Jump to content
You need to play a total of 20 battles to post in this section.
X01_ISAAC_7

How To Tell When A Match Is Lost Before It Begins

107 comments in this topic

Recommended Posts

288
[USA]
Beta Testers
360 posts

I just now realized something that I'm figuratively kicking myself for not noticing sooner. Literally in almost all the blow-out matches that I lose, I'm facing enemy teams that look like the one in this. And in all the blow out matches I win, I'm ON teams that look like the enemy team in this.

And what's even worse is divisions obviously aren't properly taken into account in the matchmaker. 2 of the 3 people in the enemy division are rank 1, and the third is rank 4.

FIVE rank 1 players on the enemy team. FIVE. Count them. FIVE.
Two rank 4s, three rank 8s, and one rank 9.

Versus...

One rank 2, one rank 4, myself at 11, and then one each at 17 and 24.

How is this balanced matchmaking again? How hard would it friggin be to code the matchmaker to take the Rank of a person from the current/previous ranked season into account? I'm not talking this silly "ranked sprint" or even the "1 vs. 1 ranked sprint" thing they had months back either. I'm talking legitimate ranked seasons. Seriously, I'm so sick and tired of losing blow-out matches, and honestly I'm sick of WINNING blow-out matches too. It couldn't possibly be that :etc_swear: hard to force a further balancing factor around a person's ranking from ranked battles so that teams are more evenly distributed.

Sure, losing close matches sucks really REALLY badly. But at least I've got a chance to do something in said match. Likewise, winning those close matches is really REALLY amazing.

 

I can't POSSIBLY be the only person here who hates constantly getting in losing matches where the loss isn't even close, but literally just one huge blowout where your team crumples like wet newspaper.

 

Edit: Keep in mind, the only reason the enemy Dallas is at the bottom is because I somehow got lucky, RNGesus blessed my shells, Stalin himself reached his cold hand up from the depths of Hell, and guided 3 of my shells into the Dallas' citadel and a few others into the hull while saying, "I like you, go to Gulag!" and I got a rare DevStrike with a battleship.

MM Balans Komrad.JPG

Edited by X01_ISAAC_7
Added the edit.
  • Cool 7
  • Meh 3

Share this post


Link to post
Share on other sites
991
[TIMT]
Members
1,707 posts
5,589 battles

I know the pain, and it is really frustrating to see this happening.

However, the issue is not with coding those constraints into the MM but the longer waiting time resulting from it. Given enough players and a good distribution of tiers/ships this might not be an issue, but at odd hours or lower tiers with less players queuing up it might take minutes before a match can start. Sure, they could also try swap between the two teams once 24 players are found, though this might not have the desired effects all around.

Admittedly, the more I thought about the issue, the more have I come the realization that the MM is flawed, yes, but making a better one is very hard without running into significant side effects.

1. Balancing on winrate/PR/ arbitrary metric defining skill puts better than average players always in an uphill battle whereas lower skill players get boosted. While this also happens in random MM, the systematic punishment for being good could have bad consequences.

2. Balancing on winrate/PR/ arbitrary metric by introducing some form of leagues/ranks etc. in the random queue would lead to excessively long waiting times. How many players with exactly your skill level are out there at any given time?

3. Attempts to balance the skill of players in a team lead to a tightening of the metric, i.e. the spread of for example win rates will be much smaller. Afterall, if all the matches are made so that the teams have equal WR the fact that high/low WR will always play against high/low WR will result in everyone having more or less the same WR after a while. At that point the system no longer works.

There are many more things to consider. I agree the MM is flawed in that the random nature sometimes confronts you with a Kobayashi Maru test, but so far I have not seen a better and practical solution to this. It just sucks.

  • Cool 2

Share this post


Link to post
Share on other sites
288
[USA]
Beta Testers
360 posts

Would it really have been that hard to swap the New Orleans for the Atlanta, the Colorado for the Gneisenau, and the Amagi for the un-divisioned No Cal in this match though? Would it honestly have been that hard to do?

