• Log InLog In
  • Register
Liquid`
Team Liquid Liquipedia
EDT 05:43
CEST 11:43
KST 18:43
  • Home
  • Forum
  • Calendar
  • Streams
  • Liquipedia
  • Features
  • Store
  • EPT
  • TL+
  • StarCraft 2
  • Brood War
  • Smash
  • Heroes
  • Counter-Strike
  • Overwatch
  • Liquibet
  • Fantasy StarCraft
  • TLPD
  • StarCraft 2
  • Brood War
  • Blogs
Forum Sidebar
Events/Features
News
Featured News
[ASL22] Finals Preview: Sweet Victory!9[ASL22] Ro4 Preview: Mirror Mirror6[ASL22] Ro8 Preview: Within Reach5[ASL22] Ro8 Preview: In A Tizzy11[ASL22] Ro16 Preview: Holy Diver5
Community News
Evil Geniuses - shutting down10ISSL (IntoTheTV x SOOP SC2 League): Premier5Weekly Cups (October 12-18): 5.0.17 era begins1[$1250] International ShieldBattery League53Weekly Cups (October 5-11): Farewell to 5.0.165
StarCraft 2
General
REPORT: StarCraft 2: Remastered In Development Weekly Cups (October 5-11): Farewell to 5.0.16 No more SC2 streamers on team liquid? Evil Geniuses - shutting down How do you feel about the mass reverts in the 5.0.17 PTR?
Tourneys
RSL goes to London! 2026 Offline Finals Nov 21-22 ISSL (IntoTheTV x SOOP SC2 League): Premier Sparkling Tuna Cup - Weekly Open Tournament ScienceCraft (October 24-25) - Live Event WORTEX Challenger Series — Open SC2 LAN in Budapes
Strategy
[H] ZvP Mid-Late Game: Stalkers Collossi HT
Custom Maps
[G] How to playtest arcade maps [A] Nemrods 1/4 players [M] (2) Sweltering Sands [M] (2) Frigid Storage
External Content
Mutation # 548 Shining Bright The PondCast: SC2 News & Results Mutation # 547 Temple Of Terror Mutation # 546 Catch the Train
Brood War
General
Do we even care about balance? [Megathread] Daily Funny Videos Rollback Netcode in Brood War BW General Discussion Young Players Tournament Announced by SOOP
Tourneys
[BSL23] RO24 Group Stage [ASL22] Grand Finals [$1250] International ShieldBattery League [Megathread] Daily Proleagues
Strategy
Cliff Jump Revisited (1 in a 1000 strategy) Simple Questions, Simple Answers Replay Review Process - What do you do?
Other Games
General Games
General RTS Discussion Thread Nintendo Switch Thread Stormgate/Frost Giant Megathread Armies of Exigo - YesYes? Dawn of War IV
Dota 2
Dota 2 Champions League Season 3 Begins April 25!
League of Legends
[TL LoL EUW IHs] Teemo shall perish
Heroes of the Storm
Heroes of the Storm 2.0
Hearthstone
Deck construction bug
TL Mafia
TL Mafia Community Thread
Community
General
US Politics Mega-thread Russo-Ukrainian War Thread Artificial Intelligence Thread The Chess Thread Canadian Politics Mega-thread
Fan Clubs
Serral Fan Club
Media & Entertainment
[Manga] One Piece Anime Discussion Thread Movie Discussion!
Sports
Football (Soccer) Thread TeamLiquid Health and Fitness Initiative For 2023 MLB/Baseball 2023
World Cup 2022
Tech Support
Computer Build, Upgrade & Buying Resource Thread
TL Community
Recent Gifted Posts
Blogs
How Much Genes Contribute to…
TrAiDoS
ShieldBattery on Mac
Kraekkling
[ASL22] Ro4 Day1 Ticket Giv…
bITt.mAN
38 yo Retired SWE loo…
PurE)Rabbit-SF
Can Bots Beat Pros?? Starcr…
namkraft
Customize Sidebar...

Website Feedback

Closed Threads



Active: 9205 users

SC2 Ladder Analysis - Hidden Ratings.

