Forced Losing Streaks are Real

But wait… Didn’t you say:

Holding my own with someone in a widow lobby is not the same to holding my own in an actual match. Again, you’re putting words into my mouth, I only said that I was improving.

If you’re seriously still saying this, I’m almost 100% you don’t know how to read.

How would my communication be determined by how I’m arguing in a post? That doesn’t even make sense. Again, making assumptions without evidence, just like your entire stance on this topic. The fact that you can not admit that we’re all in the dark shows very clearly that you don’t know what you’re talking about.


I’ve added the data to Win/Loss Simulation and Data - Google Sheets. Your data is the fourth row. You will probably need to scroll down to see it.

Unfortunately, the original post is locked, so I can’t edit that. I may repost on the new forums at some point.

Everything is consistent with the model (in the second row). The only odd looking thing is that your 4 streaks are on the low side, and your 5 streaks are on the high side, giving a little hump in the streak frequency plot. However, these effects are within 1 sigma, and as such are not statistically significant.

1 Like

What is the R^2 of the sigmoid, btw?

Here is the fit parameters that matlab is returning for Des’s data:

fitresult = 

 General model:
 fitresult(x) = 1-.5*(1+erf((x-a)/(b*sqrt(2))))
 Coefficients (with 95% confidence bounds):
   a =        2186  (2179, 2194)
   b =       289.6  (243.7, 335.6)

gof = 

  struct with fields:

       sse: 1.2548
   rsquare: 0.4251
       dfe: 215
adjrsquare: 0.4224
      rmse: 0.0764

The R^2 isn’t great, but we’d need massively more data to get it down. Also, we’d need massively more data to distinguish what kind of sigmoid is appropriate (we could use a straight line fit with this amount of data and it would work about as well).

1 Like

By the way Kaawumba, you do realize that your model will fit almost everyone’s competitive experience right? If the system was designed to force losses and wins, your overall wins and losses would look the same to something that wasn’t forced. The thing we’re arguing about isn’t that we’re losing matches, it’s how we’re losing matches.

If I gave you all of my games, you’d probably tell me, “It looks totally fine to me. Loss streaks disproven.” But you’re only dealing with 1s and 0s. Not everything going on each and every game.

For example, if I played 10 games, winning five normal games and losing five normal games, it would look fine. If someone else played 10 games and won five really easy matches while also losing 5 matches that weren’t even close, it’s still going to look fine. You’re looking at the pattern that the wins and losses are forming and relating it to what a 50/50 win/loss pattern should look like. If you’re truthfully going to run this experiment, you need to factor in the game play of each and every individual in each and every game.

If you say that match quality can be poor, I’d agree with you.

However, if you are saying the Blizzard is intentionally making matches poor (in a way that can never be differentiated from random chance), that is where you need to take off your tin foil hat and go outside.

(7) Overwatch Forums
(17) Overwatch Forums
(36) Overwatch Forums

1 Like

Of course, if you think that everyone is lying to you about everything and the only evidence that you believe is that which you can see and understand with your own eyes and brain, this direct quote regarding how the MM finds fair matches for you probably won’t sway you.

For the rest of us, yes, bad people are bad. Report them. Use the reporting system to make the matchmaking system better.

A system that I’m describing wouldn’t be able to be differentiated from random chance (just like you’re saying), which is why I know your mathematical analysis can’t disprove it. The entire “facts > feelings” argument was a total circle for you guys. Again, everyone is in the dark here.

I don’t understand how you don’t even have a slight feeling that a rigging can be happening. You don’t feel any distrust towards the matchmaker – or even the entire company? In my opinion, putting too much trust in any group that controls a product/people is naive, but if that’s what you feel then I can’t change your mind.

Haven’t we been through this already? Again, these are the same people who won’t admit when they’re wrong, won’t take action against things that have been needing it, and attacking random popular players without any real cause.

Isn’t it wrong to report someone for “being bad”? I see no option for “Bad Player”.

I’ve looked at the results, and I find them rather interesting, and, for the most part, corellating with my conclusions about my gameplay.
Streaks in your model are series of games in one direction, which were not inerrupted by games in other direction?


20 characters…

This is so true. I’m upset.

If “Rigging” gives the exact same results as the system as they describe it, non-insane people choose to believe the system is as it’s described.

You don’t believe trolls are picking your SR off a slip of paper because that’s overly complex and relies on you believing things that aren’t supported by evidence.

It COULD be true. But it sounds absolutely ridiculous to you. It’s more complex than is necessary to explain what you see. It requires a bunch of people lying to you for no good reason. And it’s not the only explanation that fits the evidence.

It’s the classic “Brain in a Vat” problem in Philosophy.

If we were in a Matrix style “Brain in a Vat” ruled by an evil demon, the world would look exactly the same as it does now. You do not “know” that you even exist. You cannot “prove” that the world is not a simulation run by a demon.

You cannot prove that something does not exist.

What we do in ALL forms of inquiry is to understand the system is as simple as possible given the evidence provided.