I for one wouldn't mind waiting longer in the queue if it meant I actually got a match where I felt like I could actually DO SOMETHING rather than getting 6 minutes in and suddenly wonder why I'm starting to get shot at from behind me only to turn around and find out that the entire other flank just was wiped off the face of the universe like it never existed, resulting in the other half of the enemy team coming to give me surprise bu-..... uhh... well, you know.

  • Cool 3
  • Haha 1

Share this post


Link to post
Share on other sites
2,682
[O7]
Members
1,513 posts
11,692 battles

Random battles will never have any competitive edge to it, it is just random battles for random people. You could have a triple o7 div on your team and the other team could have a coop only player who accidentally entered random battles, matchmaker doesnt care.

There never will be skill based matchmaking, if thats what you want play clan battles. People really need to stop complaining about this.

  • Cool 5
  • Thanks 3
  • Meh 8

Share this post


Link to post
Share on other sites
991
[TIMT]
Members
1,707 posts
5,589 battles
9 minutes ago, X01_ISAAC_7 said:

Would it really have been that hard to swap the New Orleans for the Atlanta, the Colorado for the Gneisenau, and the Amagi for the un-divisioned No Cal in this match though? Would it honestly have been that hard to do?

 

No, it would not have taken that much longer. But my point with that was that the unintended long-term side effect of such a mechanic would probably be that balanced 'skill' in both teams would lead to better than average players being constantly punished, whereas less than average players always get boosted. This would push everyone towards average, at which point the system might break down. I am not saying that is happening immediately, and you could probably make the constraints 'soft' in that they are not always applied.

Nevertheless, it would mean that the average battle result (WR) of all players is no longer based on a combination of their skill and RNG, but actively driven to mediocrity. And this seems really dangerous to me.

Edited for context

Edited by shinytrashcan
  • Cool 1
  • Meh 1

Share this post


Link to post
Share on other sites
7,343
[WORX]
Members
13,225 posts
20,086 battles

You do know OP, Rank is useless for Random... Its not and will never be a measuring stick of a good team... Its RANDOM...

Rank only means, you love pain or to feel pain is fun.. OR you like the pain of winning...

Anyone can quantify a definition of a great team VS bad team.. However, in the end, its their opinion over the mechanics of RANDOM MM.

Edited by Navalpride33
  • Cool 2
  • Haha 1
  • Meh 2

Share this post


Link to post
Share on other sites
Volunteer Moderator, Volunteer Moderator
435 posts
5,648 battles
15 minutes ago, Navalpride33 said:

Its not and will never be a measuring stick of a good team... Its RANDOM...

I agree with this statement, I've had teams with multiple "high ranked" individuals and a few purple clan peeps, but we got slaughtered none the less. Being amazing in ranked doesn't always equate to being amazing in randoms since the gameplay is so different than random battles in more ways than one, the same goes for clan battles vs randoms. Its just a different playing field. All it takes is one bad plan, open flank, etc that your team commits to and BAM, it's over before you realize. 

  • Boring 1

Share this post


Link to post
Share on other sites
1,058
[OIL-1]
Members
1,450 posts

This is what a losing battle looks like.
Do the math. My team = Rank 64. Enemy team Rank = 40. (lower is better.)

Losing.thumb.jpg.87c63d4bd0839605d7d64976e041be07.jpg

This is what a winning battle looks like.
Do the math. My team = Rank 55. Enemy team Rank = 44. (lower is better.)

Winning.thumb.jpg.d7eaa926581f6e7d336cc14ef2315044.jpg

Between this game and the tank game, I have 40,000 battles.
My buddy irl has 60,000 battles.
That's a sample of 100,000 battles right there. And this crap happens 90% of the time.
Random?

