I wonder what the odds would be for someone to win 30 games in a row in the top500? Approximately the same as losing 30 games in a row in the bottom500. That’s not a guess that’s the math of symmetric pdf. Tail-ends should have the hardest time forcing a win (at the top) or forcing a loss (at the bot).
I’m on a 20 match loss streak practically for free.
Will submit the screenshots once I get a win!
Mind showing us the codes?
Literally your profile.
They don’t export for me right now (error since latest patch), but i’ll give screenshot of the losses.
Link please? I didn’t realize I was 99% of the people.
Nice lie, it’s because you don’t want us to see you walling off your teammates in spawn and hard throwing matches and/or 6 stacking with a throwing group on LFG.
ok I’ll screenshot just before the defeat screen. that’s a lot of screenshots but np np. i guess i have to start over. too bad, since i’m upto 25+ losses in a row rn. i’ll prove it via QP since comp queues will take too long and i have to start over.
also where is your evidence for the following claims:
how about we get you to post evidence of this absurd claim before i post a single screenshot? sounds like fairplay.
Check the profiles of those who liked this post.
They are either ashamed of it and private or the ones who are open are around these guys’ sr.
10min q and bored enough to go through some:
https://playoverwatch.com/en-us/career/pc/CrackmanDan-1967/
https://playoverwatch.com/en-us/career/pc/Homura-11867/
https://playoverwatch.com/en-us/career/pc/GradeÅ-1655/
https://playoverwatch.com/en-us/career/pc/Asuchri-1410/
https://playoverwatch.com/en-us/career/pc/PainCream-2742/
Who are these people and how are they 99%?
You said:
i said:
and u bolstered your claim further by saying:
Those profiles 1. aren’t claiming to be better than they are. 2. aren’t representative of 99% “venting” group.
So please, why are you posting with such nonsense statements unless you can back them up with evidence?
This part i will check since it seems to be your evidence. Because if you like a post, you are making a statement about your skill. And profiles of random accounts not only make up 99% of people who vent, but also show how they are “deluded”.
The people who liked that post did so because they need a conspiracy theory to cope with the fact that they are not good enough. 
any documentation for this statement or just more fail?
written documentation > ‘because squirrel said so’
Not surprised they don’t answer your questions Squirrel, I have seen them avoid replying to anything that proves their wrong…
He was around the 2000 SR area and played exactly as such. He even made a new account which was placed in plat, but even then everybody could easily tell he wasn’t a plat player, only got lucky because initial placement is near the gold/plat border.
I wonder what his’ SR is at now 
This is the most blatant evidence that you don’t know what MMR is that I’ve ever seen.
mmr is a many-to-one mapping
you do stuff that makes you win, you get mmr.
sr is a many-to-one mapping
you do stuff that makes you win, you get sr
mmr-free system doesn’t try to rig the lobby. it samples around sr band (i.e. player of similar “do stuff to win” mapping), and ships match. it’s not tampering with the contest odds.
You need more math sir.
And that’s evidence you don’t math.
I have yet to see you show any type of math, other than the failed attempt in the “Rigged Healing” clip =]
The best part with SR is there is no math except random sampling. You pull from similar bracket to create fair match using a many-to-one index. It’s like mmr but without the rigging. You can 100% match people around complex multi-factorial regression/classification of skill, and not rig every single lobby.
Please - explain where this bracket comes from with “no math”.
This seemed like a good thread until people starting to get all “personal attacky”.
The bracket itself comes from normal pdf, the SR is pointer to it. The mmr-free matchmaking is random around SR range.
You start with normal pdf assumption for player distribution by SR. Then you scale the axes to whatever you want the pixels to read-out instead of [0,1]. For OW that’s 0 to 5000 iirc. Surrogate is 500 to 4500. Then you seed everyone at the median=mean (for this type of dist), say 2300 in this example. Then you let them play out, the winners fight other winners and losers fight other losers until there is sufficient partitioning (with placebo matches and error-correction matches) to bin players into something like 15 regions of the normal pdf curve.
Instead of quartiles you’re moving them to e.g. 15-iles. And that’s where they start their laddering progression, using raw SR regions (say 15 pools of 333 sr each to cover the 5000 (it’s more like [500,4500] surrogate but w/e). If players win in a random lobby from that pool they go up SR, if they lose they go down. Eventually they cross into different pool and are sampled into lobbies from that pool.
With current skill differentiation I suspect only 7 or so pools would suffice. Most experts can barely tell the difference between “high silver” and “low gold”. SR is basically just a winrate indicator against some average backdrop of people with equivalent “do what it takes to have similar winrate indicator”.
So you see mmr-free system doesn’t need fancy analytics to work. No privacy invasion or performance trail. It splays players out the way any fair contest would. By their weight class only: their ability to convert wins against others who convert wins to a similar degree (vice versa for losses). It doesn’t tamper with lobby outcomes, it converges to fair matches as people naturally move to where their SR is an accurate representation.
In the end, if Alice and Bob are both 3000 rated players, or even both within say [2800,3100], they should both have the same non-discriminatory eligibility for the lobbies sampled from that pool. They would both be randomed into lobbies in that range and converge to winning or losing against an average backdrop of that range. Neither would be discriminated for/against into cherry-picked lobbies because of some hidden criteria that just tampers with the contest, where the odds are rigged for 50-50.
So…
Math?
Also, remember: people who understand things can explain them more succinctly!
No one in statistics calls a “normal distribution” a “normal PDF”. They call it things like “Gaussian curve”, or “bell curve”, or just “normal curve”.
Yes, because I said:
I didn’t say there was 0 math.
I generalize all statistics to measure theory, information geometry, sometimes combinatorial topology.
A normal distribution is just random sampling now?