The progression of science is the gathering of evidence that does not fit the current model. Look at the transition from Newtonian mechanics to Relativism. Newtonian mechanics describes almost everything that we see. However, once we started observing the cosmos in detail, it no longer explained everything and we had to come up with a new, more complex theory that covers the edge cases as well as the common ones.

If you want to contend that the current model, the system as described by the developers, is insufficient to explain the evidence, you have to provide evidence that counters that description. Evidence, not feelings.

Otherwise, we COULD believe all sorts of crazy things that also can be explained by the evidence. For instance, a unicorn pooping in a field marked with SR would also fit the random nature of the MM system. Completely generating SR based on the RNG function of computers may fit, etc. Honestly, there are things that would fit the data that are FAR SIMPLER than “Blizzard is putting bad players on your team on purpose to make a 50% win rate.”

COULD it be happening? Sure. You COULD be a sentient robot programmed to believe that you are a human.

But if you believed that you’d be considered crazy. Stupid. Insane. Out of touch with reality.

To us, believing that Blizzard “rigs” your matches may as well be you saying that the world is actually controlled by a demon that wants us to believe that we are free humans. You are in fact saying that the game is actually controlled by a demonic company that wants us to believe that we are free players.

Does the evidence fit? Sure. But you’re certifiably insane.


What is the meaning of gaps in my autocorellation function at -6 and 6?

Hmmm. So presumably when I drop a few hundred SR I should have super easy opposition and be able to carry myself back up?

Or in my case, as I dropped 650 SR I constantly faced superior opposition with constantly sub par team mates.

In a 2k SR game, that 3 stack that stomped us, and my plat friend said seemed a bit too good to be gold? They were all at 2.9k SR a few seasons back…

Enemy team had a Reaper that could constantly 1v3 my team’s dps. And none of my team on comms so that he was able to spring out and massacre my team from the same hiding spot 4 (!!!) separate times. I had to go as yet another dps as my team’s were useless (And apparently had no common sense whatsoever).

Why wasn’t he on my team? Why didn’t I get carried hard?

Reason, as far as I can tell, is my MMR was shot from constantly dying, trying to tank for dreadful teams that didn’t heal me and didn’t even use my shields. My fault really for over estimating the competence of team mates to do even common sense stuff. (Such as stand behind the barrier. Orisa even has a voice line for it for goodness sake).

I’ve only started climbing again for 2 reasons. (Gained about 300 SR with circa 80% win rate in 20 or so games).

  1. Am now duo queuing with a very good dps. Now I’ve got at least 1 team mate that I can rely on. And in return he can rely on me to support to a competent standard.

  2. I now focus play on highly survivable heroes, such as Moira and Lucio. On Hollywood I went 44-4 with Moira. We still lost, but my MMR was definitely positively impacted. I was then able to manage 14-0 and 12-0 in 2 spawn camp stomp games with Lucio and Moira, and now average circa 3-1 kd with them. My MMR has definitely risen.

I am now getting good team mates, mostly on comms, with one or two exceptions, and games at 2150 SR are far easier than what I was facing in the dank depths of 1700. A lot of the games seem almost like forced wins. The number of games where my team has spawn camped the enemy, whilst taking the payload to the end, or taken Hanamura with nearly 4 minutes on the clock, vs. enemy being full held, is significantly higher vs. supposedly “weaker” opposition at 1700.

I know what facing harder opposition is: I’ve played (And won) games at 2500 last season. Those games were, relative to my skill level, super fast paced and frantic and I had to play at my very best to win. The games at 1700 were not harder opposition per se. Just me being set up to fail as my MMR was shot from dying far too often trying to tank for terrible teams.

Selfish survival based play with 1 or 2 self reliant heroes is how to advance in this game.

https ://

One of the legitimate complaints is that the SR system is far to granular. There is no functional difference between 2500 and 2501 SR. What constitutes a difference is up for debate, but most don’t put it less than 200 SR.

So if you dropped 650 SR, that’s a bit outside of the smallest range anyone thinks is normal an well within the range some people think is normal. (+/- 325 SR)

I tend to notice a difference in play level at 200 SR, but those that I notice might be in a valley, so maybe I don’t actually notice a play level less than 400SR. That means that you won’t get “easier” opponents until you drop about 3-400 SR.

Even then, they won’t be 100% win rate easy. You’ll still have to work for it. But unless you’re rusty you should expect a higher win rate, which you did in fact see.


After I adapted my play style for worse team mates. I mained Orisa for the last few seasons and held 2300 rank for (I believe) 4 seasons fairly comfortably. Big difference was that at gold people used my barriers, healed me and used my halts and super chargers to get kills in team fights.

Silver is an utter sh*t show and is more like a glorified death match than a proper team game at times. The opposition was far worse, but so was my team. In my experience supporting teams in silver is far more difficult than supporting teams in gold as no one peels for you, nor stays in a decent position to make healing and rezzing easy (Or even survivable). Tanking is also harder as 80% of silver dps appear to be Leeroy Jenkins in disguise.

Only way to take advantage of worse opposition is to play a character that can get kills. Last time I was in the 1700 pit I climbed to 2300 in circa 10 hours maining Torb. After 30 hours stuck as Mercy supporting terrible dps.

