You know TG Elo was broken for years due to a bug in the ratings calculation, right? As a result, many people have Elo way above their actual skill. I have seen people with 2500 TG Elo yet with 40% winrate.
If you get someone with <45% in your team there’s high chance he’s one of those and is on a losing streak until the system manages to put him where he belongs. I would not say one should leave based on his teammates winrates’ but it’s true that winrate has become a very solid indicator of the outcome of a match and I for one would not like to play a match with 3 teammates at 45% winrate (good thing I play only with premades now). Once again, stop calling people ‘toxic’ when all they do is exploit a system that just asks to be exploited.
We should usually have equal Elo on both side. There is something wrong if somebody often get matched against enemies with higher elo. You must know that he also have chance to pair with low elo players.
We can never know if he already got his correct elo or not.
It was all 3 players with lower than 45%. 1 player with that win rate isn’t a problem if he is going to try to win the game. In my experience, there is only a small chance that all of them can carry their weight, and it is a guaranteed loss if there is 1 or more smurfs on the other side. There is a great chance that one or two of them will give up early. There are really a bunch of players who rage quit and 80% of them with lower than 50% win rate according to my experience.
I can try to help you to understand this kind of behavior. One guy gave me a low elo account with 30% win rate a year ago. I solo pushed all the way to 2500 TG elo but the win rate only reached 38%. People dodged 4 of 5 games and they said how can you get such low win rate, you must be a troll blablabla. I ended up with playing with the guys who don’t care about win rate and I lost many games in a row. It shows that many people will dodge me if I don’t care about my win rate.
I won’t play the game when I know I am going to blame my teammates. I don’t want to blame them so i choose not to play with them. I guess nobody want to lose many times a row. Also, it is obviously people don’t like some types of losing such as 1 guy resign early then team resign early, got beaten by smurfs, etc.
Only solo players will get completely random matches. Premade team got huge advantage over non-premade.
Nothing will absolutely show the skill of a player, but it partially show the gaming attitude of the player. It is my choice to take the risk or not.
It can be called toxic whatever you want, but it is necessary to me. Very few high win rate player will rage quit. Although some of them talk trash which is acceptable to me but most of them will keep playing and being helpful. As I mentioned before, most of the early resign and rage quit happened to low win rate players.
Since when the word selffish can be used here? Don’t you know that you are toxic and selffish as well when you are calling others selffish ? You didn’t understand or tried to understand why people are dodging the games like this. I would say the people who don’t want to try to win TG games shouldn’t play TG. This is all my problems with TG.
Sorry but your dodge makes absolutely no sense. We have elo rating for a reason, because winning percentages just tell nothing about the skill of a player. A player can get 80 % winning percentage if he dodges all people who have higher elo than himself. But this doesn’t makes him a good/better player. It’s actually the opposite. Because he avoids challenge he will stuck at the level he is in the long run. So a negative/lower winning record actually indicates that a player may have faced a lot of stronger players and therefore has more experience to contribute to a team game.
And this works also in the different direction. If a player is more often paired with players that have higher elo than him he will have a negative winning record. But still his elo will be correct as a lot of his wins are wins against players that are actually rated higher than him.
That’s why winning percentage is absolutely no indicator of the winning chance of a player, it only shows if the player was more often paired with players that are better than him or more often paired with those who are worse than him. It’s just randomness caused by the matchmaking.
Just an example from the current rankings:
Tatoh has more than 13 % higher winrate! But JorDan just has played more often against players that are better or at the same level as him, that’s why he has the lower winning percentage, cause if he wins against better players he gets more elo than tatoh if tatoh wins against weaker players and vice versa.
The elo correctly shows, that they are currently about at the same level, but according to you, the lower winning record of jordan would indicate he would be the worse player.
Nothing is absolutely right or wrong here. You might wantna reconsider your words. We all know that a player with 100% win rate have a great chance to carry this game and a player with 30% win rate will probably be a troll. The player with 80% win rate in your example could have dodged many games before, but most likely he will win this game when he decided to play this game.
I could partially agree with your opinion, but does not change the fact that nearly all rage quiter and early resign players got low win rate.
Disagree, if they quit a game their elo goes down and they get easier opponents. Then they have higher chances of winning in the following games. Yes their winrate would be marginally lower than it would be if they wouldn’t do this. But the randomness of matchmaking has much higher influence there.
I don’t know where you got your “fact” from. Have you made a statistical test with confidence intervalls? For me this is just a hypothetical construct, that would need to be proven.
What it indeed does is lowering the elo. Rage quitters will get a lower elo than they would usually have. So actually a better indicator for a rage quitter would be someone who has higher winrate than other people in his elo range, not the other way around. (Cause rage quitting will lower your chances of winning against equal or better opponents but you will still have high win chances against weaker opponents => as you easily give up chances to get big elo points from people ranked higher than you you will overall lose elo).
So actually, a comparable low winning record for the elo would be a sign for people that don’t ragequit.
But i think this effect is largely overridden by the matchmaking randomness aswell.
Hmm. Interesting theory. These are all worst case scenario and assume random preference among maps, which are obviously flawed assumptions. (A 4 person Premade team would need to be calculated out as a single person, because they would all in theory ban the same map) (if 90% of the playerbase bans islands, then it lowers the effective percent of population baning in your calculation)
However, it certainly is directionally correct, and as far as a thought experiment really does demonstrate the enormous magnitude of the “small” change people are asking for. Very interesting indeed.
The lobby system needs to have the lobby ranked system visible similar to how the previous versions do - I know it exists, just needs to be visible to improve the manual lobby system - it would help to reduce smurfing by people with no rating.
