Who is saying they know how the match making works? The vast majority of people I see describing details are the people saying it’s rigged.
1 Like
Its an interesting experiment. I have to write it down starting from 100 rocks 100 papers and 100 scissors all starting at zero points. All the way up to say 10 rounds. And then look if a sorting effect occurs.
My first short writeup didnt show any sorting effect in the first 2 rounds but i would not rule out it will occur due to random variations within the matchmaking. (in my experiment i follow the exact odds for the matchup distribution,which in reality will always show some variation).
I also take a range for the matchmaking. Anyone within 1 point of your score can match you randomly,to more accurately reflect ladder matching. I will not consider switching to counter decks based on past matchups because then writing it all down and calculating it becomes a rather large amount of work. For that a simulation would be better suited.
But the experiment described in the op,with the small adjustment of a small range for potential matchups,is not that much to calculate and write down.
If there is someone who can explain this logically to me,then i would be happy to hear it. But i will write it all down which i guess will take halve an hour (and which i wont post today anymore but tomorrow)
And if you want to wait for that to spot my mistake then that will be perfectly fine with me as well.
My intuition tells me a counter wall is not more likely then a wall of good matchups. And writing it down suggests that your matchups will never actually change. It will always be 33% rock 33% paper 33% scissors but maybe my intuition is wrong here.
You’re going to write out 1,000 games manually?
No just 10 rounds starting from 100 rocks 100 papers and 100 scissors starting at zero points. With matching randomly vs a range of +1 and -1 point.
It shouldnt be that much work and while it is not 1000 games,it will give an indication of a trend.
I never understood why people ran simulations in statistics when it can be calculated manually without to much effort. And writing it down helps me understand it better. Then i can better see and understand if and why the sorting effect occurs.
How many games is it? 500? It’s not very clear what you’re doing. 100 rocks/paper/scissors over 10 rounds.
The reason you simulate it is because I could write a program to do that in 5 min and it won’t make mistakes.
1 Like
I cant code unfortunatly (other then basic which no computer runs anymore) and i dont want to ask you either. I need to write it down for myself,then i will learn and get better understanding.
Scrottie already ran a simulation,though i have to admit that i dont understand what his charts do show.
They just show a random distribution after 1000 rounds but from those charts i cant see the sorting effect. Maybe someone else can but i cant.
It got 12 upvotes so there must be at least a few people who can see the sorting effect occur by just looking at those charts. I cant,but i havent really tried (and scrottie doesnt give any hints either) so maybe it is there i wont rule it out.
I am genuinly trying to understand it now. Not even for you guys,i now want to understand it for myself.
what i will write down is 10 rounds. With random matchmaking within +1 and -1 of your score.
Starting from 100 rocks 100 papers and 100 scissors at zero points. And then look what happens to the papers that win and keep following the papers that win.
To see if this group of papers that keeps winning will eventually run into a wall of scissors. And if they run into a wall of scissors,to see if they not first ran into a wall of rocks.
My first impression is they wont,but i havent wrote it down completely yet.
Either way,i will drop this for today and go play some games. And tomorrow i will post my findings.
And to the responses below:no need to worry about me. Noone is getting to me and i am genuinly thankfull i got to the point that i am making a serious effort (even though my intuition tells me it wont be worth it).
i am now genuinly curious about the sorting effect,i want to see and understand it for myself since i dont think it is there. Not because someone used it as an argument.
The coding is apreciated but i dont want to ask more,i will need to think myself.
What i would like to see from the program is the distrubution of matchups of the group of papers that keep winning at every score point along the way.
Then i can see the sorting effect occur and how it forms. And then i can see that the wall of counters comes before the wall of good matchups. (which is what op claims).
I cant see this from the 5 charts in the openings post as said before,but maybe i am blind.
And since several people have suggested i am blind for not understanding it (not in those exact words) i want to give the option that i am missing something obvious here a serious consideration.
Honestly man, if you are this invested it’s probably just time to take a break. You can run a personal simulation 2 million times and they won’t care or believe you.
I mean this as a friend who isn’t really but still… Don’t let this get to you like that. Mud dwellers live in the mud and will always best you there.
2 Likes
Not near a computer for a couple of weeks but maybe someone can do something with this untested example in python.
