Jump to content
Sign in to follow this  
You need to play a total of 50 battles to post in this section.
Bratyn

Statistically impossibly bad matchmaking?

78 comments in this topic

Recommended Posts

[PDXND]
Players
32 posts
6,134 battles

Hi all,

I've been a player on WoWs every since release now, and I've never posted on these forums. Evidently, something extraordinary has to happen for me to get active on the forums... What happened wasn't a good thing, and if you'll pardon me - I'm going to rant a bit.

 

I don't consider myself a bad player. I hover around 55% winrate at over 4k games played, but this is including two years of me not really paying attention to the game while playing (i.e. youtube in the background, or other distractions that caused me to perform badly).

For years I only played seriously whenever I tried to get into ranked. The first few ranked seasons went swimmingly. I got 3 rank 1s relatively easily within the first 4 ranked seasons (I was too much of a noob to even have a high enough tier ship for the first season), and unlocked the flint. I reckoned I was only 2 ranked seasons away from the Black. Over the next ranked seasons I missed out on one, and the others... Well... I just could not get to rank 1. No matter how hard I tried. I remember one ranked season I reached rank 2.3 with 60% winrate in about 200 games... And then lost so many games in the next 400 (yes, you read that right, FOUR HUNDRED) that my end score was 43% winrate and I ended at rank 7.2. That experience was so infuriating I actually sent in a rant directly to Wargaming, knowing full well they would not be able to do anything about my case in particular. I figured at least it'd make them aware of my dissatisfaction.

That was last summer. Since October however I've gotten much more into random games. I ground up in total 12 lines to T10, and I really started enjoying non-competitive gaming. I started patching up my bad stats from years of playing with distractions, and things were going well. Until I reached 55% winrate. For the past month, I've had days where I lost 10 games straight. It seems like every time I get above 55% I just lose everything until I am back down to 54.90%-ish. The last batch of these losses made me look at my latest-played games, and, well... In the first 7 games of my Jutland I had 65% winrate. I love the ship. Turns out that now, with 33 games under my belt and having unlocked the Daring, I only have 42.42% winrate left. My Richelieu, with 26 games, only had 46.15% winrate. My Missouri, an abysmal 38.1% winrate with 42 games. Akizuki - love this ship, but 41.38% winrate (on the Kitakaze, which I enjoyed perhaps even more, but which is not substantially different in playstyle, I have a 70.83% winrate, for comparison). The list goes on. See how fickle things are in the picture below. A week ago I still had enjoyable games, games I sometimes had to carry, but games I -could- carry. Now it seems matchmaking is punishing me for them. How the hell does this keep happening?

 

image.thumb.png.f79b6d9222dcebbf7342442c68bcc2c3.png

You'd be forgiven for thinking I was a bad player, judging by my winrates in these ships. But my total winrate (even ruined through years of mis-play) shows a clear ~55% winrate. I consistently (of course with exceptions) place top of my team every battle in these losses. I consider myself very capable in any ship class except CV (which I never play). I have the skills to carry, but seemingly last week (and in periods throughout the last month) I have not even been given teams competent enough to allow me to carry them. Besides, how does my performance on one flank of the map in any way affect our other flank melting? My winrate peaked at 55.20% in December, to muddle around 54.90% for weeks after, until I managed to drag it back up to 55.15% last weekend. Now I'm back down to 54.89%. It's infuriating. It hurts to see these ships that I felt I genuinely performed well in, and which I very much enjoyed playing (Akizuki and Jutland were some of my favorite ships, and Richelieu and Missouri I very much enjoy as well), with such atrocious winrate stats. It's genuinely embarrassing. I've screamed at my computer. I've tried to 'guide' my team to victory with being helpful in chat, but usually their refusal to even acknowledge what I write, or outright hostility, mostly just means this devolves into toxicity.

Even if I was a middling player at best, you'd expect me to win at least half of my games... Middling fourties, or even upper thirties, in winrate on these ships is just atrocious. Considering I'm substantially better than the average player, how the hell is it possible that I am consistently so unlucky with my winrates? This doesn't even seem statistically possible. 

I want to play this game more, and I want to keep like this game. Why is the game doing everything it can to try to get me to quit and stop playing?

