Overwatch - SR System is BROKEN and Your Rank is Not Accurate
another youtube vid
2 Likes
At the upper tiers mechanical skills are not far apart, but at the silver region? They are VERY far apart. Diamond players can absolutely wreck Silver players purely because they wonât miss every other shot. I still remember during the off-season seeing one or two bronze/silver players paired up with high ranked ones in a diamond game; they were blatant weak-links on their team, and even as easy-mode characters like Moira their mechanics were horrendous and easily exploited.
The difference between Masters and Diamond is big enough, let alone Diamond to silver.
This is true, but the topic discussed at hand is the skill disparity within the lower brackets itself, which I fully believe can be wide. Many plat players have diamond+ mechanics, itâs just their decision making and awareness keeping themselves at their rank.
2 Likes
There isnât much a gaming company can do about that though. It happens in the lower ranks.
Plenty they can do, as they use a forced bell curve distribution to ensure what % of players are in Silver etc.
They could force that spread to be wider. Have more players in plat and above.
2 Likes
How is it forced? Please explain in detail.
Unfortunqtely explaining it would take more words than I am interested in typing. Simply put, the two times they released SR distribution to the public the % of players in each division was the same. As new players can place anywhere on the scale this implies a forced distribution.
2 Likes
You donât know what a bell curve is, do you?
Itâs a natural, mathematical distribution of events. It is impossible to force a bell curve.
If you take 100 humans and see which is the âbestâ and âworstâ at anything, say, skee ball, or writing legal briefs, or changing their underwear, you will get (roughly) 1/3 of the population in the middle, and 1/3 on either side, with those numbers tapering down to +1, +2, +3 standard deviations.
A bell curve CANNOT be forced. You can make there be no bronze or silver ranks, but there is a still a bell curve where the lowest part of gold (in this instance) would be the bottom 10%, mid gold would be the bottom 20-30%, plat would be the average, and so forth and so on.
Youâve changed nothing about the bell curveâyouâve merely labeled it differentlyâthereâs still the exact same distribution, but itâs a lot less clear to see.
TL;DR a bell curve cannot be forcedâit is a mathematical expression used to gauge any given population against an average.
1 Like
Yet how you distribute SR across player base is totally in Bliz control. And look at the distribution given by Jeff.
If you build a graph of this distribution, you will see, that right tail is very long and very low, meaning: the distribution center is shifted. No doubt, artificially.
Silver and Gold should have subdivisions. Skill difference is too far apart on the ends of these ranges.
2 Likes
HALLELUJAH hallelujah praise this post ![:raised_hands:t3:](https://d74zh8edaijub.cloudfront.net/en/overwatch/images/emoji/twitter/raised_hands/3.png?v=5" ":raised_hands:t3:")
2 Likes
There are 4502 skill subdivisions. Below 500 plus 500-5000.
There are 7 labels or groups, but those groups are completely arbitrary. They could make me a GM tomorrow, technically, by lowering the threshold of that label, but it wouldnât make my skill any better.
MOST people realize this, but Iâm still an advocate for getting rid of the overly specific SR numbers and just using âhigh goldâ, âmid silverâ, etc.
But, youâd still have a bell curve. You could center it at any rank, in theory, but why would you want a low bronze population but a high GM population? People that work hard for GM should get that recognition of being in the top 3% or whatever. No one should care of you are in the bottom 3 vs. the bottom 10%.
Yes, getting rid of numbers is ok, too.
And Iâm not saying, that we live without the bell curve. The distribution form is not an issue. Issue is - placing new accounts into gold, by making so completely diminishing the value of âaverageâ of the distribution. Issue is - having 50% of player base in two ranks lower than middle of the distribution. This, and the entire distribution shift, is clearly done on purpose, to separate extremely good players from anyone else, and not giving a slightest pinch of care about that majority of player base.
With manipulations like these you could claim (and in fact, should claim), that there is power distribution in place, and not bell. It is a well researched phenomenon - people performance forms not bell, but power law curve.
2 Likes
This 1,000,000 times. I pray the devs fix comp. I love this game and would really enjoy a proper competitive ladder.
2 Likes
They absolutely could do a gamewide reset for seasons twice a year. Thereâs really no reason for the fact that people can buy new accounts, place and maintain higher and absolutely struggle to climb on their main. The reason for this is matchmaker trying to group based on skill. The main problem with this is from accounts that have been around for 4+ seasons that have tanked in SR. You arenât playing with brand new bronze players at the same SR for the most part. Youâre playing against other people who have tanked for one reason or another.
What then happens is many people youâre grouped with are so irritated with losing so much SR, they will intentionally throw or just leave. Why? They too have another account and instead of trying to dig this one out of the grave, theyâd rather ruin the game for others.
