and they also lower rewards for playing against them
its not that much different. 
no it’s the perfect amount of difference. it’s not that much, it’s not that little, it’s perfect.
if a team is expected to win against another team two-thirds of the time, when the underdog wins, they’ll gain 150% SR had the game been a 50/50 chance.
so if 6 GMs won a game against 6 Bronzes (if the matchmaker allowed it) then the GMs would gain 0.001 SR and the Bronzes would loose 0.001 SR.
like I said the system is perfect and the win percentage determines the rewards.
now if you wanna talk about people who manipulate the SR system, that’s a different story
but it doesn’t work as soon as you have one dia and one gold. In the end result it 0 itself out, thats the intention behind it. And i want it to only look in a group for the dia SR, so it makes like 2 dias joined the waiting list.
1 Like
well if a team of 3 GMs and 3 Bronzes face off against a team 3 Masters and 3 Silvers, the odds should be 50/50
those ranks may not be exact, but I hope you see my point
It doesn’t work like that.
PBSR is on individual statistics, and if a bronze player is fighting diamond opponents, their personal statistics will be horrible, and they’ll probably lose more than the average 24 SR for the inevitable loss. There’s absolutely nothing to be gained by horribly imbalanced matches. In fact, the matchmaker will throw away the game if either team has less than a 40% chance of winning. (before even starting it)
1 Like
The change I advocate here is simple, for this calculation always take the sr of the groups highest players instead of the average.
A diamond can carry in a plat game, but it is not that difficult for plat players to use a couple brain cells and counter them. I played against a diamond Symmetra one trick once. I was playing Rein, and swapped to Winston to deal with them: completely nullified them.
In fact, I will go one step further and say as a plat, you SHOULD be able to go toe-to-toe with diamond players if even hope to climb. It’s not enough to play like a plat in a plat game to climb. You need to be playing one level above your rank if you want to climb.
So, long story short, no, just get better.
I hope that they’re weighing in the actual skill/performance difference between each tier on the ladder as well, not just the SR numbers themselves; that would just be … dumb
That’s the kind of math that the current matchmaker uses (except only within a 1000SR limit), however it’s not accurate. The GM+Bronze team will beat the Masters+Silver team at very nearly the same rate that an all GM team would beat an all Masters team.
The low skill players are no match for the high skill players and become a non-factor, leaving you with 3GM+ ‘3 warm bodies’ vs. 3Masters + ‘3 warm bodies’.
It kind of works in a typical plat+silver duo playing gold (the plat is more or less carrying a kind of ‘5.5v6’ game in gold which is reasonable) but it gets a lot more broken as you add players at the extremes of the range to the group.
1 plat + 5 silver for example will match much closer to the silver’s SR, the opponent might even be silvers depending on everyone’s exact SR, and that plat will dominate, while the silvers he is carrying will not even be dead weight, they’ll be just slightly worse than their opponents and very much able to contribute.
2 Likes
that seems to be stretching it a little, but I see what you’re driving at
kawumba says that game results are fed back into the system and SR gains readjusted as the computer learns win probabilities for different match ups
if the computer thought a certain player combination had a 1% chance of winning, but it’s seen 300 games and not seen a single victory, it would decrease the chances to 0.75% and readjust the SR rewards accordingly.
the idea is that match maker can never be wrong, because it is only fed off of inputs from the matchmaker itself
blizzard evidently thought it was a good idea to allow people from 2 full skill tiers apart to que together.
SR gap should of been 500 from the get go. You want to play with “friends” play in arcade and qp.
LOL at people who think diamonds cant carry plat games. Sure they cant carry all of them but they sure as hell can make a significant impact.
Put a plat pharah and a bronze mercy to pocket them the whole game in a low gold game and see what happens.
2 Likes
the problem isn’t you yourself, the problem is the rest of the team, not everyone cares so much.
Well, I hope that’s true, but it certainly feels as if they are simply averaging the SR values. If they have a feedback like this then it should actually be getting more accurate and work out pretty well … statistically… even if it doesn’t feel good when you lose to someone from above your rank.
a 75% win chance isn’t exactly money in the bank
OP suggestion is just as if not more unfair. Sucks having a diamond in plat because they queued with their gold buddy, but if instead it brought their gold buddy into a diamond lobby that’s definitely worse.
500 seems too tight, I’ve had good or bad days where I’ve nearly swung that far in one long sitting. Maybe 750 though.
everything would be kosher if the MMR formula were public. Because ppl would say, “help there were Plats in my game and I’m only Silver.”
and then someone would answer, “Well if you just took a second to look at the math, you only lost 5 SR.” “so if you wanna complain about time wasted, I understand. But if you wanna complain about SR and unfairness, just look at the math again.”
500 SR is quite a skill difference no matter where you are on the ladder. If you think 1500 and 1000 SR is the same you havent played down there. High silver and low silver are quite different.
The amount of people I’ve seen as of late down 500 SR is just insane. People really need to learn when to stop queing when tilted. 3 games lost in a row regardless of how close they are should be a que to stop well queing.
1000SR isn’t silver though…
…and I haven’t noticed a skill difference between a 700SR match and a 1200SR match. The only difference really, that I’ve seen, is that 1200SR has a lower chance of throwers and smurfs.
I know 1000 SR is not silver. I am simply stating that 500 SR is quite a gap in skill generally speaking across the whole ladder. 1500 and 1900 are quite noticeable. The majority of 1500 games for me are flat out trivial. 1900 is a different story.