The reason I dislike PvP is because we have a Matchmaker at all.

[DanAlcedo.3281]

Let me explain. 

First off. We need to assume for a moment that we have a healthy pvp population. I know it's hard but please do. 

The Goal of the Matchmaker is to create a balanced match up. A match that both parties have a equal chance of winning. 

Aka. A 50% win chance for both teams. 

The Matchmaker does this regardless of rank. 

Some of you think that this exactly what you want but for me, it a bad experience. 

Here is why. 

All matches are simply a coin toss. In the way that you can predict the outcome of a match before you even hit the Q button. 

If I win a fight, statistically I'm supposed to loose the next. Of course this doesn't mean I have to loose. 

But with each win, the chance of me winning the next match gets massively smaller and smaller. 

What's worse is that skill doesn't matter too. 

If I get better, sure I will win a bit more but when I reach my new rank that I belong in, it is back to 50% win rate. 

I personally have seen it when I "farmed" my legendary Amulet. 

I did 200 games and won 100. Exactly 50% win rate. The matchmaker works. 

But oh boy, I hated every second of it. Not the PvP in of itself, but the fact that I KNEW the outcome of a match before I queued up. 

All PvP is, is a Coin flip simulator. As literally as it can get. 

In Wvw, if im better at the game, I win more fights. 

In PvP, that's not the case. I can get better but I won't get more wins as the matchmaker enforces a 50% winrate in a best case scenario. 

 

And as long as the matchmaker exists, no amount of balance could get me into it. 

 

Feel free to disagree but that's the way I feel about it. 

Edited April 23, 2022 by DanAlcedo.3281
Fixed some words.

[Ragnar.4257]

That's not how flipping a coin works.

If I flip a coin two times, the outcome of the second flip is not determined by the outcome of the first.

Not saying the matchmaking is fine, but your reasoning is flawed.

[Multicolorhipster.9751]

In Ranked or Unranked?

Unranked could prolly do fine without, but with Ranked the only way you earn rating is based on the rating of other people in your match.

That could easily be like +3 for wins or -20 for losses without matchmaking, because sometimes its like that with matchmaking.

 👀💦

[phokus.8934]

On 4/23/2022 at 7:05 AM, Ragnar.4257 said:

That's not how flipping a coin works.

If I flip a coin two times, the outcome of the second flip is not determined by the outcome of the first.

Not saying the matchmaking is fine, but your reasoning is flawed.

A lot of people don't know how the match maker works nor simple math.  The OP is one of these people.

Edited April 24, 2022 by phokus.8934

[ArthurDent.9538]

So you are proposing completely randomized team? 5 man team q only? or something else?

[Reikou.7068]

The current rating system that gives MMR based on personal rating vs enemy team average would actually benefit from removing matchmaking completely, making it a complete random arena.

[Nymthalas.4019]

Tbh wouldn't make sense the matchmaking placing temas that are not kind of the same rate, but for me it would get fixed if there was no solo Q and you would have to search your 5 people party to play with, when you win or loose it has a meaning, you can analyze what went wrong or what went right and improve, with random people is absurd you can analyze yourself and thats it new team and nothing changed... get boring and sucks the fun out of the game mode.

[Xentera.4560]

On 4/23/2022 at 7:05 AM, Ragnar.4257 said:

That's not how flipping a coin works.

If I flip a coin two times, the outcome of the second flip is not determined by the outcome of the first.

Not saying the matchmaking is fine, but your reasoning is flawed.

 

Exactly. Which makes tossing a coin vastly superior to the current kitten matching system in Gw2. Every match should have a 50:50 outcome, and that should not be influenced by the result of your previous match. 
When your win rate is more than 50%, you know for a fact the MMR will place you with a team that is supposed to lose the next match. If this is not discouraging players from playing more ranked matches then I don’t know what is.

Then we wonder why the population is so small.

[Infinity.2876]

I don't understand why don't you play necro then you'll get a 75 percent chance instead

(After posting I reconsidered the importance of making this comment)

Edited April 25, 2022 by Infinity.2876

[Spellhunter.9675]

For all of you people who think matchmaker is pushing you to 50% winrate. Do you think the game will pull you to 50% winrate if you will just stay afk at spawn for dozens of matches? I can guarantee it will not. So stop this thrashtalk about matchmaker and learn to play.

[Ragnar.4257]

59 minutes ago, Xentera.4560 said:

 

Exactly. Which makes tossing a coin vastly superior to the current kitten matching system in Gw2. Every match should have a 50:50 outcome, and that should not be influenced by the result of your previous match. 
When your win rate is more than 50%, you know for a fact the MMR will place you with a team that is supposed to lose the next match. If this is not discouraging players from playing more ranked matches then I don’t know what is.

Then we wonder why the population is so small.

This is clearly disproven by the fact that people quite often go on long win/loss streaks.

If winning the last match meant you were guaranteed to lose the next match, win/loss streaks would be impossible.

[Xentera.4560]

35 minutes ago, Ragnar.4257 said:

This is clearly disproven by the fact that people quite often go on long win/loss streaks.