But since you started it, go ahead and compare the definitions between “hypothetical” and “prediction” in the dictionary. They are two different things. That spreadsheet was a mathematical representation of “hypothetically”.
if we gave them the power to ban infinite maps, and 100% of the player base used that power to ban 90% of the maps, we would see… " Now that we’ve done the math, we consider whether that is actually realistic at all. We look for assumptions that are hidden in the math and whether we agree with those assumptions. We critically think about whether that means the answer will be higher or lower. We might even use some of the other math to estimate by what order of magnitude it might be off by.
If you re-read the response, I pointed out that it was unrealistic in magnitude, and I pointed out a couple of flaws in the logic as to WHY it reached unrealistic results as far as specific numbers.
Soo… while you are correct about the fact that it is EXTREMELY unlikely to ever see a 30 million minute queue, you were very Incorrect in your analysis of what I was saying. You were also used incorrect logic to say the math was wrong because you didn’t like the outcome. (The math very well may be correct, but the assumptions for that specific section of the hypothetical matrix don’t align very well with what we can expect the community to choose based on several factors.)
First, you concluded your post with “Very interesting indeed.”, so no, you did not “point out something that was unrealistic in magnitude”, you pointed out something that was completely incorrect, without doing basic checks, and implied it should be trusted, because you liked the outcome.
Second, I did the math in another thread, someone who bans all but Arabia should get around 1 minute queue time, as expected from the intuition, very far, very far from the 30 min queue supposedly in the spreadsheet. So no, I did not say the math was wrong because I didn’t like it, I said the math was wrong, because it is wrong, as one could guess from the nonsensical 30 million minutes predicted, and even OP admitted it was wrong.
Third, I have no idea who you are, and your history shows I replied to you only once before, so your repeated “Leave me alone” makes even less sense than the rest of your post, unless of course, you are a smurf account.
Haha alrighty then. Looks like you understand everything perfectly. you still missed a step in the logic and therefore misinterpreted both the intent and content of the message. But that happens. I wouldn’t sweat it too much if i were you.
In an ideal system if they could filter and match people by map first and then by elo you could reach times around a minute. Great. But that solution is not “giving infinite bans” . It is also not what his comment is about if you read it and the context he linked it to.
We are discussing just “giving infinite bans” and the catastrophic failure it would be if just implemented as so many people think is good. Like go into the code and change the variable “max_map_bans” to 9 instead of actually fixing the fundamental flaw.
If this company wants to know if age of empires community likes or dislikes this update with the elo system , maybe they should make a pole easily accessible to everyone and see where the new patch stands,
my vote = this patch is the worst patch to date, 2nd worst is removal of DM whats next?
It’s not a software programming forum… When people say “give unlimited bans”, they mean, “rework the system to allow infinite bans”, not “set the variable map_bans to 9”. Devs will know how to do it, no need to debate the implementation for them… The important point is that, there IS a technical solution. Whether it’s presented in the form of Arabia/Arena/Random queue or unlimited bans is just a UI thing.
In fact, unlimited bans is already implemented in quick play (you can check, it’s presented as unlimited bans, not as multiple queues for each map, and we don’t wait 30 million minutes), so even your assumption that it’s not just one variable to change.may be incorrect. It might actually be just that simple.
Hmm. First off I checked two days ago while I was looking into this, and you do NOT have unlimited bans in quickplay. it is the same rules as the standard, just presented differently. I highly doubt they changed that within the past two days. But feel free to screenshot a picture of the quickplay queue with all maps banned except for one. I’ll gladly admit I’m wrong if you can present the evidence it was changed within the past two days.
As far as your comment on “when people say” … you and I have had different experiences on that obviously. (admittedly I have been active on steam forums rather than this particular age forum, but in the steam forums at least that conversation has happened countless times).
In my experience the conversation normally goes something like this:
“we want unlimited bans! TODAY!!! it should take like 5 seconds to implement!”
“it will be a relatively large change that will take a lot of structure changes because they’ll have to have it match by map first and then elo instead. And there are always chances for bugs with a change that big.”
“no, they just need to change it from 3 bans in 1v1 to 6. and 4 bans in teams if I solo queue. Or at least just add one to each.”
“you can’t do that because then it is possible that every map will be banned.”
“well, if it matches you and all of them are banned, then it should just drop that matchup and rejoin the queue for each of you until one of the matchups works out.”
Then someone tries to explain using words that it would multiply queue times, especially if a large percentage of the population bans all maps than one.
Then the first guy says something like “nah. it wouldn’t affect anything”
That is where Floydroid’s chart would be useful. it would show the mathematical probability of a given percentage chance with a few key population percentages.
That is the type of scenario that Floydroid’s math is set to estimate and demonstrate why it is a terrible idea. I’m sure you can see the flaws with that particular solution, thus why you assume they would never do it in the way that is mathematically represented here (at least hypothetically other than the specific adjustments that would need to be made for incorrect logical assumptions). (Of course they would follow a solution that makes it match by map first rather than just requeue… right? RIGHT??)
Part of me agrees with you that we shouldn’t be the ones coming up with the real details on the solution, but the other part of me realizes: you’ve see in the past where they just take what the forums say and run with it. (i.e. just timeout) without really thinking of all of the angles and possible outcomes and pre-emptively solving those issues. Soo I feel we do have at least some responsibility to be vocal about particularly disastrous outcomes from particular solutions. I think we’re both on the same side here in general. We both think they need to match map first and THEN elo/teams. We just have differing opinions about our role in highlighting potential issues, and there has been a bit of misinterpretation in the thread. That’s all.