If it was really random, like these shills want you to believe, then why did WG hide it?
You 'used' to be able to see ranks before the start of a battle, now you can't.
It's because too many players were catching on, and most of them are too lazy to check the team score at the end.

It's the exact same 90% if you go by win ratios.

  • Cool 2
  • Meh 1

Share this post


Link to post
Share on other sites
3,423
[META_]
Members
2,282 posts
7,823 battles
1 hour ago, X01_ISAAC_7 said:

I just now realized something that I'm figuratively kicking myself for not noticing sooner. Literally in almost all the blow-out matches that I lose, I'm facing enemy teams that look like the one in this. And in all the blow out matches I win, I'm ON teams that look like the enemy team in this.

And what's even worse is divisions obviously aren't properly taken into account in the matchmaker. 2 of the 3 people in the enemy division are rank 1, and the third is rank 4.

FIVE rank 1 players on the enemy team. FIVE. Count them. FIVE.
Two rank 4s, three rank 8s, and one rank 9.

Versus...

One rank 2, one rank 4, myself at 11, and then one each at 17 and 24.

How is this balanced matchmaking again? How hard would it friggin be to code the matchmaker to take the Rank of a person from the current/previous ranked season into account? I'm not talking this silly "ranked sprint" or even the "1 vs. 1 ranked sprint" thing they had months back either. I'm talking legitimate ranked seasons. Seriously, I'm so sick and tired of losing blow-out matches, and honestly I'm sick of WINNING blow-out matches too. It couldn't possibly be that :etc_swear: hard to force a further balancing factor around a person's ranking from ranked battles so that teams are more evenly distributed.

Sure, losing close matches sucks really REALLY badly. But at least I've got a chance to do something in said match. Likewise, winning those close matches is really REALLY amazing.

 

I can't POSSIBLY be the only person here who hates constantly getting in losing matches where the loss isn't even close, but literally just one huge blowout where your team crumples like wet newspaper.

 

Edit: Keep in mind, the only reason the enemy Dallas is at the bottom is because I somehow got lucky, RNGesus blessed my shells, Stalin himself reached his cold hand up from the depths of Hell, and guided 3 of my shells into the Dallas' citadel and a few others into the hull while saying, "I like you, go to Gulag!" and I got a rare DevStrike with a battleship.

MM Balans Komrad.JPG

this is what a bad day in ranked actually looks like... if it wasn't horrific MM it was horrific MM coupled with [edited] up [edited] RNG

it was, at one point, so bad a green player asked me how i could miss a shot on a broadside CA... i didn't, they all hit. it was all overpens, for 1310 damage, again. 

shot-20_06.30_01.05_06-0205.thumb.jpg.943689ff159b239c9f8bbec02fd413f5.jpg

NONE of the losses were close. on a few occasions we got off to a good start, capped a couple, in 3 cases capped all 3....  only one match went the distance, and went 5 battles in a row with no kills.

Edited by Spud_butt
  • Cool 6
  • Sad 1

Share this post


Link to post
Share on other sites
3,576
[-K-]
Beta Testers, In AlfaTesters
7,913 posts
10,311 battles
1 hour ago, X01_ISAAC_7 said:

FIVE rank 1 players on the enemy team. FIVE. Count them. FIVE.

Two rank 4s, three rank 8s, and one rank 9.

Versus...

One rank 2, one rank 4, myself at 11, and then one each at 17 and 24.

Ranks aren't a good indicator of skill, though there is a correlation.

Use MMM if you're looking for than informatkion, it'll give you a much better idea of the expected match outcome.

I recently had this game; the average account win rate for teams differed by more than 20 percentage points (!). Naturally, we won.

4pMXxxW.png

But of course, you should still try to do your best. Upsets are always possible.