This fallacy that everything evens out as far as leavers on both sides is one huge disservice to the state of the game. Yesterday I lost 8 games with a leaver on my team. In 3 of those games, multiple people left. Know the amount of leavers on the other team for the day? 3. I ended up climbing in SR from having more wins overall, but this garbage is absolutely destroying lower ranks. Thatâs 6-8 losses (3-6 due to the wins from the leavers on the enemy team) in one day that I have to now win due directly to leavers on top of what I already need to actually climb.
Itâs the exact same thing with people intentionally trolling, though thatâs harder to spot on the enemy team depending on what theyâre doing. Thatâs just one day. Every single day Iâve played this season, I have had more leavers on my teams than the enemy teams Iâve faced.
Acting like Blizzard canât do anything about this is foolish. People have given suggestions. Some would work, some wouldnât. People outright denying that thereâs an issue with comp period, though⌠thatâs just burying oneâs head in the sand.
2 Likes
By using MMR to achieve a theoretical 50/50 outcome it is pushing everyone to 50% saying win percentage has nothing to do with an algorithm designed to achieve a 50/50 outcome is beyond disingenuousâŚItâs like a casino saying the games arenât fixed against a single playerâŚwell technically no the entire casino is riggedâŚwhy not give players the option on how they would like to be matched? MMR adjusted ladder and a strictly SR balance ladder?
2 Likes
A bell curve canât be forced? Who told you this? Iâm not saying this is how OW matchmaking works, Iâm responding to your outspoken statement that a bell curve canât be forced, seriously who told you that? I could actually draw a bell curve right now and take everyone who has posted in this thread and randomly distribute them across the bell curve and say that itâs my ranking system of how well people can construct posts.
Are you meaning to say something else?
1 Like
Nature told me that.
A bell curve is a natural by-product of probability. To see it in action, roll 2 dice 100 or so times, record your data (how many 2s, how many 6s, how many 10s etc.) and you will see a bell curve. You may have some noise in it after 100 rolls, but do it a million times and it will be accurate to 100s of percentage points.
Overwatch ranking is very similar. There is a median skill (variance). Itâs pretty easy to figure out what the median skill is for several reasons, but the easiest to understand is that most players will gravitate toward it since, by definition, average is exactly average. Just as two dice rolled 10000s of times will tend toward 7.
You can stretch the curve however you like. You can flatten it (everyone trending toward 2500). You can inflate one end (only 1% of people in bronze and 20% in GM). You can spike the middle (90% of people in gold with very 10% in all the other ranks).
But it is still the same curve with the exact same standard deviations. The only thing that changes is the definition of your rank. In the last example, for instance, the difference between a 2000 and a 2500 would insanely large since 90% of the population would be thereânot a particularly good system.
Bell curves are naturally occurring phenomena, and we used maths to figure out how to express them.
As to you ârankingâ posts and putting them in a bell curve (gonna assume Iâll get a zero from you), thatâs just opinion. But IF you were somehow able to rate every single post on an empirical and impartial scale taking into account grammar, logic, humor, toxicity, follow-through, etc etc etc etc, then you would see that not only would all posts âgradeâ follow a bell curve, it is impossible for them to do anything else.
Even if there more FAR more s-posts than âexcellentâ posts, the s-posts would STILL be better or worse than one another.
2 Likes
lol so you donât know. If you are going to claim the system is rigged you need to provide proof that it is.
what?!?! how does the distribution among ranks and new account placements prove a forced bell curve?
First: the % of players in each division are absolutely not the same⌠not even close
Previous released player distribution:
https://i.imgur.com/LoldcuU.png
Second: it is more a log-normal distribution than normal distribution(see links below) and itâs just a display % in X rank. There will always be a peak in any of the distribution types
Log-Normal Distribution:
https://en.wikipedia.org/wiki/Log-normal_distribution
Normal Distribution:
https://en.wikipedia.org/wiki/Normal_distribution
Third: new players do not place anywhere, they place where their MMR(from QP) and placement matches put them. This is just too ridiculousâŚ
I meanâŚI assume you realize that âaverageâ by definition means that 50% is lower?
How can it possibly be any kind of problem that the middle is in the middle? Plus, the average is actually mid-gold, so you have 2.5 of 7 ranks below average and 4.5 above average. Which, let the GMs have their rank. Itâs ok to have a top rank be a small percent.
Now, IF we grouped by rank rather than MMR I could see this as a problem, but we donât do that. If youâre 1999, the chances are high that your team and opponents will be between very low gold and very high silver. Theyâre not 1500-1999. Except for groups, of course, which screws up the matchmaking.