33 replies to this topic

[Mechovy muzicek]

Right, so I think I was on the losing side in 18-19 out of last 20 games and the game is losing the fun aspect quite quickly. It's not like the games are lost because of me (I still climb towards a higher tier, often having near-top match score). But the fact it's a loss after loss, day after day, seems like it's beyond random.

Has anyone else had such experience?

[martian]

Mechovy muzicek, on 26 July 2020 - 12:42 AM, said:

Right, so I think I was on the losing side in 18-19 out of last 20 games and the game is losing the fun aspect quite quickly. It's not like the games are lost because of me (I still climb towards a higher tier, often having near-top match score). But the fact it's a loss after loss, day after day, seems like it's beyond random.

Well, it's hard to say unless you keep exact records regarding your games.

Mechovy muzicek, on 26 July 2020 - 12:42 AM, said:

Has anyone else had such experience?

Not exactly.


Nemáš chodit Russu Bullockovi za ženou ...
Jo, a nazdárek.

[Tarl Cabot]

MM has nothing to do with it. The only real constant is you. With that said, considering the time of this post the population is likely low compared to NA primetime but you are still doing well, MS-wise. For the month of July ya at 29 W and 49 L with an avg 298 MS. And as you noted, even on losses you seeing more up arrows then down arrows. Without knowing what mechs/loadouts being used, how you are running your mechs could potentially be part of the issue, such as running LRM/ATM boats, thus spreading damage instead of concentrating the damage. Via your stats you are winning more games in heavies than mediums, while running the same avg MS.

[Nesutizale]

I have to update my record with the latest games but so far my W/L is around 1 (1.1 I think it was). Meaning that I loose and win about the same amount of matches.

The longest loosing streak I had was 4 matches and the longest winning streak was 5. So also pretty balanced.
What makes you feel like loosing, even while winning, can be how fast you died during a match.
I mean when you die pretty early on and then just have to wait till the match is done because you want to use that mech again then it can feel like you lost.

In the end how it feels and what actualy happens can be different things.

[Mechovy muzicek]

martian: Sakra, kdo to prokecnul?

Tarl Cabot: Thanks for the reply! Yeah, I think the month was fairly balanced (with more wins initially), but then this long streak of near-pure losses came and left me wondering what's wrong... The point you raise on mediums/heavies is an intriguing one - because the reason I switched from Night Gyr to Hunchback was that I felt I wasn't doing well enough, not making enough of an impact in matches (many would have been won even if I wasn't there to be honest). But the difference seems to come mainly from difference in teams - I had more balanced matches in the heavy mech than in the medium, where quite many matches were horrendously unbalanced skillwise.

ATM use can be a part of the problem and I'm aware of that skewing the scores a bit, but I think I can tell when I'm doing well/badly, and when it's the team, and in my recent streak of losses, I don't think I screwed up that often (of course I did in some games too)... I get that non-prime time can cause unbalanced matches in general, but am not sure why it was pretty much always against me :/

Nesutizale: Agreed, there are all sorts of observation biases coming into play - pity I cannot get the exact stats!

[KodiakGW]

Matchmaker has always hated me. Seems like it always will. Only time it didn’t was for a short time when I convinced my old War!Online teammates to finally download the game. A few months later, they dropped the Skill Shubbery and engine desync, which made them leave.

If you think you are being shafted by the matchmaker, take a screen shot and compare names to Jarl’s.

[Sjorpha]

Long streaks are normal in large samples of random numbers. For example if you flip a coin a thousand times you should not be surprised to find a streak of 20 heads in there. Yet if you were to flip it 20 times and get all heads you'd feel like something really weird was happening. But the fact is that someone is out there flipping a coin 20 times, getting heads "every time" and perhaps falling into the trap of thinking there is some spooky reason for this, when in reality they are just the inevitable random dude that happened to get 20 heads in a row.

Point is, a losing streak of 18 matches doesn't have to indicate a damn thing about anything. It feels super weird but it's actually a fairly normal thing to happen even if the odds were close to 50/50 in each of those matches.

Edited by Sjorpha, 26 July 2020 - 05:14 AM.

[Mechovy muzicek]

Sjorpha, on 26 July 2020 - 05:13 AM, said:

Long streaks are normal in large samples of random numbers. For example if you flip a coin a thousand times you should not be surprised to find a streak of 20 heads in there. Yet if you were to flip it 20 times and get all heads you'd feel like something really weird was happening. But the fact is that someone is out there flipping a coin 20 times, getting heads "every time" and perhaps falling into the trap of thinking there is some spooky reason for this, when in reality they are just the inevitable random dude that happened to get 20 heads in a row.

Point is, a losing streak of 18 matches doesn't have to indicate a damn thing about anything. It feels super weird but it's actually a fairly normal thing to happen even if the odds were close to 50/50 in each of those matches.