  • Meh 1

Share this post


Link to post
Share on other sites
1,058
[OIL-1]
Members
1,450 posts
11 minutes ago, Spud_butt said:

shot-20_06.30_01.05_06-0205.thumb.jpg.943689ff159b239c9f8bbec02fd413f5.jpg

I've seen that more times than I can count. Same for the tank game. If it was random, the odds would be about 50%. That doesn't look like 50% to me.

some say it will balance out. How often do you flip a coin and see tails come up that much?

  • Cool 3
  • Boring 1

Share this post


Link to post
Share on other sites
Beta Testers
1,830 posts

one thing I do dislike about the current MM is how it will match divisions of 2 t8 players against a division of 3 TX players.

  • Cool 1

Share this post


Link to post
Share on other sites
3,576
[-K-]
Beta Testers, In AlfaTesters
7,913 posts
10,311 battles
4 minutes ago, z9_ said:

If it was random, the odds would be about 50%

First of all, the fact that outcomes for a certain cluster doesn't mean that these events aren't random. Refer to this Wiki article: https://en.wikipedia.org/wiki/Clustering_illusion

Second of all, for a large number of games (starting at ~hundreds), odds are only 50% if you're an average player. If you're below average, your odds are below 50%, if you're above average then they're above 50%.

  • Cool 3
  • Meh 1

Share this post


Link to post
Share on other sites
10,396
[INTEL]
Members
13,459 posts
39,002 battles
5 minutes ago, Dareios said:

one thing I do dislike about the current MM is how it will match divisions of 2 t8 players against a division of 3 TX players.

heh...and produce games in which one side has one more division than the other. Again and again... 

...all easily fixable. 

  • Cool 2

Share this post


Link to post
Share on other sites
10,396
[INTEL]
Members
13,459 posts
39,002 battles
24 minutes ago, Spud_butt said:

this is what a bad day in ranked actually looks like... if it wasn't horrific MM it was horrific MM coupled with [edited] up [edited] RNG

it was, at one point, so bad a green player asked me how i could miss a shot on a broadside CA... i didn't, they all hit. it was all overpens, for 1310 damage, again. 

shot-20_06.30_01.05_06-0205.thumb.jpg.943689ff159b239c9f8bbec02fd413f5.jpg

NONE of the losses were close. on a few occasions we got off to a good start, capped a couple, in 3 cases capped all 3....  only one match went the distance, and went 5 battles in a row with no kills.

Man, how come you didn't quit after the second loss? That's when my tolerance runs out, and I go play another game. 

Share this post


Link to post
Share on other sites
7,343
[WORX]
Members
13,225 posts
20,086 battles
On 6/30/2020 at 1:01 AM, vak_ said:

Second of all, even for a large number of games (starting at ~hundreds), odds are only 50% if you're an average player. If you're below average, your odds are below 50%, if you're above average then they're above 50%.

Actually, statistics for all WOWS server states....

  • The %45 to %55 WR block account for %90 of the server population...

This means, AVG. below AVG. Above AVG players... All reside in this block...

Anyone below this block account for %5 of the population.. Which means they're starting out or not even trying.

At the %56 and above account for the other %5 of the population... This means they're doing other measures (like DIV up with other players).

The deviation is very small across all server regions... Which means, in your playing session... You will find players in the %90 percentile block in the MM queue population.

The statistics do not mean/suggest what your statement^^^ above  suggest/state.

Edited by Navalpride33
  • Cool 3
  • Meh 2

Share this post


Link to post
Share on other sites
3,576
[-K-]
Beta Testers, In AlfaTesters
7,913 posts
10,311 battles
8 minutes ago, Navalpride33 said:

The %45 to %55 WR block account for %90 of the server population...

In the context of the discussion, when I say average, I mean truly average -- i.e. if you get random perfectly matched teammates and enemies, your odds of winning are exactly 0.50.

Point is, the assumption about 50% is incorrect unless we're talking a very thin portion of the skill distribution, right in the middle of the curve.

