Halo 3 Forum
This topic has moved here: Subject: In depth explanation of the Halo 3 skill ranking system.
  • Subject: In depth explanation of the Halo 3 skill ranking system.
Subject: In depth explanation of the Halo 3 skill ranking system.

Posted by: o MANNY o
P L E A S E R E A D A N D R E P L Y
The true skill system is broken. I'm a 40 in lone wolves. and only a 28 in team slayer. this would be fine with me if my win/lose weren't so drastic. I have lost a total of 11 games in team slayer but i have won 126 games. I have 66 steaktacular metals. That mean that 66 games we have won by at least 20 kills. Team Slayer have become boring because the competition is so weak. My friends and I get together and play team slayer all night but lately we just stop playing together and play lone wolves or something else because, as stated prior, The competition is weak being matched up with other 28s. If my friends all have the jump on my rank. they are in the 40s or so. I try to play team slayer solo but I have the wears luck with random teammates. my 11 loses are mostly from doing just that. almost every lose I lead both teams in kills. I'm human and I'm not always playing smart but I feel like I'm in some sort of matchmaking purgatory.
Any Ideas???????????????????????????????????????????????????????? ???????????????????????????????????????????????????????


You are playing in a mixed party and the system is having trouble with mixed parties. This quote was posted earlier in the thread but 10 pages is a lot to read through so it's no surprise you didn't read it. Here it is again for your convenience with a link to the original source.

Posted By: Ferrex - Original Source

How? What are you fixing?
First, there is a general matchmaking issue affecting parties of mixed skill levels, which slows (or halts) the progress of players in those parties. This is made worse in large teams, and doubly so in Social playlists were skill levels are not displayed. Top men at Bungie are investigating the cause for this issue (tentatively identified as an unintentionally vigorous guard against boosting exploits that were a problem in Halo 2) and effective fixes for it. All of our playlists will benefit from dialing this in, as soon as we feel good about the changes.

  • 11.23.2007 11:35 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

Posted by: TwoFaced1680
Posted by: matt1242
15 games straight ---> 3 ranks, at LOW (18) level in doubles not to mention a perfection without a rank up. Take into account those 15 games that we won in team doubles are more total than I have played in the rest of team doubles while still achieving the rank of 18 in less than 15 games. My teamate and me did around same (but I consistently had a much better ratio, usually only a FEW deaths per game and was usually mvp) and he advanced 23 levels...... hmmmm so he's at a 34 and I'm at a 21, that doesn't make sense at all. Another PROVEN testimonial that the ranking system IS BROKEN. I just don't know how else to convince the kids that swear by it..... but it's coming for you sooner or later, and you will curse it and realize that you were talking out of your ass, I promise you that much.


I checked the gamertag linked to your account...Is that the correct one? Also your friend is Ix Rumz xI?Seems to me like you are a little mixed up with how many games you played. I looked through all 75 pages of games you have listed (that was pretty boring by the way, but I do what I can to get at the truth) and before the streak you had played 69 games. I'll say that again...You played 69 games of Double Team BEFORE the win streak. Your friend had played 6 games before the win streak...Once again...Your friend played only 6 games BEFORE the win streak!

Seems to me this fits right into the system as we currently understand it. Your friend moved much faster because the system was far less certain of his skill level. He had also played far fewer games then you which would mean he is going to move faster then you.

Correct me if any of my facts are wrong, but please look back at your games before you try to tell me otherwise. Most of the Team Doubles games you played occured on page 40 and back. That should help you speed up the search a little.


Thank you, TwoFaced, I had been looking through his games, but I got bored at around page 25, and decided to come back to it later. Yup, the 69 prior games would lower your Sigma significantly compared to your friend's 6. That's why you didn't rank up as fast, because the game was more certain of your skill level.

TwoFaced1680 has hit the explanation right on the money. I'd like to thank him and all the other dedicated posters in this thread for their help in the defense of my explanation. Jay120171 and SatansReverence come to mind. If I've not mentioned you here and you've been helping, my sincerest apologies. I appreciate the effort.

Also, how many individual OMG THE SYSTEM DOESN'T WORK cases do we have to easily slot into the mathematical construct before the doubters are silenced? Do your research before you claim the system doesn't work, or if you don't feel like doing the research, ask someone to explain to you why things happened the way they did in your case. Or send me a PM. I've been responding to them, albeit slowly. 10+ pages and nobody has come up with a case that DOESN'T fit into the system.