My first hunch was "I know, I'm a statistician", but given how weak one's intuition can be, I simulated this now out of curiosity - generated 100k instances of random binary sequences of length corresponding to my game count (1550), checking for maximum length of loss streaks. Looks like the proportion of instances with a streak of at least 18 losses is around 0.3% (mean value is 10-loss streak) - so I wouldn't say it's really normal for an individual, but it's true that with the player base, it is extremely likely to happen to someone.

I think what made me think something is strange is the fact that on an individual level (kills, match score), I did reasonably well (at least given my not-bad-not-great skill), often scoring highly - given the state of the game at the moment, I dare say I'm quite above-average in performance, so the chance of loss should be less than 0.5, which would make the likelihood of long streaks markedly lower...

A caveat in the other direction is that my counting was vague - I started when it felt like 10 losses in a row, so I cannot rule out it was actually 15 or 20.

(btw. since writing the post, I had 5 losses and 1 win (where I didn't have to be there at all, because the enemy team was weak) - still feel grumpy)

Sjorpha, on 26 July 2020 - 05:13 AM, said:

when in reality they are just the inevitable random dude that happened to get 20 heads in a row.

That must have been intense!

Edited by Mechovy muzicek, 26 July 2020 - 06:58 AM.

[Nightbird]

MS you earn has little to do with how much value you have to a team. For example, you can easily hide for half the match to clean up damaged mechs later and earn lots of MS but this helps the steamroll against your team and leading to a loss more often than not. There are other ways to earn MS that is just as bad for your team.

[Hrix]

Most recent drop, 5 assaults to NONE. Not a single one on our side, We lasted 2 minutes. War Thunder it is. Email me when you fix things, mmk?

[Mechovy muzicek]

Nightbird, on 26 July 2020 - 07:06 AM, said:

MS you earn has little to do with how much value you have to a team. For example, you can easily hide for half the match to clean up damaged mechs later and earn lots of MS but this helps the steamroll against your team and leading to a loss more often than not. There are other ways to earn MS that is just as bad for your team.

That is of course true, but I'm no back-sitting potato - I stick with the team, support pushes, and am engaging at normal distance (like 200-400 m)...

I agree that MS is not equivalent to how much value you have to a team, but in most cases, I think there is a reasonable correspondence...

If you hide for a half of the match and your team is wiped out, you will get wiped out soon after, and with minimal damage for good measure (you'd have to just shoot "safe" weapons like LRMs, which some people try to do, but not my case).

[Sjorpha]

Mechovy muzicek, on 26 July 2020 - 06:50 AM, said:

My first hunch was "I know, I'm a statistician", but given how weak one's intuition can be, I simulated this now out of curiosity - generated 100k instances of random binary sequences of length corresponding to my game count (1550), checking for maximum length of loss streaks. Looks like the proportion of instances with a streak of at least 18 losses is around 0.3% (mean value is 10-loss streak) - so I wouldn't say it's really normal for an individual, but it's true that with the player base, it is extremely likely to happen to someone.

Well, I didn't mean it's likely to happen to everyone. Just that it can and does happen from time to time, so it can't straightforwardly be taken as evidence of something being wrong.

You do have a positive win/loss ratio historically, especially in June, so evidently you're on average contributing to wins. I bet if you play 100-250 matches like you did in June your w/l for July is going to be positive even despite this streak of losses you're describing.

That's just speculation of course, just saying that the historical stats indicate that you've been getting too weak opponents rather than too strong. Pretty much no one should have a 1.43 wlr over 243 matches in a functional matchmaker, so that was matchmaking failure in your favour during June.

Edited by Sjorpha, 26 July 2020 - 07:34 AM.

[martian]

Hrix, on 26 July 2020 - 07:13 AM, said:

Most recent drop, 5 assaults to NONE. Not a single one on our side, We lasted 2 minutes. War Thunder it is. Email me when you fix things, mmk?

Russ Bullock and Paul Inouye are absolutely okay with this fact.

[Mechovy muzicek]

Sjorpha, on 26 July 2020 - 07:28 AM, said:

Well, I didn't mean it's likely to happen to everyone. Just that it can and does happen.

You do have a positive win/loss ratio historically, so evidently you're on average contributing to wins. I bet if you play 100-250 matches like you did in June your w/l for July is going to be positive even despite this streak of losses you're describing.

Totally agreed - if it's around 0.3% per person, it's very likely it will happen to several people in the player base.

Ha, yeah, I agree it should go back, it's just the game is not a lot of fun when one loses all the time :/ (I don't mind losing 6+ vs 12, when it's because the enemies do well, but when it's that you say "take top" on HPGM, and 3 mechs do that, 4 sit back hiding, and 5 rotate, and it ends 1:12, it's no fun at all).

Btw. how can you find stats over time? I don't use the website extensively, and have found no way of separating the results by months or something. Thanks!

Edited by Mechovy muzicek, 26 July 2020 - 07:42 AM.

[East Indy]

Nightbird, on 26 July 2020 - 07:06 AM, said:

MS you earn has little to do with how much value you have to a team. For example, you can easily hide for half the match to clean up damaged mechs later and earn lots of MS but this helps the steamroll against your team and leading to a loss more often than not. There are other ways to earn MS that is just as bad for your team.