Forum Index > Closed
Post a Reply
1 2 3 4 5 6 7 Next All
darmousseh
Profile Blog Joined May 2010
United States3437 Posts
Last Edited: 2010-11-02 03:13:07
October 31 2010 23:36 GMT
#1
Mod Edit -
The OP has not responded to the arguments of known community ladder analysts and statisticians like Excalibur_Z, Vanick or Shadowed and blindly continues to expound his theory in the face of facts that prove him wrong. This thread is closed in the interest of containing misinformation. Please continue discussion in the original ladder analysis thread.


SC2 Ladder Analysis - Hidden Ratings.


Huge Edit

Apparently one of the fundamental assumptions I made was wrong. It seems that leagues fluctuate constantly and that the threshold for a league is not a static number. Therefore it is impossible to determine the size and scope of a league.

On November 02 2010 06:08 Excalibur_Z wrote:
Show nested quote +
On November 02 2010 04:58 darmousseh wrote:
On November 02 2010 04:53 Excalibur_Z wrote:
On November 02 2010 04:45 c0ldfusion wrote:
On October 30 2010 22:17 Cade wrote:
On October 30 2010 20:43 rozina wrote:
Sure there is a 20/20/20/20/20 split. However there is more divisions in brozne than in diamond, therefor there is more people in brozne. A division is a ladder with 100people. And you have the same named division across all the leagues and each has 100 people. So there is obviously 20% in each. However not all divisions have a diamond version since there is not as many diamond players. SC2 ranks is quite accurate imo.


ya I think almost everyone in this thread is not understanding the difference between a

20/20/20/20/20 division split

and a

20/20/20/20/20 league split

there IS a 20/20/20/20/20 division split, there is NOT a 20/20/20/20/20 league split.


This is possible. But I distinctly remember Dustin saying that if you're in Diamond league you're in the top 20% of your region, which implies a league split. Of course this doesn't mean that sc2ranks is inaccurate. It could be that the ladder system is designed to eventually approach 20/20/20/20/20. There are reports on threads here saying that the percentage of diamond league players as reported by sc2ranks increased over time.


I have a huge correction on this. It actually is a 20/20/20/20/20 league split, but with a caveat: the split applies only to active players!. I don't know what the inactivity period is and Blizzard didn't elaborate, but that's the explanation. Naturally the league distribution will fluctuate a bit over time, only roughly hovering around 20/20/20/20/20, but nowhere near the pyramid that we're seeing at SC2ranks. That's the reason why -- those league numbers aren't adjusted for inactivity like Blizzard's are.


Wow, that is a huge piece of information. That kinda throws my whole MMR theory out of whack if true. Where did you get this information?


From Bashiok via email.

Show nested quote +
Hello again. I was completely wrong.

Diamond is 20% of all players. Each league represents roughly 20% of active players. So, Diamond is rather huge. It attempts to keep 20% at all times although it could very well fluctuate, BUT, not enough to scale all the way down to 7% as sc2ranks shows.

It’s likely that they’re so far off because they’re not accounting for active players (or rather, inactive), which we do.

I’ll count this as the last time I open my big yapper before investigating.


Without being able to assume that leagues are static it is very difficult to tell the point differential within a league.

The only thing that I am left with then is to say that it is still likely that MMR of top platinums is near the MMR of low diamonds and that the phenomena of the 1000-1600 rule is probably scaled based on the league. This means that it probably is something more like


1000silver = 1000 mmr
1600 silver = 1400 mmr
1000gold = 1400 mmr
1600 gold = 2200 mmr

Normally i could figure out the scale if you could create a game between a 1000 player and a 1500 player, however, the matchmaking system is too clever and will never match players with that much of a mmr difference.

Blizzard is a lot cleverer than i thought.

I will be taking down sc2stars.com as it now is meaningless.






Edit:

I made a website real quick so you can estimate your MMR based on the formula mentioned.
http://sc2stars.com
It's not fancy, i just made it in 15 minutes.



Blizzard, in order to encourage players to play the game, start all players with 0 rating and increase over time. As you gain leagues you will sometimes lose points and this varies. Also it is impossible to rate yourself against anyone else other than by comparing points and divisions.

Tonight i present the unlocking of the mystery of MMR and a formula which i believe is very accurate at determining your MMR.