 

How do you deal with periods like this? How do you deal with atrocious winrates, or evaporating teams? Am I insane for starting to suspect Matchmaking was randomly updated a week ago to ruin it for me?

If you read this far, well... Kudos to you for making it through my rant :D

- Bratyn

  • Cool 1
  • Boring 1

Share this post


Link to post
Share on other sites
Beta Tester
4,381 posts
9,616 battles

You're refering to WRs in ships you have so few games in that a single game is worth multiple percentage points up or down. There's nothing statisticly odd about the error margin being large with such small samples, as you're only a few battles away from being at your usual WR, or higher. You do also conveniently ignore that there's ships you have that are higher than your average.

 

As for getting to a certain WR and not going higher? That happens to everyone. You're just getting to the point where you can't improve much yourself anymore and where you are is where you are, barring some massive change in your personal skill, or in you selectively starting to grind for WR (sealclubbing, only playing the ships you do best in), or in you starting to division up with players as good as you or better.

  • Cool 2

Share this post


Link to post
Share on other sites
[PDXND]
Players
32 posts
6,134 battles
1 minute ago, AgarwaenME said:

You're just getting to the point where you can't improve much yourself anymore and where you are is where you are, barring some massive change in your personal skill, or in you selectively starting to grind for WR


I disagree. My winrate is not blocked by my own level of play - when I consistently perform best of my team, I am not the one dragging my team down. You could argue that if I were better I could carry harder, but how are you even supposed to carry when an entire flank melts and you're 5 kills behind? I get that this happens every so often, but surely it cannot keep happening day after day, almost every single game?

I'll concede these are low numbers of games, but I named a number of ships I very much enjoyed and felt I was performing well in, the total number of games played are well on the way to 200, and yet the average winrate for all of these ships combined is well below 50%. This should not happen. Yes, with ships you hate playing due to not understanding them, or not knowing how to play them. But not with ships you do well in and enjoy.

Share this post


Link to post
Share on other sites
[TCNF]
Players
39 posts
5,528 battles

If you're average, I'd say 45% winrate is just as likely as 55% winrate. I would not recommend staring winrate too much, as you're just one player in a team of 12 ships. There's only so much you can do, but try to improve your own play.

 

One thing to add, is that you need preferably over 100 games, if not several hundred games in a ship to make statistical conclusions about the ship or your play. Just google "law of large numbers" for more info.

 

However, if you want to improve your winrate, play in a division of three players of the same caliber.

Share this post


Link to post
Share on other sites
Beta Tester
4,381 posts
9,616 battles
5 hours ago, Bratyn said:


I disagree. My winrate is not blocked by my own level of play - when I consistently perform best of my team, I am not the one dragging my team down. You could argue that if I were better I could carry harder, but how are you even supposed to carry when an entire flank melts and you're 5 kills behind? I get that this happens every so often, but surely it cannot keep happening day after day, almost every single game?

I'll concede these are low numbers of games, but I named a number of ships I very much enjoyed and felt I was performing well in, the total number of games played are well on the way to 200, and yet the average winrate for all of these ships combined is well below 50%. This should not happen. Yes, with ships you hate playing due to not understanding them, or not knowing how to play them. But not with ships you do well in and enjoy.

 

Erm. You do need to realise that just because you can't get to 100%, it's still limited by your own skill. The better you are the more you can carry, and just because there are games that pretty much no one can carry (making it nigh impossible to get to extremely high WRs) there's still a delta of games where better players than you can do what's needed to win, as are there (a lot more) worse players who consistently don't.

 

And yes, seeing varying WRs on ships with 200 games is still not something that "shouldn't happen". At 200 games you're still changing your WR by 0.5% per game, so being 20 games within your usual spread isn't special and that's +-10% WR.

 

5 hours ago, Breezewind said:

If you're average, I'd say 45% winrate is just as likely as 55% winrate. I would not recommend staring winrate too much, as you're just one player in a team of 12 ships. There's only so much you can do, but try to improve your own play.

 

One thing to add, is that you need preferably over 100 games, if not several hundred games in a ship to make statistical conclusions about the ship or your play. Just google "law of large numbers" for more info.