Players can earn more than zero but not enough to mimic top-tier performances, based on a quick scan of known culprits. Better than average at best. Whatever lucky draws they get, of half a dozen crippled and open 'Mechs, are outnumbered by games where they get toasted by 8 players at once.

Ultimately, players with the very highest matchscores — especially those who play solo or in small groups — are in fact constantly engaged, and simply out-shooting, out-rolling, outmaneuvering and overwhelming opponents.

Hrix, on 26 July 2020 - 07:13 AM, said:

Most recent drop, 5 assaults to NONE. Not a single one on our side, We lasted 2 minutes. War Thunder it is. Email me when you fix things, mmk?

"One weight class per player" would really simplify things for the matchmaker. It'd also give players in groups a little less pressure to up-ton.

[Willard Phule]

Mechovy muzicek, on 26 July 2020 - 12:42 AM, said:

Right, so I think I was on the losing side in 18-19 out of last 20 games and the game is losing the fun aspect quite quickly. It's not like the games are lost because of me (I still climb towards a higher tier, often having near-top match score). But the fact it's a loss after loss, day after day, seems like it's beyond random.

Has anyone else had such experience?

It's not a coincidence, it's actually the way the whole thing is designed. Honestly, the solution to getting stomped in QP is the same as it was in FP. Join a group. That's how they're doing the stomping. 4 guys that are coordinating their efforts together are always better than 1 guy doing his own thing.

Why do you think they added groups to QP in the first place? It's not like they didn't know what the result would be, groups have been in QP before. Same thing happened then as is happening now, except now we're all jockeying to get out of T3.

[Too Much Love]

Mechovy muzicek, on 26 July 2020 - 12:42 AM, said:

Right, so I think I was on the losing side in 18-19 out of last 20 games and the game is losing the fun aspect quite quickly. It's not like the games are lost because of me (I still climb towards a higher tier, often having near-top match score). But the fact it's a loss after loss, day after day, seems like it's beyond random.

Has anyone else had such experience?

Yes, it works that way. I had exactly the same experience across many years in MWO.

It's always like that: 10 or so matches with the team that literally has no chances, then several matches on the strong side. So called winning and losing streaks.

Why it happens? Big question. I don't know, and I doubt that even PGI knows.

[Mechovy muzicek]

drunkblackstar, on 26 July 2020 - 07:46 AM, said:

Yes, it works that way. I had exactly the same experience across many years in MWO.

It's always like that: 10 or so matches with the team that literally has no chances, then several matches on the strong side. So called winning and losing streaks.

Why it happens? Big question. I don't know, and I doubt that even PGI knows.

As discussed with Sjorpha - streaks are quite a common pattern, but too much can be too much
Yeah, the problem is that losing when your team has no chance sucks - and while being on the other side is somewhat more fun, it's not great either.

Not sure what are all the factors that are considered in a matchmaker, and I don't pretend it's an easy problem in general - but I'm finding that as much as I do enjoy MWO when it works, the ratio of cool games versus the near-ragequit ones has become much worse for me recently.

[Nesutizale]

Mechovy muzicek, on 26 July 2020 - 04:17 AM, said:

Nesutizale: Agreed, there are all sorts of observation biases coming into play - pity I cannot get the exact stats!

You can either look at Jarls Lust -> https://leaderboard.isengrim.org/stats
Just type in your pilots name in search. Also it sometimes isn't up to date. I think it updates every month?

Then you could also go into your profile on this site. Click Leaderboard and at the lower end is your name and your rank. Click on the number of your rank and you can find your current seasons Stats.
While it dosn't list every match at least you can see how many wins and losses you had so far as well as some other stats.
Should give you a much better impression.

Last thing you could do is make a screenshoot at each match end and make yourself an excel table and note your wins, losses and whatever else you want to keep track of.

[Snowhawk]

Oh yes… I knwo this.

Here is a copy from my old text (old matchmaker solo Queue):

Long time ago I had a loosing streak. Some Players here in the Forum have the opinion that when you have a loosing streak over several games then it's your own fault because you are the only "constant" in this Matches.
Well, the only constant in this lost Matches was that my Performance was good and I often managed to get top scores of my Team. Then I decided to compare my Teams with the data from Jarl's list and I was shocked. The Winning Team had often Players from the 90% percentile while my Team the best Players were from the 80% percentile area. Nice... isn't it?

Also, there was often a clear cut in this Teams. My Team had no one from the 90% percentile area. My Question is… how can it be that there is a such a clear difference between the two Teams? Even the dumbest Matchmaker would give you by chance at least 1 top Player in your Team. But... not the matchmaker from pgi…..

There is something wrong with the matchmaker and many Players have reported this. Now I understand why Bowser and Proton (top-Players) are 3 times in the same Team (quickplay old soloqueue with no groups, and they are not syncdropping)…. it's the matchmaker who opens a game for Tier 1 Players and then fills the second Team with everything he can find, even with Tier 3 Players or below.

Edited by Snowhawk, 26 July 2020 - 11:03 AM.