Placement match Abandonment counts?

[Kindeller.3072]

Just curious, playing my first placement match there and we got abandoned after the first mid fight leaving a snowbally 4v5.. we lost. I know D/C gives rating protection if you lose but it still showed a loss in my placement matches... will this affect my league placement? If so should it be null and void for placement matches, even if rating is protected?

[Bazsi.2734]

Your first ten matches always count, for some reason the rating protection doesn't apply to those. Sadly you have to put up with the unfair losses.

[BeLZedaR.4790]

It doesn’t count. Your 11th match (where you already see the rating) will be your 10th placement.

Even in placements you have rating, you just don’t see it because it goes up/down by like 100 each game and would be super depressing to lose.

[MyPuppy.8970]

You'll get where you belong eventualy so don't worry too much about it, and have fun!

[Farkon.2170]

What's worse is when we have verified bots instead of DCs.

[Math.5123]

@Bazsi.2734 said:Your first ten matches always count, for some reason the rating protection doesn't apply to those. Sadly you have to put up with the unfair losses.

This is straight up false. Connect your API to gw2effi and you can see them exact MMR gain / loss, even in placements.

[Kindeller.3072]

@"AngelLovesFredrik.6741" said:

@Bazsi.2734 said:Your first ten matches always count, for some reason the rating protection doesn't apply to those. Sadly you have to put up with the unfair losses.

This is straight up false. Connect your API to gw2effi and you can see them exact MMR gain / loss, even in placements.

So with this knowledge, has anyone tested if rating protection kicks in when in placement and someone d/c's or quits?

[Ben Phongluangtham.1065]

Placement matches are pretty much the same as any other season ranked match. You don't lose skill rating if someone on your team has been disconnected for at least 90 seconds.

[zoopop.5630]

@"Ben Phongluangtham.1065" said:Placement matches are pretty much the same as any other season ranked match. You don't lose skill rating if someone on your team has been disconnected for at least 90 seconds.

However You Lose Rating If said Player Decides to Just "troll" for the first 90 seconds + of a game by standing in base but not going afk.

Theirs needs to be a system where the system can see if a player even trying by reading if their just in base and not even playing. Theirs a few holes to the "afk" honor system.

[Kindeller.3072]

@Ben Phongluangtham.1065 said:Placement matches are pretty much the same as any other season ranked match. You don't lose skill rating if someone on your team has been disconnected for at least 90 seconds.

Thanks for the clarification Ben :)

[Demonftw.4705]

@Ben Phongluangtham.1065 said:Placement matches are pretty much the same as any other season ranked match. You don't lose skill rating if someone on your team has been disconnected for at least 90 seconds.

still shows as a loss during placements. I tested this on 2 accounts last season, same region. I had a bunch of games where people DC'd while on the 1st account and none on the 2nd account when doing placements. I ended with 250 points higher than account1 even though I had more losses on account2.Account1 had 5W 3L 2DC(not me)Account2 had 5W 5LThis was last season though. I refuse to play this season just because how frustrating placement games are, no discredit to you as a dev, you're the only one I see that communicates and I'm sure you give your best.

[phs.6089]

@Demonftw.4705 said:

@Ben Phongluangtham.1065 said:Placement matches are pretty much the same as any other season ranked match. You don't lose skill rating if someone on your team has been disconnected for at least 90 seconds.

still shows as a loss during placements. I tested this on 2 accounts last season, same region. I had a bunch of games where people DC'd while on the 1st account and none on the 2nd account when doing placements. I ended with 250 points higher than account1 even though I had more losses on account2.Account1 had 5W 3L 2DC(not me)Account2 had 5W 5LThis was last season though. I refuse to play this season just because how frustrating placement games are, no discredit to you as a dev, you're the only one I see that communicates and I'm sure you give your best.

I can support this, whatever happens in placement matches -COUNTS

[Khalisto.5780]

This is more like a technical question, because it barely affects how high you gonna be placed. I say that because i let my little nephew that have never played the game do my placement matches. He went 4/6 a was placed low silver. When I started palying the game kinda knew I was better the other players and skyrocketed me to gold 3 where I belong. Took me just 15 matches to climb there. I was getting 25-33 pts every game, a lot different than the 10-17 I'm getting now. It's 2/3x faster when you playing out of your bracket of skill. The game seems to adjust for ppl not play out of their rank for much time, and not let less experienced players get frustrated.

[Exedore.6320]

@"zoopop.5630" said:However You Lose Rating If said Player Decides to Just "troll" for the first 90 seconds + of a game by standing in base but not going afk.Theirs needs to be a system where the system can see if a player even trying by reading if their just in base and not even playing. Theirs a few holes to the "afk" honor system.If you base the AFK rule on being AFK in base, then those AFK people will now run out, stand off-point, and die a few times. Catching throwers isn't that simple.

@"Demonftw.4705" said:still shows as a loss during placements. I tested this on 2 accounts last season, same region. I had a bunch of games where people DC'd while on the 1st account and none on the 2nd account when doing placements. I ended with 250 points higher than account1 even though I had more losses on account2.Account1 had 5W 3L 2DC(not me)Account2 had 5W 5LThis was last season though. I refuse to play this season just because how frustrating placement games are, no discredit to you as a dev, you're the only one I see that communicates and I'm sure you give your best.The system has to take credit for the match somehow. ANet's system can only display "win" or "loss", even though it tracks matches which don't count.

Your example is flawed. Who you play when you win or lose matters more than the total number of wins and losses. 250 rating is close to the margin of error in the rating system. It's likely that account1 suffered from less information than from counting the DC games.