 

45-55% is a huge spread. And it truly takes someone rottenly bad to get as low as 45%. Most of the players in the entire game sits in that range.

 

Actually "average" players would sit within 49-51% or thereabouts.

Share this post


Link to post
Share on other sites
[PDXND]
Players
32 posts
6,134 battles

Agreed - I am probably too hung up on the number. But even though the number itself shouldn't matter, it is not fun to sink 3 hours into the game and be rewarded with a game log with only loss after loss. This does not encourage me to keep playing.

As for the law of high numbers, I am aware of this. If it was one ship I wouldn't be complaining - the problem is that it's been almost all of my recently-played ships, the combined number of games being closer to 200 on all of them.

 

5 hours ago, AgarwaenME said:

And yes, seeing varying WRs on ships with 200 games is still not something that "shouldn't happen". At 200 games you're still changing your WR by 0.5% per game, so being 20 games within your usual spread isn't special and that's +-10% WR.

Fair enough. You're probably right. :) It sure doesn't feel that way, though. I can't wait for Lady Luck to give me 200 games of nigh-consecutive wins to make up for this, in this case.

Share this post


Link to post
Share on other sites
[TCNF]
Players
39 posts
5,528 battles
3 minutes ago, Bratyn said:

Agreed - I am probably too hung up on the number. But even though the number itself shouldn't matter, it is not fun to sink 3 hours into the game and be rewarded with a game log with only loss after loss. This does not encourage me to keep playing.

As for the law of high numbers, I am aware of this. If it was one ship I wouldn't be complaining - the problem is that it's been almost all of my recently-played ships, the combined number of games being closer to 200 on all of them.

One question to ask is if the percentage of wins to losses is a good indicator in random battles for your personal skill progression? Maybe average experience received would be better indicator?

Share this post


Link to post
Share on other sites
[PDXND]
Players
32 posts
6,134 battles
2 minutes ago, Breezewind said:

One question to ask is if the percentage of wins to losses is a good indicator in random battles for your personal skill progression? Maybe average experience received would be better indicator?


Perhaps. And my average xp per battle has been steadily increasing every game -despite- all of my losses...

Share this post


Link to post
Share on other sites
[OV]
Players
64 posts
6,447 battles

I feel you brother - this is exactly as I feel.

Almost lost interest in WoWs because of bad matchmaking.

It is one thing to have 55% winrate, completely different when you end the game on top 5 positions all the time, and still have 55% winrate - this is really, really, REALLY frustrating.

I stopped playing for ~1 year (in the meantime my 9yo nephew used my account), and I guess I will take another 2 year brake very soon. Hopefully they will change the minds behind the game and fix it.

Share this post


Link to post
Share on other sites
Beta Tester
4,381 posts
9,616 battles
5 hours ago, Breezewind said:

One question to ask is if the percentage of wins to losses is a good indicator in random battles for your personal skill progression? Maybe average experience received would be better indicator?

 

Average xp is pretty much in direct correlation to whatever tier you prefer to play though. So it only works on at best a per tier basis and really just on a per ship basis. Also your average xp is also tied to your WR (ie, the more you win the more you get).

 

In my mind it's much alike to survival rates, average damage, hit rates etc. It tells a fair bit, and over a few amount of battles might be fairly relevant. But it's more an indicator than anything else.

 

5 hours ago, Ukio said:

I feel you brother - this is exactly as I feel.

Almost lost interest in WoWs because of bad matchmaking.

It is one thing to have 55% winrate, completely different when you end the game on top 5 positions all the time, and still have 55% winrate - this is really, really, REALLY frustrating.

I stopped playing for ~1 year (in the meantime my 9yo nephew used my account), and I guess I will take another 2 year brake very soon. Hopefully they will change the minds behind the game and fix it.


You might want to not just go on the forum to the game you're playing and say you're in breach of the games rules.

  • Cool 1
  • Funny 1

Share this post


Link to post
Share on other sites
Beta Tester
362 posts
2,920 battles

A tip: Your survivability is extremely low, try to improve this one. This should help you to carry more games.
Don't focus too hard on your win rate. I am at 67%, but just as dissatisfied as you.