If winning the last match meant you were guaranteed to lose the next match, win/loss streaks would be impossible.

The MMR placing you on the team that is intended and expected to lose, is very different from you actually losing the match. Also very often if you solo Q, your win streaks are evened out by losing streak as well. Which at the end results in a win rate of around 50% unless you beat the system.

[Wild.1705]

On 4/23/2022 at 1:18 AM, DanAlcedo.3281 said:

I can get better but I won't get more wins as the matchmaker enforces a 50% winrate in a best case scenario. 

That is fine bye me and should be fine by most of the community. However, match quality must also be high for this to be easily accepted. Currently, match quality is so bad it almost fees like the MM is a parity of a working MM system.

Players can be way, way, way, lower or higher in ranking on a single team or opposing team, players that consistently afk (and don't get banned), bots, match manipulators etc. This is the real reason climbing sounds like a good idea is to get out of the "bad match" ranking. But unless you match manipulate, all rankings have these problems.

 

P.S. No, ArenaNet match maker, I did not want 2 deadeyes and a core ranger on my forest of Nifhel match last night.

 

Edited April 25, 2022 by Wild.1705

[JusticeRetroHunter.7684]

12 hours ago, Ragnar.4257 said:

This is clearly disproven by the fact that people quite often go on long win/loss streaks.

If winning the last match meant you were guaranteed to lose the next match, win/loss streaks would be impossible.

Which ironically also proves that rating also doesn’t place you in the appropriate rating because of said volatility. For example you can flip 10000 coins, and in one instance you get a sequence of coin flips where it’s 50 losses in a row and 50 wins in a row. During that time period say you were -15 or +15 rating for each win or loss than at that instance you were either 750 points above or below your 50% win rate skill level. 
 

These instances in volatility are not inconsequential they have an effect on the entire systems behavior and it’s a problem. Like a kind of butterfly effect…because person X was in your game with a rating of 1200 but his actual skill level is that of a plat player, he’s gonna be placed on a team in which the match maker believes he is a noob. Because you have 9 other players here their rating is thus correlated to this persons botched rating…and if all players have botched ratings like this, then the entire system is indeed a statistical soup of highly correlated crap data.

 

So in a sense the game is a kind of coin flip…but not really perhaps that’s the wrong analogy…it’s worse than that. I make the analogy often but it’s like the YouTube or Instagram Algorithm…where we are sort of bound to getting a certain kind of outcome rather than what we think the outcome should actually be…You know… being recommended pointless, low effort hot tub streamers with onlyfans that call themselves gamers rather than…well genuine, high effort and meaningful art on these platforms. The matchmaking algorithm is like this but in another sense…where the algorithm is pulling you to a certain rating like a magnet…this magnet is the overall statistical average of the entire soup of players, because all these players ratings are highly correlated rather than independent. 

Edited April 26, 2022 by JusticeRetroHunter.7684

[Spellhunter.9675]

8 hours ago, JusticeRetroHunter.7684 said:

Which ironically also proves that rating also doesn’t place you in the appropriate rating because of said volatility. For example you can flip 10000 coins, and in one instance you get a sequence of coin flips where it’s 50 losses in a row and 50 wins in a row. During that time period say you were -15 or +15 rating for each win or loss than at that instance you were either 750 points above or below your 50% win rate skill level. 
 

These instances in volatility are not inconsequential they have an effect on the entire systems behavior and it’s a problem. Like a kind of butterfly effect…because person X was in your game with a rating of 1200 but his actual skill level is that of a plat player, he’s gonna be placed on a team in which the match maker believes he is a noob. Because you have 9 other players here their rating is thus correlated to this persons botched rating…and if all players have botched ratings like this, then the entire system is indeed a statistical soup of highly correlated crap data.

You seem to like smart words. I recommend you to read about "Mathematical expectation" and "Law of large numbers".

[Shao.7236]

Too much overthinking. The lack of players in general with a good will for them to actually play the game is the reason why matches are so random and stupid.

 

MM does what it's supposed to, it does not think or reason with anything but the concept of having to make numbers even.

[zonias.1083]

On 4/23/2022 at 9:18 AM, DanAlcedo.3281 said:

What's worse is that skill doesn't matter too. 

If I get better, sure I will win a bit more but when I reach my new rank that I belong in, it is back to 50% win rate. 

This is exacly the goal you have. If you are in your skill level then you are supposed to get a 50% win rate in the long run and stay in the same rating. But if you improve then your win rate will get higher, the more you improve the higher it will be, but unless you become a top player overnight, then your winrate will eventually get to a 50%. 

 

I have exclusively played pvp, only for more than 3000 matches in ranked arena only and i have been able to maintain a slightly higher win rate than 50%.

 

Skill is what differenciates you in a match, of course a single player can not carry outragiously bad players, you are going to lose matches no matter you do.

[JusticeRetroHunter.7684]

On 4/26/2022 at 6:14 AM, Spellhunter.9675 said:

You seem to like smart words. I recommend you to read about "Mathematical expectation" and "Law of large numbers".

 

There's no such thing as a "smart word." You either know what words mean or you don't.