First off let me say that MMR contains two components, your mean and your deviation. In most systems your displays rating is mean - 2 * deviation since 2 deviations is about 95% correct. Over time, your deviation will increase. Initially you might be 1500 mean with 150 deviation so your rating will be about 1200. If you play 200 games and maintain the same mean, your deviation will drop to something like 25 so now your rating will be about 1450. Unfortunately it is impossible to determine the deviation without more information, however, it is possible to determine the MMR.


Background

First off we know that blizzard is using an ELO type of system. It is highly likely that it is either trueskill (made by microsoft) or glicko-2. Let us assume that they are using trueskill (it doesn't make that much of a difference).

Also, the numbers I am using are the numbers I came up with based on the data. It is very likely that these are not the "actual" numbers, however I believe they are very close. Also, these numbers do not work out perfectly for people with less than 50 games or so, therefore i have added a modifier to adjust it.


First off, please read the following thread.

http://www.teamliquid.net/forum/viewmessage.php?topic_id=142211

From this we know that it is highly likely that the leagues represent certain skill levels.
Also the bonus pool is used as a means to encourage players to play.



Now please read

http://us.battle.net/sc2/en/forum/topic/628075819

The important pieces we get from here are.

Q: How do I increase my bonus pool? Do I have to not play?
A: Your bonus pool increases at the same rate regardless of whether or not you're playing. The rate of gain is about 1 point per 2 hours.


From http://us.battle.net/sc2/en/forum/topic/540694801

Q. What is the Bonus Pool and how are bonus points acquired?
A. The Bonus Pool is an accumulation of points that every player receives whether they're online and playing or not. They're essentially used as a means to help give a player a catch-up boost if they haven't played in a while. The pool does have a cap, but it increases slowly until the end of a season.


These two combined mean that there is a cap on bonus points and that everyone gets the same bonus points.



Q: Is the bonus pool considered in any calculations? How can that be accurate if it constantly inflates ratings?
A: It has been confirmed that the bonus pool is not factored into any MMR or displayed rating calculations.


Bonus pool is used for something else.



Q: I'm in Bronze with 0 rating and a poor win ratio. I'm only getting 2 or 4 or 6 points -- including bonus pool -- for a win, but losing 22 points for a loss! Why is this happening, and how will I possibly get above 0 rating as long as this is happening to me?
A: This one was tough to figure out because it doesn't make a whole lot of sense on the surface. Point gains and losses that are that extreme can only mean that you're Favored in every single match, but how is that possible if you're already in Bronze with 0 rating? The answer is that your MMR can go below Bronze 0. Because your rating -- which is zero -- is higher than your MMR which is somewhere far below zero, you will constantly be Favored. Because of the way MMR gains work, however, as long as you maintain a positive win ratio, you will eventually find matches where you are Slightly Favored and ultimately Teams Even and beyond. At that point, you will start to gain more points than you lose.
Speculation: This may be a remnant from Beta, either from when Copper league was removed or when players started at 0 rating when they used to start at 1000 rating. It could be that Bronze 0 is equal to Beta's 1000 rating (and therefore Bronze 0 = 1000 MMR).

More info: Allowing MMR to dip below the minimum point value is one way to help ensure that inflation beyond the bonus pool does not occur. The player who loses will have his MMR dip below Bronze Zero, which will cause him to be Favored in every matchup and therefore, because points drift toward MMR, he will remain at zero. This means that any points that are gained are quickly taken away by any losses, thus vacuuming up any points that would have been not really "earned" by the points floor until his MMR rises above the equivalence of Bronze Zero. This stops the production of extra points from the floor of 0.


What we can take away from this is very very important. It means that the system gives you points not based on your rating vs your opponents rating, but on on your rating and opponents points. Why do they do this?


More broadly, your points drift towards your MMR. Because of this mechanism any inflation that occurs will be counteracted by the game outcomes. That is, if point inflation in fact existed, then your points would eventually outpace your MMR and you would then start gaining less points on a win and losing more points on a loss, thereby negating any "inflation" that might occur. Note that this assumes that MMR does not inflate.


Eventually your points will be the same as MMR. but does that mean a 1000 brozne is the same as a 1000 diamond? Absolutely not. It just means that you have 1000 rating in the bronze division.

Now after getting all of this information I was curious about the divisions. Is it possible that your current points can be used to determine your MMR? (or at least MMR-2*sigma)

My conclusion is yes. The following is the formula i came up with.