Share this post


Link to post
Share on other sites
1,058
[OIL-1]
Members
1,450 posts
34 minutes ago, vak_ said:

First of all, the fact that outcomes for a certain cluster doesn't mean that these events aren't random. Refer to this Wiki article: https://en.wikipedia.org/wiki/Clustering_illusion

Second of all, for a large number of games (starting at ~hundreds), odds are only 50% if you're an average player. If you're below average, your odds are below 50%, if you're above average then they're above 50%.

The article you cited refers to small sample sizes, and the examples it gave were completely flawed.
V-1 flying bombs can be adjusted simply by changing the amount of fuel you put in them.
Stock market chartists are interesting, but only offer 1 piece of information, although that piece is important,
because if enough people believe it, it will move a market, not to mention that the bulk of the money
is managed by strict funds, that are required to buy/sell based on a 120 day moving average, etc.

Now let's look at players who are above/below average.
Are you familiar with the Bell Curve?
If the odds in Vegas are 3% for the house, and you gamble long enough, you will probably lose 2.5-3.5% of your money.
That's true for over 99% of the players.
There are a very rare number of players who win all the time, and a very rare number of players who lose all the time.

So, 100,000 sample battles from me, and countless battles from players who long ago left WG.
As a side note, spoiler alert, WG isn't the only game that pulls stunts like that.
And stacking the teams is just one stunt. There are a whole lot more.

  • Meh 1

Share this post


Link to post
Share on other sites
3,576
[-K-]
Beta Testers, In AlfaTesters
7,913 posts
10,311 battles
25 minutes ago, z9_ said:

The article you cited refers to small sample sizes

Which doesn't invalidate the random occurrence of win and loss streaks in WoWS.

25 minutes ago, z9_ said:

V-1 flying bombs can be adjusted simply by changing the amount of fuel you put in them.

I think you misunderstand what is randomness. While the exact location of V-1s strikes is dependent on a myriad of deterministic factors (technically there is quantum indeterminacy, but let's pretend it doesn't exist for now); however, the combination of these many-many factors produces results that appear random to an independent observer, and can be treated as such.

25 minutes ago, z9_ said:

Are you familiar with the Bell Curve?

Given that I've published papers where results were based on various statistical analyses of my experimental data -- I would dare say, yes, I'm quite familiar with the normal distribution curve, among other things.

25 minutes ago, z9_ said:

If the odds in Vegas are 3% for the house

What's this analogy supposed to demonstrate? What's the "house" here, draws? We hardly ever get them these days.

25 minutes ago, z9_ said:

So, 100,000 sample battles from me, and countless battles from players who long ago left WG.

Your stats are hidden, so I can't validate this statement of yours. However, this would imply that you played, on average, more than fifty battles every single day since Open Beta. You sure you wanna stand by that?

25 minutes ago, z9_ said:

And stacking the teams is just one stunt. There are a whole lot more.

You can believe that WeeGee is out to get you all you want. Unless you can provide objective proof of your claims, it's nothing more than just another ill-informed conspiracy theory.

  • Cool 1

Share this post


Link to post
Share on other sites
3,423
[META_]
Members
2,282 posts
7,823 battles
14 hours ago, Taichunger said:

Man, how come you didn't quit after the second loss? That's when my tolerance runs out, and I go play another game. 

stupid. no, stubborn. NO, stupidly stubborn.... ahhhhh soothingly refreshing.

actually, i have limited time to play this week, and ranked sprint won't last forever.

also, i have a morbid fascination that i turn loose when the losses start rolling in....

mostly, i have my best battles in the losses. some super unicum type battles that i wonder how they would look if it was a win? 

it's hard to carry a great team to a win, they will do it with or without you. those few times i pull a victory out of a sure fire slaughter defeat, i KNOW it was a great effort. 

never had a winning streak over 6, ever, in ranked, clam, or random. routinely get losing streaks over 6 (8 posted here for ranked, and it's been as long as 15 in randoms) absolutely LOVE it when the statisticians start posting how it is normal, and average, to have losing streaks twice as long as winning streaks. lmbo, sure it is. 

i think the funniest number of the whole string was when i got 187 secondary battery hits, and started 1 fire for zero damage. 142 2ndary hits with zero fires was a close second. i know, i know, if they have the captain's skill, it reduces the 8% fire chance down to negative numbers, it should be starting fires on my own ship, and i should be happy to have them! brrrr  it gets cold in the southern pacific...