 

Now...I imagine that you actually don't know what those words mean....because if you did you'd realize that they don't prove or show the point you probably wanted to make. That's kind of what makes a person smart (in a subject) is whether they understand concepts, well enough to articulate and talk about them to any great depth.

 

For those that don't know, The Law of Large Numbers is the notion that events that occur with low probability...like say being struck by lightning strike, ... occurs more often, because the sheer number of events that could have otherwise happen are so many, that rare ones occur with more perceived frequency.

 

For example, about 40 people die every year due to lightning strikes, which you have about a 1/500,000 chance of being struck by lightning. There's billions of people in the world, and so you are not very likely to get hit by lightning let alone die from it but out there in the vast set of events occurring, 40 people got the "winning lottery ticket" so to speak every year. 

 

Let me put it this way. If you played 10,000 games, and every 100 games you get 50 wins and 50 losses in a row, you still have a 50/50 chance of winning or losing statistically...that is equivalent to just winning a game, and losing a game every 2 games. This is what Ragnar makes painfully obvious to the OP...and I guess it wasn't made obvious to you. If you play 10 million games, then the perceived frequency of these 50 game win and loss streaks go up not down because of the law of large numbers...in other words you are more likely to get these freak of nature win and loss streaks the more games that are played.

 

The point is that basically 1:1's and 50:50's happen at equally proportional frequencies, and that's exactly what Ragnar's statement means...the fact wins and loss streaks happen is because the two scenario's are equivalent. If you played only 50 games, then winning 50 games in a row is seems like extremely rare occurrence, because well... you only played 50 games. If you played 10 million games, than winning or losing 50 games in a row happens more often...because of the law of large numbers.

 

[JusticeRetroHunter.7684]

On 4/26/2022 at 6:14 AM, Spellhunter.9675 said:

You seem to like smart words. I recommend you to read about "Mathematical expectation" and "Law of large numbers".

 

Just an additional example to the above comment...to drive the point home and just end the discussion before it even starts because I can jus  feel the juju for how argumentative people get on this forum.

 

You've got a coin flip like game, where 0's are losses, 1's are wins. And you want to know the frequency with which a 3 win and loss streaks occur. You start by playing 3 games. the number of possible states are :

 

000, 001, 010, 100, 110, 011, 101, 111.

 

So if you played 8 sets of 3 games, you've got a 2 out of 8 chance to get a 3 game win or loss streak.

Now you up the amount of games you played, from 3 to 4, and again you want to know the frequency of 3 game win or loss streaks.  The number of possible states are:

 

0000, 0001, 0010, 0100, 1000, 0011, 0110, 1100, 0111, 1110, 1001, 1101, 1011, 0101, 1010 and  1111.

 

So if you played 16 sets of 4 games, the chances you'll get a 3 win or 3 loss streak is 6 out of 16 chances. It's also interesting to note the frequency of the idealized W/L scenarios: 1010 or 0101 which is a 2 out 16 chances...which occurs less often than the win/loss streaks.

 

The more games you play, the more the idealized 1 to 1 scenario drops to 0% chance of happening, because getting a the perfect string of 010101010101 and so on is just as unlikely as getting a perfect string of 11111111111 and 0000000000...because both scenarios are equivalent to each other...and in the space of possibilities, you are more likely to get equally proportionate distributions of wins and loss streaks...like 00011011110010111000110110 and so on.

[CalmTheStorm.2364]

On 4/26/2022 at 6:44 AM, Shao.7236 said:

Too much overthinking. The lack of players in general with a good will for them to actually play the game is the reason why matches are so random and stupid.

 

MM does what it's supposed to, it does not think or reason with anything but the concept of having to make numbers even.

This is only partly true. You're right that the MM is designed to work with a larger population, and that if we had more people (of all levels of skill) queuing at the same time, we would reliably have better quality matches.  But we don't live in a world with a massive PvP population at all hours of the day, so having a MM algorithm that depends on this is fundamentally flawed.

 

The problem is that the MM is coded such that it will find a game for everyone, even if the match quality is poor (see: https://wiki.guildwars2.com/wiki/PvP_Matchmaking_Algorithm ). It actually pads your MMR the longer you are in queue, making you eligible for more matches, even though you are less and less of a good fit for those matches the more padded your MMR becomes.  In other words, the matchmaker cares more about finding a match for everyone (and finding it as quickly as possible) rather than ensuring match quality. That's how you get games with legendary duo-queues against gold 2 players. If that legendary duo waits long enough, they will get a match, even if it's slaughtering a bunch of gold players.

 

And that's the core problem with MM. A small amount of padding could be ok, but there needs to be a hard limit to the padding. Legendary players, for instance, shouldn't get matched against anyone lower than p1. G3 shouldn't play against people lower than g1, etc.

 

People may complain that this could make queue times too long. Maybe it would; but too long is a matter of opinion. I for one would gladly wait 5-10 minutes for a good match. But in any event, that's changing the topic. If we are concerned about match quality (and we should be), we need a MM algorithm that says "no" to bad matchups, period.