[Abazigal.3679]

@Demonftw.4705 said:

@Ben Phongluangtham.1065 said:Placement matches are pretty much the same as any other season ranked match. You don't lose skill rating if someone on your team has been disconnected for at least 90 seconds.

still shows as a loss during placements. I tested this on 2 accounts last season, same region. I had a bunch of games where people DC'd while on the 1st account and none on the 2nd account when doing placements. I ended with 250 points higher than account1 even though I had more losses on account2.Account1 had 5W 3L 2DC(not me)Account2 had 5W 5LThis was last season though. I refuse to play this season just because how frustrating placement games are, no discredit to you as a dev, you're the only one I see that communicates and I'm sure you give your best.

I think the order of your wins/loses could also influates on the outcome. You know, 10 games following placement do have a huge impact. It can be like + 30 -27 -25 + 25 +22, etc until it settls around -12/+12.Maybe there is a kind of similar thing on the 10 placements. Like first matchs counts for +/-150, then +/-100 then +/-75 ( these are obviously random numbers). Then, if you start with 5 wins followed by 5 loses, you will be placed much higher than if you start with 5 loses followed by 5 wins.

That's however only an idea i have, this might be totally wrong.

[Zenix.6198]

it makes sense to have some form of calibration matches.if you want any algorithm to accurately rate player performance you have to feed it with some form of data....it cant just randomly place people somewhere.

That being said however, I would be interested in what the "default rating" is.Imagine a compeltely new account with. PvP level 20, quing for ranked the very first time.Lets assume the default ranking is 1.2k rating.So first game is played on ~1.2k rating. You won ==> try vs 1.5k (300 more). Game lost ==> try vs 1.3k (200 less). Game won ==> try vs 1.4k (100 more) ...etc .etc.Personally I think this process makes a lot of sense.

For people that already played last season, it also makes sense to modify their "default ranking" ...cause hard-resets are very messy.So I assume, that last seasons rating has some impact on the players starting ranking.For somebody finishing at 1.7k last season it could look something like this.1200 (=assumed default rating) + ((1700 - 1200) / 2 ) = 1450 and then start the calibration from there.

[Exedore.6320]

@"Zenix.6198" said:That being said however, I would be interested in what the "default rating" is.1200 is correct.For people that already played last season, it also makes sense to modify their "default ranking" ...cause hard-resets are very messy.Hard resets are a mess and are only necessary when the rating system had major bugs or was prevented from working properly (such as going from pips to pure rating from season4 to season5).

ANet does a soft reset for the start of the each season. It averages you rating from the last season with 1200 (the middle rating) and I believe resets deviation and volatility to their maximums. I question the need for doing that, however.

[Demonftw.4705]

ANet does a soft reset for the start of the each season. It averages you rating from the last season with 1200 (the middle rating) and I believe resets deviation and volatility to their maximums. I question the need for doing that, however.

for having played this game mode for 6 years and consistently playing at plat/ legend, seeing the match making evolve over all these seasons. I can very well assure you that in my experience that it doesn't feel like there is any soft reset. Your logic on whole is flawed and you have no examples or proof to back these claims you make.

The system has to take credit for the match somehow. ANet's system can only display "win" or "loss", even though it tracks matches which don't count.

Your example is flawed. Who you play when you win or lose matters more than the total number of wins and losses. 250 rating is close to the margin of error in the rating system. It's likely that account1 suffered from less information than from counting the DC games.

If you've played any game which has a decent matchmaking system for competitive modes, you will notice that if you've had someone abandon a game in your placements then you have to play another game till you finish the required quota. This may or may not be hard to implement and that is something the player base cannot control. The point of my original post was to point out that placements don't actually work as Ben claims, which still stands.

[Exedore.6320]

@"Demonftw.4705" said:

ANet does a soft reset for the start of the each season. It averages you rating from the last season with 1200 (the middle rating) and I believe resets deviation and volatility to their maximums. I question the need for doing that, however.

for having played this game mode for 6 years and consistently playing at plat/ legend, seeing the match making evolve over all these seasons. I can very well assure you that in my experience that it doesn't feel like there is any soft reset. Your logic on whole is flawed and you have no examples or proof to back these claims you make.Please, make up more shit based purely on your own observation bias. ANet said in a forum post that they do the soft reset every season. If that's not proof enough, nothing is.

Soft reset just means that high gold and legendary start the season with closer ratings. After 20 or so games, they'll spread out to close to their original ratings.

The system has to take credit for the match somehow. ANet's system can only display "win" or "loss", even though it tracks matches which don't count.

Your example is flawed.

Who

you play when you win or lose matters more than the total number of wins and losses. 250 rating is close to the margin of error in the rating system. It's likely that account1 suffered from less information than from counting the DC games.

If you've played any game which has a decent matchmaking system for competitive modes, you will notice that if you've had someone abandon a game in your placements then you have to play another game till you finish the required quota. This may or may not be hard to implement and that is something the player base cannot control. The point of my original post was to point out that placements don't actually work as Ben claims, which still stands.Again, you've provided no proof, just a claim backed only by observation bias. If you supply the order of wins, losses, and DCs, it would help more in determining what happened. Losing early placement matches has a lot bigger impact on final rating than losing later placement matches.

And to counter your claim, I believe Overwatch still counts matches where people rage quit for placements. The only matches which don't count are where someone doesn't load in initially or leaves in the first minute or so and it completely cancels the game.