  • Cool 1

Share this post


Link to post
Share on other sites
[IRQ]
Players
2,753 posts
6,638 battles
17 minutes ago, Breezewind said:

One thing to add, is that you need preferably over 100 games, if not several hundred games in a ship to make statistical conclusions about the ship or your play. Just google "law of large numbers" for more info.

You need about ten times that for the stats to be statistically significant. It depends a little on how accurate you want to be, and what margin for error you want.

 

19 minutes ago, Bratyn said:

I'll concede these are low numbers of games, but I named a number of ships I very much enjoyed and felt I was performing well in, the total number of games played are well on the way to 200, and yet the average winrate for all of these ships combined is well below 50%. This should not happen. Yes, with ships you hate playing due to not understanding them, or not knowing how to play them. But not with ships you do well in and enjoy.

That's still cherrypicking. I do "well", counting personal stats like damage and score placement, in ships I have crap winrate in, but considering the number of ships I've played, there are bound to be a few that have winrates that stick out from what I "should" have.

 

2 minutes ago, AgarwaenME said:

In my mind it's much alike to survival rates, average damage, hit rates etc. It tells a fair bit, and over a few amount of battles might be fairly relevant. But it's more an indicator than anything else.

Those don't need need as many games to be statistically significant as WR, if you count for single ships (since overall they're not all that indicative), since they hit their average a little bit easier. WR is a 1 or a 0, while something like damage is more often close to the average damage you deal per match.

  • Cool 1

Share this post


Link to post
Share on other sites
[PDXND]
Players
32 posts
6,134 battles
1 minute ago, LemonadeWarrior said:

Your survivability is extremely low, try to improve this one.


Agreed! I have a tendency to be too aggressive, and usually when I consciously stop myself I perform much better. Got any tips on how to stop yourself from impulsively going in to secure a kill?

Share this post


Link to post
Share on other sites
Beta Tester
4,381 posts
9,616 battles
Just now, AnotherDuck said:

 

Those don't need need as many games to be statistically significant as WR, if you count for single ships (since overall they're not all that indicative), since they hit their average a little bit easier. WR is a 1 or a 0, while something like damage is more often close to the average damage you deal per match.

 

Indeed, it was a point I almost made in that post. However for avg xp if you have had a loss streak you'll also by necessity end up with a bit lower avg xp. And average damage (and hit rates since it's such a telling stat for how people play) can be high without you playing well. So to an extent, ye I'd say they tell a bit more early, but not the entire story.

  • Cool 1

Share this post


Link to post
Share on other sites
Modder
3,989 posts
7,098 battles

Everytime I read "impossibly bad matchmaking" in this forum, I expect a person with no clue in maths in general and probability in particular complaining about things the player never understands (and most of the time not even want to understand) and on top mixing up the influence of the matchmaker and player decisions on the result and outcome of a match.

 

And up to now I got never disappointed.

  • Funny 1

Share this post


Link to post
Share on other sites
Beta Tester
362 posts
2,920 battles
3 minutes ago, Bratyn said:


Agreed! I have a tendency to be too aggressive, and usually when I consciously stop myself I perform much better. Got any tips on how to stop yourself from impulsively going in to secure a kill?


What drives you to go in and secure the kill? What makes you think it is wise to sacrifice your ship for another? You said it yourself; you think you are a better player, so it is not so smart to trade yourself for an average player, right?

For pushing there are rules that you should not push further than your base concealment and that you should trade damage wisely. Always have your exit ready.
Letting ships come into your base concealment is called overextending. These mistakes are easily recognizable on the minimap.

Share this post


Link to post
Share on other sites
Beta Tester
4,381 posts
9,616 battles
5 minutes ago, principat121 said:

Everytime I read "impossibly bad matchmaking" in this forum, I expect a person with no clue in maths in general and probability in particular complaining about things the player never understands (and most of the time not even want to understand) and on top mixing up the influence of the matchmaker and player decisions on the result and outcome of a match.

 

And up to now I got never disappointed.

 

To be fair, he did admit he was probably wrong. Which I give him credit for.

  • Cool 2

Share this post


Link to post
Share on other sites
Modder
3,989 posts
7,098 battles
2 minutes ago, AgarwaenME said:

To be fair, he did admit he was probably wrong. Which I give him credit for.