Formula to determine MMR

division [bronze = 0, silver = 1, gold = 2, platinum = 3, diamond = 4]

(division - 1) * 500 + (points) + (bonus pool remaining) + (1500 - total bonus pool)

For example i am a 358 platinum player with 1004 bonus points remaining. Currently all players have received 1322 bonus pool
so my MMR is (3 - 1 ) * 500 + 358 + 1004 + 178 = 2540.


How I came up with this formula.

First off, I need to clarify that this is only possible if it is true that your points will converge to your actual rating over time.

Thanks to sc2ranks.com I noticed the following patterns.

In all leagues except bronze, there was few or no players under 1000 points as long as they had used up their entire bonus pool.
In all leagues except diamond, there were few or no players over 1500 points as long as they had used up their entire bonus pool.

First off I will assume that a 0 bronze is actually -500 rating. (There are reasons for this, I will explain later on)

So a 0 bronze = -500 rating.
1500 bronze = 1000 rating.

Ok now let's assume that silver starts at 1000. Notice above that people in silver that have no bonus pool remaining should have between 1000-1500 points so
1000 silver = 1000 rating
1500 silver = 1500 rating.

And so forth

1000 gold = 1500 rating
1500 gold = 2000 rating

1000 platinum = 2000 rating
1500 platinum = 2500 rating

1000 diamond = 2500 rating
1500 diamond = 3000 rating.

Now, for those who still have a bonus pool left, simply add their bonus pool to their score.


For me at 2540 it means i am very close to getting promoted to diamond, however, probably won't be promoted unless i maintain that rating for a certain number of games or period of time.


After examining a number of different people with different ratings, bonus pools, divisions I came to the conclusion that this was absolutely right. There is some overlap between divisions (a 2600 player might be in platinum and a 2400 might be in diamond), however, it was easy to tell that the overlap was very small.


What i am proposing.
That the points you gain are not simply based on your points compared to your opponents MMR, but is based on your points + (division - 1)*500 compared to your opponents MMR. This is the way to ensure that your rating will always converge towards your actual rating.



Things to note:

These groupings could be wrong. I had a hard time deciding whether it was 500 points a group or 600, but i noticed that people who were rated near the 1600 mark had been winning a lot of games recently and were likely to get promoted.

In my calculations I also noticed that bronze must be a very very wide group.

That the distirbution of points in the different divisions never changes, it is always 1000-1500 in silver through platinum. This means that rather than the rating cap changing constantly, that the number of people in each division changes drastically.

The reason i started bronze at -500 is because of how the chess ELO system works. In chess the average player ELO is 1500, however, it is important to note in chess that people who have ratings are people that take it seriously. In starcraft 2 it seems that blizzard did not take this into account as there seem to be a significant number of people with a rating of 0. Based on information from sc2ranks.com The average player MMR would be about 1200 which makes sense considering that a lot of players don't play seriously.

Also it seems that blizzard believed, based most likely on the results from the beta, that the groups should have been distributed about 20% each, however, pros have gotten signficantly better, while noobs are still noobs. This is the reason for players like Huk to have a 4200 rating according to my formula.

If this is true, then it also means that blizzard underestimated just how much of a gap between a noob and a pro would be. Ratings are simply mean to be a way of comparing players. You can estimate the chances of beating a player based on this.

In chess, ratings are not nearly as exaggerated because of 2 reasons. 1: The only people that play are those that are serious. 2: The skill gap between a grandmaster and an amateur is pretty small.




My thoughts on the future:

Blizzard will likely change the groups to reflect their desired distribution. Current distribution is like 42/18/18/14/8. My guess is that it will be 30/20/20/20/10 in the future so they will adjust the boundries accordingly.

Blizzard will likely lower the rating for 0 bronze, by proably another 500 points.

Blizzard has said that it will only take 1 placement match for the next season. This is probably because MMR will remain between seasons.

Final Notes:

I am willing to accept my theory is wrong, but please provide evidence.





TLDR;

My conclusion is yes you can determine your approximate MMR. The following is the formula i came up with.

division [bronze = 0, silver = 1, gold = 2, platinum = 3, diamond = 4]

Rating = (division - 1) * 500 + (points) + (bonus pool remaining) + (1500 - total bonus pool)

For example i am a 358 platinum player with 1004 bonus points remaining. Currently all players have received 1322 bonus pool
so my MMR is (3 - 1 ) * 500 + 358 + 1004 + 178 = 2540.