To sum up, if you don't think your case fits, present us with a listing of all your games in that hopper (or at least the date ranges on which they were played) and tell us your problem. We'll be able to figure it out and explain to you how your experience fits into the system, and probably we'll be able to offer suggestions on how to alleviate the trouble you've been having.

  • 11.23.2007 11:55 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

true satansrevenge, I was just a little aggrivated after the lack of ranking up and didn't remember all of the early games before I had any idea of what I was doing in halo 3. Both my posts were actually wrong compared to what two faced said though so it seems I might have been mistaken on my approximation and it was my fault in the first place, it can just be irritating.... but haha two face I can't believe that you went through all of my games played and kept a count, that's awful, you really have nothing better to do haha.

[Edited on 11.24.2007 12:00 AM PST]

  • 11.23.2007 11:56 PM PDT

El Kafungus,

Someone earlier in this thread posted a quote by Ferrex about the system having issues with mixed parties. I quoted it again a couple posts back because I believe it was possibly part of the problem for that particular person. I noticed you don't have any info about this issues in the main post. Perhaps adding the quote and a link to the source would be a good idea. Just a suggestion.

I feel like this thread is actually making some headway. Either that or people have just gotten sick of complaining :) Either way it's a great thread and it's nice having something I can just point others to. I said it way back on the first page but I'll say it again..."Very nice post"

[Edited on 11.24.2007 12:35 AM PST]

  • 11.24.2007 12:18 AM PDT

I went thorugh some of Achronos' pevious post and found this Quote

Posted by: Achronos
You're not paying attention.

It DOES NOT MATTER how much you beat them by. It only matters whether you won, or loss. Other input variables include your level, your team members level, how many games each of you have played, and similar values for your opponents.

The game does not care at all that you beat the other team 50-1 or 50-49. It cares only that you won. It also doesn't care if you were the MVP, how many kills you got, or anything else.

Posted by: No iTz not luck
Did you even look at our games LMFAO, we str8 killed those kids 50-20 or 50-10 we got a steak almost every game. We went more then positive 10..


p.s Not a problem mate, your thread helped me so I will help you and your thread.

  • 11.24.2007 1:18 AM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

very nice well thought out post, thanks for writing. Hopefully this will help counter all the complainers whining about their rank on the forum.

  • 11.24.2007 11:36 AM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

hey thanx 4 that in depth talk. its been killing to no how the Ranking System worked 4ever. but thanx 2 u i dont hav to strangle the info out of my buddies anymore. HIGH-5

  • 11.24.2007 12:31 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

very nice, it actually helped me alot

  • 11.24.2007 3:52 PM PDT

-----------------------------------------------------------------------------------
"Strength and honour..." Maximus Decimus Meridius,
The Gladiator

Interesting......guess I should've paid more attention in my high school stats class...... I didn't realize it was this complex.......

  • 11.24.2007 5:17 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

what does "mu" and "sigma" mean.
Sorry if this has already been posted and replied to, its just i cant be bothered to read over 200 replys.

  • 11.24.2007 5:28 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

Posted by: CaptainCrunch117
hey thanx 4 that in depth talk. its been killing to no how the Ranking System worked 4ever. but thanx 2 u i dont hav to strangle the info out of my buddies anymore. HIGH-5


your not sending a message of a phone, write in english. not some jumbled up, un-understandable words. it dont understand "ftw" it makes no sense.

  • 11.24.2007 5:30 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

Too bad it doesn't explain how you can have 10 games on your 49, lose 1 game to 47-49s and drop back to a 48.

  • 11.24.2007 9:09 PM PDT

Posted by: J3TTY
Too bad it doesn't explain how you can have 10 games on your 49, lose 1 game to 47-49s and drop back to a 48.
I believe it does answer that. Is this what you are looking for?




"Why do I win 10 games and not level up, then delevel when I lose 2? Help!"

There are several possible explanations for this. The easiest one, and the one that fits the Trueskill system EXACTLY as explained by Microsoft, is that losing against a team that is X (I believe 6 or more, but experimentation is needed to determine the true value) skill points below you can cause you to lose as much as 5 times the Mu value that winning against even level teams gains you. Of course, winning against a team that has X more skill levels than you can net you as much as 5 times the Mu gain. The reverse is true as well. Winning against a weak team can count as little as 1/5th of an even skill win.

Second, and this is enlightened conjecture, your Sigma value was low at the beginning of your streak, and since your win streak is raising it, the conservative ranking system (Remember, Rank = Mu MINUS (K times Sigma)) is kicking in to reduce your apparent skill level. Your real Mu is higher, but your higher Sigma is having a negative impact.


