Just delete this and get rid of my headache.
5 Likes
Thank you for this informative post. You are of course 100% correct but there are some who still deny these plain truths.
It took me around 40 hours of forced grind to get back to Diamond.
The best solution I have found to combat algorithmic handicapping is to group with one other person.
1 Like
I don’t get your point honestly. As long as the matchmaker is consistent (which it is), it doesn’t matter whether there is MM bias. You get better, you climb.
3 Likes
I mean… thanks… experiments with outcomes.
Honestly, it’s all a system that feeds into itself.
Content creators benefit cause they appear like they know something.
Coaches benefit cause they appear like they know something.
Forum folks that prop up the “get gud”.
And the only way to circumvent this? Is to get a team. It’s that simple and they know it which why we’ll never get a solo q.
1 Like
It isn’t, as the author of this thread discovered with his experiment:
Next experiment you have to tackle a couple things.
- High ping
- Endorsements
Do you mean, their effects on the those of players you get matched with? Or their effects on SR during placement?
You’re coming to this post with EXTREME bias but some of your results are interesting. It’s like having an argument using half-truths.
I’m tempted to just call you a “cuthbert cultist” and walk away, but I will engage your data while ignoring your clear bias. My anecdotal experience is just as valid as yours, right?
Experiment #1: Throwing.
I’ve thrown and dropped like a rock on a second account. It worked to get me to ~1000sr when even my 1800sr skill rating made it hard to drop any lower even if I was putting in the most MINIMAL effort possible. Your 2k rating was probably the same.
Experiment #2: Multiple Accounts
If you split your playtime then your main account is going to suffer as the game takes quite a long time to increase SR if you do get better. The game rewards consistency over longer periods of time some of which is out of your hands. Splitting that focus doesn’t help. Up to a 500sr difference is standard due to the way that matches are put together (Each game can have a 300sr+ split on a typical basis). While 800 is extreme it would also show as a difference in the starting account data.
Experiment #3: Using a Low-SR account
My alt was 1200sr when I started playing it after improving on my main. The account shot up to my main sr in almost no time and after the 5th win or so I was already playing against people that were around my current skill level.
If anything: What you’re seeing is that your MMR boosts far too quickly and SR is unable to catch up fast enough with your extreme increase. You don’t see the improvement in your SR as quickly with your MMR, which matches you regardless of your SR. Everything that cuthbert theorizes as “bias” in the matchmaker is just this and not some scheme to hold people back. More gradual increases in SR leads to more stable gameplay where large jumps result in more unstable experiences.
4 Likes
It’s true that I’m biased, but my results, as well as information on the patent for the matchmaker, as well as common sense make it very clear that the system is flawed in terms of real competitive fairness.
For my first experiment, it’s possible my lowest effort wasn’t that much worse than my max. I’ll grant that, but the minor difference in SR in my results show me that at least, in some cases, trying harder does not equate equal benefit. I am only one person and not a team of statisticians, so of course my results aren’t the makings of grand science. I’m aware of that. But I believe a lot of people can attest to the strange difference of SR between accounts. You can’t deny that, it’s a fact.
For the second experiment, I used both new accounts for the experiment. And the large amount of time required to show a difference in SR is the point I was making in my post. In a fair matchmaking system, players should be paired with true-equals in skill, and sr gains should be small and consistent, like the Glicko and Glicko2 rating systems used with chess. Instead, pairing individuals with partners and enemies with sporadically different skill levels without accounting for the individual’s actual input in the game skews the resulting adjustments toward a mean. This is why it takes so long to climb; it minimizes the player’s ability to realistically effect games of their own caliber because they don’t get games of their caliber.
And for my third experiment, MMR is irrelevant in this case. Of course by design, MMR changes faster than SR, but that’s the problem. SR should move just as fast as MMR; otherwise you get the problem of people like me still playing against people rated over 1000 sr below me after 30 games.
I’m not sure what you think my argument is. I’ll say it plainly here:
1: Matchmaking pairs you with players of highly different real-skill levels while giving them nearly the same visible SR, making SR a (somewhat) ineffective tool at determining skill-level.
2: Because SR is ineffective as a tool for determining skill-level, and because it takes so long to increase despite true improvements in skill, many players feel frustrated when the matchmaker, which creates games to have a 50/50 percent win-chance on each side, puts them in games they have less impact on because they’re being paired with people of different skill level’s than them.
So I’m not sure what you’re disagreeing with here. Are you saying that SR is a better indicator of skill than MMR? Or are you saying the matchmaker does not create matches to have a 50/50 win rate? Because we’ve known both to be true for a long time.
2 Likes
While it is true that most of us including you have our own biases it may be argued that an underhanded personal attack does little to convey a sense of neutrality from your side.
2 Likes
I have been doing same for years and its nonsense. Its not rigged.
4 Likes
I want to know what I’m wrong about here though.
I personally know people, including myself as well as friends, who are in different ranks on different accounts, and hard stuck between them. What about this are you refuting? That they forget how to aim when they switch accounts? Or maybe they forget positioning skills, or hero picks? If you’re highly ranked, sure, you tend to play very consistently and therefor rank similarly between accounts. But for people between low gold to low diamond, consistent play isn’t realistic. I don’t know what you’re refuting here.
That’s a lot of words just to say that you want to be carried and not have to try in your games.
Experience isn’t facts
That just means that you weren’t Diamond skill level before that.
5 Likes
Pretty much whole thing.
You talk about experiments yet you dont present any real data. I did that multiple times in past. Then you talk about some experience from your friends, also no real data shown. All we have are your stories and biased experience because you were already conviced of rigged matchmaker before you done these experiments.
Its also funny you didnt post from your main acc and are hiding behind alt.
3 Likes
That wasn’t my point here. Please thoroughly read my claims before you strawman me.
And if he wasn’t Diamond skill level before then, then why was he placed in diamond? You can’t vouch for the efficacy of the system sometimes and refute it at others when it’s convenient. If the results were inconsistent, that’s the point we’re making.
1 Like
Just posting something so you can react again…
1 Like
His peak was Diamond, you are average skill level isn’t your peak. Just because I peaked Top 194, doesn’t mean I play at that level reguarly. My skill level is High Masters/Low GM.
Also a lot of people have reviewed Hulk’s VODs and they also estimate him to be around Low to Mid Plat, even a Coach (Temporal) agreed on that.
2 Likes
Why would I post my main account? I play with my friends on there. I’m not posting that on here. And no, I don’t have the data. Fine, call me wrong. But I’ve been asking specific questions here and haven’t gotten anything besides “You’re wrong” with no evidence directly in opposition to my claims. Even without my experiments, it’s a logical conclusion for simple observation.
So please tell me what I’m wrong about, and explain to me how I’m wrong. Point it out directly so I know what you’re disagreeing with.
1 Like
What is that supposed to mean?
1 Like
Well he obviously dont want to post from main since he admitted to throwed 40 games. So i guess its valid reason. But still not cool at all to throw games with people who want to win.
3 Likes