EDIT; After doing some testing I have added the following to my formula.

Rating = (division - 1) * 500 + (points) + (bonus pool remaining) - (2 * 150 * (1-games played/50) #if you have played under 50 games.)

The reasoning being that i noticed that players will a really small number of games will have artifically inflated ratings. By punishing players who have played fewer games i noticed that you can even the distribution significantly to the point where it's an even distribution across the whole division. 150 is a standard for chess as well (ironic) and 50 games is based on an article i read about trueskill which is near the point where your deviation starts to flatten signficantly. These numbers seemed to work well in all divisions, but seemed to work more significantly in the lower divisions (which i'm guessing is because noobs improve more significantly over time than pros do. A noob might go from 0 rating to 1000 rating in a few weeks. A pro might go from 3000-3200 in a few months)


This actually leads me to believe that ratings blizzard is using actually adds 2 sigmas to your score when calculating your rating rather than subtracting them as normal.


EDIT2; Please note that i didn't come up with these numbers on my own. I used information from sc2 ranks, applied the formula to 1000 random people from each division and checked the distribution to make sure it looked correct. I will present the excel file with the information when i clean it up a bit and make it look better.
Developer for http://mtgfiddle.com
darmousseh
Profile Blog Joined May 2010
United States3437 Posts
November 01 2010 00:05 GMT
#2
please also note that i will be adjusting the numbers once i mine the data a little more, however, i believe the formula by itself is very close to accurate.

Also note be accurate, i mean accurate to one of your MMRs.

There are 3 likely MMRs.

Mean - 2 * sigma
mean
mean + 2 * sigma

I have feelings that the points displayed represent the 3rd one so that players have a little bit higher points.
Developer for http://mtgfiddle.com
FiWiFaKi
Profile Blog Joined February 2009
Canada9859 Posts
November 01 2010 00:08 GMT
#3
If you actually found how to determine the MMR that would be an impressive feat for sure. So how do you verify it now? Do you put up many players against each other and see if they are favoured or not? Or what's the plan?

Really cool nontheless. Congrats
In life, the journey is more satisfying than the destination. || .::Entrepreneurship::. Living a few years of your life like most people won't, so that you can spend the rest of your life like most people can't || Mechanical Engineering & Economics Major
darmousseh
Profile Blog Joined May 2010
United States3437 Posts
November 01 2010 00:10 GMT
#4
On November 01 2010 09:08 Skillz_Man wrote:
If you actually found how to determine the MMR that would be an impressive feat for sure. So how do you verify it now? Do you put up many players against each other and see if they are favoured or not? Or what's the plan?

Really cool nontheless. Congrats



I used the data from sc2ranks, applied the formula i discovered and made sure that the distribution was even across the division and across all the divisions put together. That's the only way i can think of to verify it. Also i played some games today, checked how many points i won or lost, checked the point differential expected from the number of points i won or lost and compared it to my opponents rating. It also has shown to be very accurate.
Developer for http://mtgfiddle.com
FiWiFaKi
Profile Blog Joined February 2009
Canada9859 Posts
Last Edited: 2010-11-01 00:19:15
November 01 2010 00:19 GMT
#5
On November 01 2010 09:10 darmousseh wrote:
Show nested quote +
On November 01 2010 09:08 Skillz_Man wrote:
If you actually found how to determine the MMR that would be an impressive feat for sure. So how do you verify it now? Do you put up many players against each other and see if they are favoured or not? Or what's the plan?

Really cool nontheless. Congrats



I used the data from sc2ranks, applied the formula i discovered and made sure that the distribution was even across the division and across all the divisions put together. That's the only way i can think of to verify it. Also i played some games today, checked how many points i won or lost, checked the point differential expected from the number of points i won or lost and compared it to my opponents rating. It also has shown to be very accurate.


Hmm, well that's pretty sweet. Now I'm trying to find a practical place for where you could apply it O.o... I can see a subsection in like sc2ranks where they could use the hidden MMR to determine the rating rather than the ladder points, no other applications though?

