective at winning games where a team has 1 pro and 3 weaker players. If the pro is dead then the other 3 lose morale and will be killed easily. However, from the perspective of the losing team, they had no fun because it's gg 6 minutes in, and there is nothing you can do to stop it. This is the root of most problems. And the reason for small playerbase is because of unwelcoming hosts, then it goes back to the smurfing and imbalance problems. So a viscious circle is formed. The only way out of it is to make 3v1 less viable or less effective so that there is less motivation for smurfing. This requires some changes in the game mechanic. Good suggestion.

We should start a new thread about improved rating system. However, an idea is that rating gain could be a function of the player's score. For example, rating gain = A * ln(B * score + 1) Where A and B are some constants to be determined as appropriate. This model has the advantage of no ridiculous rating gains from just one game, due to the property of logarithm (f''(x)<0). The +1 is necessary so that a player with 0 score does not get a negative infinity rating. Using this model, taking A=1 and B=1, a 9000 score player would gain 9.1 points, a 30000 score player would gain 10.3 points; a 800 score player would gain 6.9 points. You may argue that this is unfair as the 30000 score player did much better than the 800 score player. So let's consider some function that takes account of the player's individual performance and the overall performance of their team: f(player score, team average) We can deduct points if a player plays too badly with respect to their team, and reward the overperforming players. So we can have a comparing term player score / team average or, player score - team average And then we operate on the result of this comparison to give a rating change. One such model is rating gain = A * exp(B * (player score / average score of their team)) - C C is a threshold constant, which determines whether someone gains or loses points after comparing with their teammates. The advantage of this model is that a player in the losing team who played very well can still gain a lot of rating in spite of their noob teammates. Suppose a 1800 player is paired up with 3 cosmic noobs, the 1800 player can still gain quite a lot of points if he scores 20000 and the noobs score 5000 each. On the other hand, a noob player in the winning team may still lose points.

Players only gain rating if they win a stated 1v1. This is not good enough. There are extremely skilled team game players with very low ratings and 1v1 players with ridiculously high ratings that don't suit them. So rating has 0 meaning now; many 1300 players can beat 1700 players. Some frequent smurfs never had a rating above 1500 although they themselves can beat any 1900. Only rated 1v1s are counted in a player's profile, which is ridiculous. Proposed changes: Any game a player plays is counted. So when someone clicks on the player profile they can see exactly how experienced someone is. Ratings change even for team games. Remove the 'exit' button. Only leave the resign button in the interface. If the host terminates the game, then the player with the highest score gain rating.

I was Gauls... I made 7 cavalry and had infantry everywhere in my base... But 30 cavalry is too much at minute 6. They just go straight in for the farms then woodline. Yes, if I make cav from the very beginning, then I could have defended it. But, any fight with 30 enemy cav will result in huge losses, and considering that I am pocket to a weaker player threatened by someone with 200 ratings more than them, I can't afford to go full defensive on my own. I expected to be rushed since the very beginning because I know 1 of their player is a rusher, but the other one in his team just convinced them to coordinate a rush on me because he thinks I am an easy target. My teammates did help me to defend by send a few of his Cavs, and I thank them for it. But the result was he reached P3 at minute 17, at that time there were 2 rams at his cc. The ge was poorly balanced because certain players are underestimated. This rings a bell for another issue: ratings.

@Stan` I would really like you and any other moderators to be aware of these problems. I can't offer a solution for every one of them, but some can be solved by better code.

Now onto malpractice in the lobby 1. Spamming. This can vary from spamming any nonsense to profanity. Some hackers are proud of their spamming skills, and they creat sophisticated spams that can temporarily block out the lobby chat or game chat. The most notable examples are 'snow people mountain people' and 'follow the white rabbit'. The spammer may join any game and start spamming so that the player's screen is full of white patterns and they can't see the actual battle anymore. Fortunately this hacker has been inactive for a while. This made many hosts ban specs and makes it less friendly for new players who want to learn from pros. 2. Trolling the bots. Some people repeated ping the ratings bot or wfg bot. 3. Cheating. This can be done in many ways but generally it either involves quickly changing the game from rated to unrated or vice versa before starting or trying to gather enemy information not through scouting but an observer PC . Cheat codes need to be banned.

Now let's talk about DDOS. I myself don't know much about the technical side of this attack, but it seems to me that whenever more than 2 players are losing connection or the host crashes, people whine about DDOS. Then they start to point fingers at each other and accuse each other of ddos even though no such attack has ever happened; host crashes may be due to bad laptop or bad broadband, sometimes even 0AD bugs. However, the quest to pull out the ddoser has never stopped; the 'who is ddosing' argument has caused many feuds in the lobby, for absolutely nothing.