Ok, admitted.

Share this post


Link to post
Share on other sites
[PDXND]
Players
32 posts
6,134 battles
5 hours ago, principat121 said:

Everytime I read "impossibly bad matchmaking" in this forum, I expect a person with no clue in maths in general and probability in particular complaining about things the player never understands (and most of the time not even want to understand) and on top mixing up the influence of the matchmaker and player decisions on the result and outcome of a match.

 

And up to now I got never disappointed.


Do you mind keeping your condescending tone out of my thread? I went to great lengths to tone down my emotions when writing my post, at a time I wanted nothing more than to throw my glass through the room in frustration. Perhaps you could have the emotional maturity to do the same?

Furthermore, this entire thread is about -not- having any impact on the outcome of the match. How no matter how good I play or which decisions I take, I lose game after game. (Granted, this is probably to some degree observation bias)

 

5 hours ago, LemonadeWarrior said:


What drives you to go in and secure the kill? What makes you think it is wise to sacrifice your ship for another? You said it yourself; you think you are a better player, so it is not so smart to trade yourself for an average player, right?

For pushing there are rules that you should not push further than your base concealment and that you should trade damage wisely. Always have your exit ready.
Letting ships come into your base concealment is called overextending. These mistakes are easily recognizable on the minimap.


All in all, it probably is tunnel vision. A year ago I switched my minimap to full size, after having dabbled in middling size a lot before that, and it immediately had a great impact on my performance. However it still happens too often that I think I will be safe in chasing an enemy, and yet seemingly out of nowhere comes another ship that was just minutes earlier spotted halfway across the map. Map awareness is probably my single biggest issue, as most of the other gameplay aspects I have nailed down.

  • Cool 1

Share this post


Link to post
Share on other sites
[CATS]
Players
15,239 posts
11,004 battles
8 minutes ago, Bratyn said:


Agreed! I have a tendency to be too aggressive, and usually when I consciously stop myself I perform much better. Got any tips on how to stop yourself from impulsively going in to secure a kill?

Looks like this thread accomplishes something.

 

Looking more at the minimap helps in that regard.

26 minutes ago, Breezewind said:

One question to ask is if the percentage of wins to losses is a good indicator in random battles for your personal skill progression? Maybe average experience received would be better indicator?

Yes. Because winning should be your aim in a match.

No. XP is influenced by the Tier you play. The higher the Tier the more XP you get. Someone playing more high Tier proportionally would be considered a better player, even while playing less battles there absolutely. There is also the problem that WG counts the premium bonus for XP. Buy premium, be considered a better player? And another important factor is that actions that win the game do not get the appropriate amount of XP. It makes a huge difference if you do something at the start of battle or at the end. XP does not consider the time of an action. Some important actions are barely considered.

Share this post


Link to post
Share on other sites
[PDXND]
Players
32 posts
6,134 battles
9 minutes ago, ColonelPete said:

Looks like this thread accomplishes something.


I still disagree - this thread was not about how I myself can improve (as I have been aware of being too aggressive). It's about how, when I consistently play well and do manage to be careful, I still lose game after game for weeks. How my team folds like a house of cards no matter what I do. How I hold off an entire flank of 6 ships with only a single other battleship to support me, and yet the other flank, where we completely outnumber the enemy, dies to a man. How I get games with multiple caps, Kraken, and 150-200k damage, where I play the objective hard, and yet still lose. How this keeps happening day after day, with teams that are utterly impossible to carry because what are you supposed to do when it's just you left against 8 enemies. How you deal with that measure of bad luck and 'hopelessness'.

People mentioned my survivability - I argue part of that is due to two years of playing 'with distraction' and occasionally just YOLOing in because I wasn't aware of the game (at least in random battles). It has gotten a lot better now that I take random battles much more seriously, but it's hard to get a good survivability rate when it's landslide defeat after defeat. It's not that easy to stay alive when you're 1v6-8-10, and I refuse to cower in the corner of a map to pad my survivability rate.