I do really think it's cool finding the hidden formula that Blizz wouldn't release if that's really it. So yeah, good job! I hope your thread gets more attention too
In life, the journey is more satisfying than the destination. || .::Entrepreneurship::. Living a few years of your life like most people won't, so that you can spend the rest of your life like most people can't || Mechanical Engineering & Economics Major
Ketara
Profile Blog Joined August 2010
United States15065 Posts
Last Edited: 2010-11-01 00:25:07
November 01 2010 00:24 GMT
#6
One way you could try to verify this is when one of the top 200 lists come out, since those are done by MMR rather than points, you could look up the ratings of all those players and see if your formula organizes them in the same order that Blizzard organizes them.

Of course, this would assume that the lists come out the same day they are calculated and none of these players plays any games between when the list comes out and when you check SC2ranks.
http://www.liquidlegends.net/forum/lol-general/502075-patch-61-league-of-legends-general-discussion?page=25#498
Subversion
Profile Blog Joined April 2010
South Africa3627 Posts
November 01 2010 00:35 GMT
#7
Excalibur_Z... A challenger steps forward.

Interesting read.
Xapti
Profile Joined April 2010
Canada2473 Posts
Last Edited: 2010-11-01 01:03:25
November 01 2010 01:01 GMT
#8
current league division is 20/20/20/15/5 AFAIK, not what you stated.

IMO it should be more like 40/30/20/8/2 (although master league may affect this)

You also proposed that when playing a game, the final point change is based off your rating vs their points? I think that's mistaken, I think it's more like your points vs their rating, although definitely still a mix of both for each, not just one or the other.
"Then he told me to tell you that he wouldn't piss on you if you were on fire" — "Well, you tell him that I said that I wouldn't piss on him if he was on Jeopardy!"
darmousseh
Profile Blog Joined May 2010
United States3437 Posts
November 01 2010 01:02 GMT
#9
On November 01 2010 09:35 Subversion wrote:
Excalibur_Z... A challenger steps forward.

Interesting read.


lol. i'm still waiting for criticism and people's thoughts. if someone would like to verify this independently please do so.
Developer for http://mtgfiddle.com
darmousseh
Profile Blog Joined May 2010
United States3437 Posts
November 01 2010 01:07 GMT
#10
On November 01 2010 10:01 Xapti wrote:
current league division is 20/20/20/15/5 AFAIK, not what you stated.

IMO it should be more like 40/30/20/8/2 (although master league may affect this)

You also proposed that when playing a game, the final point change is based off your rating vs their points? I think that's mistaken, I think it's more like your points vs their rating, although definitely still a mix of both for each, not just one or the other.


To be exact it's
your points + (division - 1) * 500 + bonus pool + (1500 - total bonus pool) against their MMR.

This is the way in order to make sure that YOUR points go towards the correct MMR.
Developer for http://mtgfiddle.com
Rahlekk
Profile Joined September 2010
United States107 Posts
November 01 2010 01:08 GMT
#11
Lemme see if I got this right. I'm in Bronze, 1632 points, 36 bonus points.

(division - 1) * 500 + (points) + (bonus pool remaining) + (1500 - total bonus pool)

(0 -1) * 500 + (1632) + (36) + (1500-36)

And that equals: 2632

Is that good, bad? I don't really know what that number tells me.
viel gluck TLO ^^ | 행운을 빌어요 BoxeR
vesicular
Profile Blog Joined March 2010
United States1311 Posts
November 01 2010 01:12 GMT
#12
On November 01 2010 10:08 Rahlekk wrote:
Lemme see if I got this right. I'm in Bronze, 1632 points, 36 bonus points.

(division - 1) * 500 + (points) + (bonus pool remaining) + (1500 - total bonus pool)

(0 -1) * 500 + (1632) + (36) + (1500-36)

And that equals: 2632

Is that good, bad? I don't really know what that number tells me.


Total bonus pool is 1322, so your rating is really:

(0 -1) * 500 + (1632) + (36) + (1500-1322) = 1346

STX Fighting!
darmousseh
Profile Blog Joined May 2010
United States3437 Posts
November 01 2010 01:14 GMT
#13
On November 01 2010 10:08 Rahlekk wrote:
Lemme see if I got this right. I'm in Bronze, 1632 points, 36 bonus points.