 

  • Cool 5

Share this post


Link to post
Share on other sites
654
[NSEW]
Members
2,477 posts
12,059 battles
16 hours ago, ITZ_ACE_BABY said:

Random battles will never have any competitive edge to it, it is just random battles for random people

This is very true.  Even WG themselves used the wording/usage of "Random" as an excuse (for the lack of better word).  To justify/explain the reasoning behind "Random" matches.  

Now if it was relabelled as "PvP", it would be difference I hazard to guess.

 

 

Share this post


Link to post
Share on other sites
3,286
[GOB]
Members
2,736 posts

The ONLY way to truly KNOW that a game is lost even before it stats is people think like this and start spamming chat with  "This is a horrible team..". "We are going to lose"   SURE loss!  

The people who not only QUIT as soon as they see a ship die but also feel the need to drag everyone else down with their loser attitude.. yeah as soon as I see them it's MAJOR Carry time or lose!

Share this post


Link to post
Share on other sites
97
[BOTES]
[BOTES]
Members
28 posts
15,133 battles

Sometimes you get stuff like your pic, and then you get stuff like this. Such is the joy of playing randoms.

1146802388_yanggang.thumb.PNG.a6e24487cda66e5eab5ed66009fe5d88.PNG

Share this post


Link to post
Share on other sites
609
[BOTO]
Members
1,460 posts
18,147 battles
22 hours ago, X01_ISAAC_7 said:

I just now realized something that I'm figuratively kicking myself for not noticing sooner. Literally in almost all the blow-out matches that I lose, I'm facing enemy teams that look like the one in this. And in all the blow out matches I win, I'm ON teams that look like the enemy team in this.

And what's even worse is divisions obviously aren't properly taken into account in the matchmaker. 2 of the 3 people in the enemy division are rank 1, and the third is rank 4.

FIVE rank 1 players on the enemy team. FIVE. Count them. FIVE.
Two rank 4s, three rank 8s, and one rank 9.

Versus...

One rank 2, one rank 4, myself at 11, and then one each at 17 and 24.

How is this balanced matchmaking again? How hard would it friggin be to code the matchmaker to take the Rank of a person from the current/previous ranked season into account? I'm not talking this silly "ranked sprint" or even the "1 vs. 1 ranked sprint" thing they had months back either. I'm talking legitimate ranked seasons. Seriously, I'm so sick and tired of losing blow-out matches, and honestly I'm sick of WINNING blow-out matches too. It couldn't possibly be that :etc_swear: hard to force a further balancing factor around a person's ranking from ranked battles so that teams are more evenly distributed.

Sure, losing close matches sucks really REALLY badly. But at least I've got a chance to do something in said match. Likewise, winning those close matches is really REALLY amazing.

 

I can't POSSIBLY be the only person here who hates constantly getting in losing matches where the loss isn't even close, but literally just one huge blowout where your team crumples like wet newspaper.

 

Edit: Keep in mind, the only reason the enemy Dallas is at the bottom is because I somehow got lucky, RNGesus blessed my shells, Stalin himself reached his cold hand up from the depths of Hell, and guided 3 of my shells into the Dallas' citadel and a few others into the hull while saying, "I like you, go to Gulag!" and I got a rare DevStrike with a battleship.

MM Balans Komrad.JPG

There's no correlation between a player's rank and their skill level.  Even the worst potato can grind/get carried to rank 1 if they play enough games.

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.

×