And that, folks, is why so many dps instalocks…

Torb is sadly done in the current meta (Worthless vs. pirate ship, and clearly inferior to Junkrat’s mobility and massive damage output. You see Junkrat a heck of a lot more in OWL than Torb…)

I now point blank refuse to play Mercy anymore, and am sticking with my 2 mobile, survivable, somewhat dps based heroes.

When it comes to describing how or why things work the way they do, you are forced to put faith in their statements, as they are the ones that have access to the intimate details of what is going on. After statements are made, it becomes a matter of relativity among the reasonableness of each opinion. For example, it’s more reasonable to say the matchmaker is at least mostly impartial when trying to make fair matches, but that sometimes unlucky things happen than to suggest the system is rigged and streaks are being forced, because they’ve expressly stated the system is not rigged. Not knowing the exact mechanisms and calculations in place to create each match isn’t required to point out that the people that actually know how it works have said it isn’t rigged.

The moment you adopt a stance that you do not/cannot trust what the devs say, we lose the ability to have a meaningful conversation, because we don’t really have a choice but to operate under the assumption that what they tell us, while not necessarily all-inclusive, is at least truthful. Everything is speculation at the point you adopt the stance they’re (un)intentionally communicating misinformation. Arguing that rigged matches are happening becomes only marginally more reasonable than someone that claims they have an army of hamsters running on their wheels to spin a larger globe filled with what are basically bingo numbers, and that process governs matchmaking for quickplay.

It means that if you win game zero, you were more likely to lose game 0+6. Or that if you lose game zero, you are more likely to win game 0+6. The autocorrelation function is symmetric forward and backward in time because of the way it is calculated, but it only really makes sense to interpret it forward. Again, though, because of the error bars, it is not statistically significantly different from zero. It’s a little more than two sigma. If you took more data, or if I combined your data with porkypines, we’d expect the value to revert to zero.

I think it’s even more insane to trust a company. Did you trust Mark Zuckerburg when he said he would never steal and sell your personal information?

Then why do you come here and try to trash on what I’m saying? If you can’t prove that I’m not experiencing these things, why are you here at all?

You guys keep bringing this up, but what I’m stating in this post aren’t “feelings”. They’re actual, real, experiences. I can see the quality of game play switch very quickly between games. I can honestly say I’ve had, so far on Overwatch, around ten fair, close games. The rest of them is either stomp or be stomped, and it happens at a pattern that I can identify. I can almost see it before it happens.

So I guess Elon Musk is “certifiably insane” for believing we are in a simulation? And I guess all religious people are also certifiably insane? Not believing in the mainstream idea is not insane.

I am inclined to not put faith in their statements. Like I asked above, did you believe Mark Zuckerburg when he said he wasn’t collecting your personal information to sell?

See, I totally see where you’re coming from with this. At first, this is what I thought. After bad games would happen I would just say, “Hey, it’s just my luck”. But I could start identifying a pattern where I wouldn’t have to try in some games, yet sometimes have trouble getting my team in team chat. But then you say:

Well they’re obviously not going to say, “Well it’s rigged! You got us!” You can’t trust someone just because they said they didn’t do it.

The only people who know how it works are the devs, and again, they’re not going to admit it’s rigged.

I understand, but if you feel that you can not have a meaningful conversation here, then why bother? I totally get why you feel that way, and I would feel that way too, but if one side isn’t trusting sources from the same place that the other side is trusting from, it would be a waste of time for both of us.

Exactly, but your side is also speculation. Even though you trust them, you still don’t know if they’re telling the truth. I trust our president because he’s our president, but I don’t know 100% if he’s telling the truth. If he makes a statement, I’m still skeptical about it because I don’t have all the information, even though he said it. The same with Blizzard. When they say certain things, I feel inclined to still question it, not fully agree. I think it’s at this point that your side and my side disagrees, but I wish there was a way to still have a conversation that gets us somewhere. Maybe I’ll recreate my competitive ideas post and we can all actually contribute in the hopes of giving ideas to Blizzard to change competitive as so many people are asking, rather than arguing over what we believe to be correct.

  1. Did I trust Facebook was following their own public business model and selling information? Yes. Of course.
  2. Misinformation is dangerous. People take to conspiracy theories like yours because it’s easier to believe that someone or something else is the problem rather than themselves. When people believe, en masse, that they have no control over the win or loss of the match that’s a very bad state for the game.

Also, I think the better question is: “If you truly think Blizzard is lying to you an rigging the games, why are YOU here or even playing at all?”

  1. “I can almost see it before it happens” is the feeling. Your theory is testable. If you could watch someone else’s play and predict wins or losses before the match begins that would be something to see. Not your own matches though, because you can make it a self-fulfilling prophecy.

  2. Related to your asking for proof that THIS system is not rigged, if you don’t trust what people tell you (which religions do) or a priori logical proof (what Elon is doing and some religious persons do also), you’re not going to be able to show that a conspiracy theory is wrong. There’s always another conspiracy that can be layered on top to explain away evidence. You have only your own (biased) experience. Not logic, not authority.