(division - 1) * 500 + (points) + (bonus pool remaining) + (1500 - total bonus pool)

(0 -1) * 500 + (1632) + (36) + (1500-36)

And that equals: 2632

Is that good, bad? I don't really know what that number tells me.



Close.

(0 - 1) * 500 + (1632) + (36) + (1500 - 1331) = 1337

So you will likely be promoted to silver soon.
Developer for http://mtgfiddle.com
Rea
Profile Joined June 2010
Germany88 Posts
Last Edited: 2010-11-01 01:28:19
November 01 2010 01:26 GMT
#14
according to ur formula someone with more points, including bonuspool, has a higher mmr...
thought its well known since early beta that this is not the case


just go through any matchhistory and compare the players where one got +12 or -12p before bonuspool and compare their ladderpoints, they wont always be similar
(`.*(C=(`.´Q)
darmousseh
Profile Blog Joined May 2010
United States3437 Posts
November 01 2010 01:27 GMT
#15
On November 01 2010 10:26 Rea wrote:
according to ur formula someone with more points, including bonuspool, has a higher mmr...
thought its well known since early beta that this is not the case


Well, blizzard told us that points will converge towards your rating, so it is true over time.
Developer for http://mtgfiddle.com
dave333
Profile Joined August 2010
United States915 Posts
November 01 2010 01:28 GMT
#16
^1337 is a cool MMR :D
rauk
Profile Blog Joined February 2009
United States2228 Posts
November 01 2010 01:29 GMT
#17
On November 01 2010 10:27 darmousseh wrote:
Show nested quote +
On November 01 2010 10:26 Rea wrote:
according to ur formula someone with more points, including bonuspool, has a higher mmr...
thought its well known since early beta that this is not the case


Well, blizzard told us that points will converge towards your rating, so it is true over time.


blizzard also said they have a moving average that's only valid for the last 30ish games, so any system that takes into account every game is automatically an invalid guess as to how mmr works.
Rea
Profile Joined June 2010
Germany88 Posts
Last Edited: 2010-11-01 01:32:13
November 01 2010 01:31 GMT
#18
On November 01 2010 10:27 darmousseh wrote:
Show nested quote +
On November 01 2010 10:26 Rea wrote:
according to ur formula someone with more points, including bonuspool, has a higher mmr...
thought its well known since early beta that this is not the case


Well, blizzard told us that points will converge towards your rating, so it is true over time.


?
rating can be measured in 0 = newb and 10 = the best, or from 100 = newb to 100³³ = best
until someone proofs me that ladderpoints are directly linked to the rating you just cant use either to determine the other

alone the fact that someone gets 12points and his opponents loses only 11 should tell everything
(`.*(C=(`.´Q)
darmousseh
Profile Blog Joined May 2010
United States3437 Posts
November 01 2010 01:36 GMT
#19
On November 01 2010 10:29 rauk wrote:
Show nested quote +
On November 01 2010 10:27 darmousseh wrote:
On November 01 2010 10:26 Rea wrote:
according to ur formula someone with more points, including bonuspool, has a higher mmr...
thought its well known since early beta that this is not the case


Well, blizzard told us that points will converge towards your rating, so it is true over time.


blizzard also said they have a moving average that's only valid for the last 30ish games, so any system that takes into account every game is automatically an invalid guess as to how mmr works.



The way trueskill works is that the last 30 games have more of an impact on the previous games, absolutely it is impossible to find the actual MMR, but for players who have not fluctuated recently, points will converge to your rating. That was probably the single most important piece of information that i found out recently which i referenced in the OP.
Developer for http://mtgfiddle.com
Medzo
Profile Blog Joined June 2009
United States627 Posts
Last Edited: 2010-11-01 01:45:07
November 01 2010 01:40 GMT
#20
I just want to point out that your hidden rating isnt determined by your shown rating at all. If you look at the top 200 a lot of those people are not in the top 200 on scranks (shown rating ranks). It is because their hidden ratings are higher than other people with higher point ratings. Basically what it seems like to me is you are ranked in your division by your hidden rating and the system will correct itself and give you the correct amount of shown rating points so that this is accurate.