Share this post


Link to post
Share on other sites
[CATS]
Players
15,239 posts
11,004 battles
2 minutes ago, Bratyn said:


I still disagree - this thread was not about how I myself can improve (as I have been aware of being too aggressive). It's about how, when I consistently play well and do manage to be careful, I still lose game after game for weeks. How my team folds like a house of cards no matter what I do. How I hold off an entire flank of 6 ships with only a single other battleship to support me, and yet the other flank, where we completely outnumber the enemy, dies to a man. How I get games with multiple caps, Kraken, and 150-200k damage, where I play the objective hard, and yet still lose. How this keeps happening day after day, with teams that are utterly impossible to carry because what are you supposed to do when it's just you left against 8 enemies.

People mentioned my survivability - I argue part of that is due to two years of playing 'with distraction' and occasionally just YOLOing in because I wasn't aware of the game (at least in random battles). It has gotten a lot better now that I take random battles much more seriously, but it's hard to get a good survivability rate when it's landslide defeat after defeat. It's not that easy to stay alive when you're 1v6-8-10, and I refuse to cower in the corner of a map to pad my survivability rate.

But it is about how you improve.

You claim to play well, but according to your words, that assumption is flawed. If you improve that, you get better results.

Share this post


Link to post
Share on other sites
[IRQ]
Players
2,753 posts
6,638 battles
26 minutes ago, AgarwaenME said:

Indeed, it was a point I almost made in that post. However for avg xp if you have had a loss streak you'll also by necessity end up with a bit lower avg xp. And average damage (and hit rates since it's such a telling stat for how people play) can be high without you playing well. So to an extent, ye I'd say they tell a bit more early, but not the entire story.

Yeah, single stats don't tell you much how well a player plays. WR is the most significant, but as mentioned, also takes the largest number of games to reach an accurate number.

 

XP is only relevant for single ships, and only after you adjust for WR. Even then, there's a lot of variation between how much XP matches give. Score placement is a better indicator, since that at least shows how well you do within the team. You know, as long as you discount how crap WG's XP rewarding system is in correlation to game-winning actions as opposed to damage farming.

 

Damage can be farmed after the game has been effectively decided, and you can do very well while still doing low damage. That's particularly noticeable with IJN DDs.

 

A low hit rate can indicate a useless sniper, but also someone who fires the moment the guns reload to focus on more chances rather than better chances.

 

Survival rate is a fairly good indicator of how aggressive you are, but aggression isn't something consistant in matches. Sometimes you need to play passively, sometimes that will doom you.

 

So yeah, there are a lot of things you can indicate and guesstimate, but you can never really be certain. In the end, there's really only one stat that matters: Are you having fun?

 

10 minutes ago, Bratyn said:

Map awareness is probably my single biggest issue, as most of the other gameplay aspects I have nailed down.

You could try getting a timer that beeps every 10 seconds, and look at the minimap (at least) every time it beeps. Or every time you shoot your guns (although for anything reloading slower than 6-7 s, probably a few seconds before you fire, in case of new targets).

 

You can also try to force yourself to never go into a situation unless you have an exit strategy. Before you engage any target, and every 30 or so seconds, take a moment to think about how to extract yourself if things go south.

Share this post


Link to post
Share on other sites
[TCNF]
Players
39 posts
5,528 battles
8 minutes ago, ColonelPete said:

Yes. Because winning should be your aim in a match.

No. XP is influenced by the Tier you play. The higher the Tier the more XP you get. Someone playing more high Tier proportionally would be considered a better player, even while playing less battles there absolutely. There is also the problem that WG counts the premium bonus for XP. Buy premium, be considered a better player? And another important factor is that actions that win the game do not get the appropriate amount of XP. It makes a huge difference if you do something at the start of battle or at the end. XP does not consider the time of an action. Some important actions are barely considered.

 

I think it's not that simple. Winning is of course what one should aim for, but winrate tends to be a bit problematic in showing your skill progression. For example, if you play mainly DDs, your skill has way more impact than if you play in a cruiser and you'll be seeing much higher winrates (not to mention pre-rework CVs). Average XP was more of a thought provoker and as you've demonstrated it is not a good measure of individual's skill progression, for the reasons stated above. There might not necessarily even be a good enough measure.

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

×