# player class
class Player:
def __init__(self, id, rating = 0):
self.id = id
self.rating = rating
def play(self):
choices = ['rock', 'paper', 'scissors']
return random.choice(choices)
def update_rating(self, outcome):
if outcome == 'win':
self.rating += 1
elif outcome == 'lose':
self.rating -= 1
# function to play a match between 2 players
def play_match(p1, p2):
p1_choice = p1.play()
p2_choice = p2.play()
print(f'{p1.id} chose {p1_choice} and {p2.id} chose {p2_choice}')
if p1_choice == p2_choice:
print("It's a tie!")
elif p1_choice == 'rock' and p2_choice == 'scissors':
print(f'{p1.id} wins!')
p1.update_rating('win')
p2.update_rating('lose')
elif p1_choice == 'paper' and p2_choice == 'rock':
print(f'{p1.id} wins!')
p1.update_rating('win')
p2.update_rating('lose')
elif p1_choice == 'scissors' and p2_choice == 'paper':
print(f'{p1.id} wins!')
p1.update_rating('win')
p2.update_rating('lose')
else:
print(f'{p2.id} wins!')
p2.update_rating('win')
p1.update_rating('lose')
print(f'{p1.id} new rating: {p1.rating}, {p2.id} new rating: {p2.rating}')
# list of players
players = [Player(f'player{i}') for i in range(100)]
# simulate matches
for i in range(10)
players.shuffle()
for p1 in players[:50]:
p2 = random.choice([player for player in players if player != p1 and p1.rating == player.rating])
players.remove(p2)
play_match(p1, p2)
This should output results you can analyse in excel. Output needs tweaking.
1 Like
Sadly, for Goramier this is about what (s)he considers personal victories. Personally I’m just trying to help you understand something you asked for help with.
2 Likes
I know you are trying to help. And it is genuinly apreciated!.
Off now to play some games.
Edit:and goramier i think also has the best intentions with his post.
1 Like
Best of luck man. Don’t get dirty with muck dwellers, they will always win.
Goramier is like an abusive partner
1 Like
That is very offensive. I find that to be truly a personal attack worth reporting. That is crossing the line.
I find muck and mud dwellers offensive too. And the attached statements.
Too late. Writing a ticket. That crossed the line. Nothing we debate is worth that. That crosses the line.
1 Like
Knock yourself out 
The players at the top never faced a counterwall,because they would not have been able to pass it.
They ended up on top not because they overcame a counterwall but because they never faced a significant counterwall. They ended up there because they faced less counters and more favorable matchups then all other players.
So where is the counterwall for the players that kept winning and ended up on top in this experiment?
The counterwall (over representation of counter decks) was there for the players that kept loosing.
What winrate did you use for favorable matchups in this experiment ,100%?
And can you show the matchup history for the 5 players that ended up on top of each run and the 5 players that ended up on the bottom
So that we can actually see where they met this counterwall and how big these counterwall(s) where in relation to the totall amount of games. And compare those with the counterwalls of players that ended up at the bottom?
Showing the matchup history of the winning players is obviously essential to judge your statement about counterwalls.
But you did not post the most important data of this experiment.
Is there any particular reason for leaving out this key information?.
So if you would be so kind scrottie,and make this experiment count.
Pls post the matchup history for the top and bottom players at the end of these runs.
I know you will ignore this request,because it would expose what actually is going on in this experiment.
And while you may not be rude,you have a strong tendency to go silent or start trolling when faced with critizism.
Goramier was right i am starting to think.
Edit:maybe the spreadsheet shows the matchup historys,i didnt look at the spreadsheet yet. I will have a look just to be sure. But i already know what i will see if it does show the matchup.
Goramier was right.
I wont respond to replies.
Time to uninstall this rigged game 4head.
1 Like
We have multiple millions of games of output.
So by way of analogy, let’s say there’s this machine and sausage comes out of it. Millions of people have eaten the sausage and parts of it has been directed under microscopes. We can’t see what’s in the machine itself, but from the output we know, without a doubt, that it’s a sausage making machine. If there’s something other than sausage in the sausage, there ain’t much of whatever that would be. It’s 99.9% sausage.
In the same way, we KNOW matchmaking is by rank/MMR. If it’s rigged at all, it’s less than 0.1% rigged, which is basically not rigged.
Ben Brode 11:07 how and why matchmaking is rigged. He speaks in general but u can assume he references HS.
2 Likes
Bee, that’s almost trolling of a post because you’re bringing something completely irrelevant from a whole different topic to this one. Stop stalking my replies on other topics.
Also stop implying I’m playing that deck to climb. What is wrong with you?
And “I don’t want to be mean” is “No offense, but”. Be quiet, Bee. Sorry that someone in your life scorned you somehow that you have to toxic-post on RNGstone forums, but for fel sake.