I have 2 reasons for this hypothesis. First, In my own division pylon kappa last week everyone rank 9 and above was in the blizzard hidden rating top 200 in the same order of their ranking. The 9th ranked person was 1965 points which is about rank 400 in shown rating. My second reason is just from playing at a ladder and profiling my opponents when im favored or unfavored. If you have a high hidden rating you will fight other players who you are favored over who have a higher shown rating, they are usually rank 1 or 2 in their division with about 2k points but have hidden ratings similar or lower than 1600-1800 in harder divisions. At the same time you can play another player with lower or equal shown and hidden rating in equally hard or harder (medivac alamo) divisions.

This is easily noticeable when you have a very high hidden rating.
1 2 3 4 5 6 7 Next All
Please log in or register to reply.
Live Events Refresh
Next event in 1h 17m
[ Submit Event ]
Live Streams
Refresh
StarCraft: Brood War
Britney 25085
Calm 4318
Shuttle 1880
firebathero 699
Bisu 600
EffOrt 160
ZerO 149
Soma 143
Leta 114
Mong 105
[ Show more ]
Dewaltoss 96
Mini 82
Hyun 63
Rush 58
Killer 43
Barracks 40
Sharp 38
NaDa 36
910 34
NotJumperer 26
Trap 23
Hm[arnc] 21
Shinee 13
ajuk12(nOOB) 11
HiyA 11
Free 9
Noble 9
sorry 7
Terrorterran 6
Dota 2
Gorgc5348
Fuzer 35
XaKoH 6
League of Legends
JimRising 434
Counter-Strike
olofmeister1597
allub401
dupreeh321
gobbtv127
edward110
Heroes of the Storm
Trikslyr44
Other Games
summit1g11328
ceh9640
crisheroes251
Livibee251
Sick43
ZerO(Twitch)6
Organizations
Other Games
gamesdonequick451
StarCraft: Brood War
UltimateBattle 78
[ Show 14 non-featured ]
StarCraft 2
• Berry_CruncH289
• LUISG 27
• AfreecaTV YouTube
• intothetv
• Kozan
• IndyKCrew
• Migwel
StarCraft: Brood War
• iopq 4
• FirePhoenix1
• BSLYoutube
• STPLYoutube
• ZZZeroYoutube
League of Legends
• Stunt589
• Jankos505
Upcoming Events
IntoTheTV X SOOP
1h 17m
Reynor vs Zoun
Neeb vs Rogue
OSC
13h 17m
FEL
1d 23h
Patches Events
2 days
AI Arena Tournament
2 days
FEL
3 days
Sparkling Tuna Cup
3 days
Replay Cast
3 days
WardiTV Weekly
4 days
IntoTheTV X SOOP
5 days
ByuN vs Lambo
MaxPax vs Bunny
[ Show More ]
PiGosaur Cup
5 days
The PondCast
6 days
Kung Fu Cup
6 days
Liquipedia Results

Completed

Proleague 2026-10-20
Blizzard Classic Cup 2026
HCC Season 3

Ongoing

Acropolis #5
IPSL Winter 2026-27
KCM Race Survival 2026 Season 4
Acropolis #5 - GSC
1win Private Club #2
Thunderpick World Champ. '26
ESL Pro League Season 24
Stake Ranked Episode 4
1win Private Club #1
Logitech G Play Connect 2026
SL StarSeries Fall 2026
FISSURE Playground #3
BLAST Open Fall 2026
Esports World Cup 2026

Upcoming

BSL Season 23
SC4ALL II: Brood War
BSL 23: Non-Korean Championship
HSC XXX
Stellar Fest 2: Lunar Cup
SC4ALL II: StarCraft II
Kung Fu Cup 2026 Grand Finals
RSL Offline Finals
eXTREMESLAND 2026
PGL Major Singapore 2026
Stake Ranked Episode 6
BLAST Rivals Fall 2026
IEM Beijing 2026
Stake Ranked Episode 5
PGL Masters Bucharest 2026
TLPD

1. ByuN
2. TY
3. Dark
4. Solar
5. Stats
6. Nerchio
7. sOs
8. soO
9. INnoVation
10. Elazer
1. Rain
2. Flash
3. EffOrt
4. Last
5. Bisu
6. Soulkey
7. Mini
8. Sharp
Sidebar Settings...

Advertising | Privacy Policy | Terms Of Use | Contact Us

Original banner artwork: Jim Warren
The contents of this webpage are copyright © 2026 TLnet. All Rights Reserved.