If not why don't you post the range of games and the playlist you are questioning. I'm sure the OP or someone else will do there best to help answer.

[Edited on 11.24.2007 9:27 PM PST]

  • 11.24.2007 9:26 PM PDT

AN EYE FOUR AN EYE

yo i have a question. i dont know if this was asked or answered before. can someone tell me how i ranked up when when i lost the match. (im talking about ranked team slayer)

  • 11.24.2007 9:42 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

A very good summary, Kafungus, and kudos to sticking it out in the thread; I would have thrown my hands up and left the thread after some of the flames).

A couple of additional points, both for the OP and anyone else who's interested in how TrueSkill deals with team games (I'm not with Bungie or Microsoft, so I'm just parsing information from here):

1) In team games, the TrueSkill algorithm treats a team as a single player with mu (estimated skill) equal to the sum of the mu for each player, and sigma (uncertainty) equal to the sum of the squares of each player's sigma. This means that if one player on your team has played a lot of games (and TrueSkill is very certain about his skill level), that certainty means you see a smaller change in your skill after the game.

2) Likewise, if you're a 19 playing with a 25 in Team Doubles, you're going to need to play against a team with a skill of approximately 44 or better, since the TrueSkill algorithm is comparing only the sum of your team's skill level. You might feel proud of beating a pair of 22s, but the game compares only the teams' collected skill levels.

3) Speaking of which, it bears repeating that the ONLY input to the TrueSkill algorithm is the skill level of each competitor, and whether each competitor won or lost. If you're the 19 in the previous example, and you got every single kill and never died, TrueSkill sees exactly the same thing as if your partner carried you to a 25-24 win.

4) A lot of people are complaining about Team Slayer, and I don't exactly blame you--I've won 90-95% of my Team Slayer matches for weeks, and just recently got past 20. Keep this in mind: TrueSkill (like any other ranking system) is at its best in a situation like Lone Wolves, where one player is matched against multiple opponents. This is because it gets to make five comparisons at the end of every match instead of one (it compares you to each other player), and can therefore arrive at the hypothetical true skill level for you much faster.

Two teams with multiple players on each team is a much worse situation. Not only is there only one comparison available, but that comparison is complicated by the fact that there are other people to take into consideration; since each person's performance naturally varies from their actual skill, there's an additional question as to whether the final result is due to the winning team having more skill, or due to the random elements of the game. The upshot of all this is that team games take much longer to level up in than individual games--and the larger the team, the longer it takes. This is an unavoidable consequence of the way skill is tracked in TrueSkill/Halo 3, and it's never going to turn into the more linear system we had in Halo 2.

  • 11.24.2007 9:55 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

Not sure if this has been said but in the first post you say that ingame stats don't affect someone's ranking. However, bungie has stated that if you start becoming a headshot god then you will rank up faster

  • 11.24.2007 10:11 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

okok tht is call an all but how do u make up ur own topic?

  • 11.24.2007 10:53 PM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

Thanks for your indepth thought and time taken to type this up but I after reading I am still left puzzled. Why would bungie switch the ranking system (from Halo 2 to Halo 3) to this complicated mathematical calculations which probably only a very small minority of the Halo gaming community can understand, even IF they take the time to, from the previous more simple ranking system? Hindsight is 20/20; but, Shouldn't bungie have expected thousands of disgruntled Halo gamers? What were the motives behind bunige overhauling the ranking system and the game interface to Microsoft based/developed programs and calculations? Thanks.

  • 11.24.2007 11:12 PM PDT

Posted by: BaDShoT is BK
Not sure if this has been said but in the first post you say that ingame stats don't affect someone's ranking. However, bungie has stated that if you start becoming a headshot god then you will rank up faster
Bungies statements were misunderstood. Better headshots skills was intended to represent a general increase in skill. This was corrected by Achronos in numerouse different threads and I believe quotes from him have been posted in this thread. In game stats are in fact not used at all to determine skill level changes.

  • 11.24.2007 11:52 PM PDT

Posted by: A_SexyPenguin
Thanks for your indepth thought and time taken to type this up but I after reading I am still left puzzled. Why would bungie switch the ranking system (from Halo 2 to Halo 3) to this complicated mathematical calculations which probably only a very small minority of the Halo gaming community can understand, even IF they take the time to, from the previous more simple ranking system? Hindsight is 20/20; but, Shouldn't bungie have expected thousands of disgruntled Halo gamers? What were the motives behind bunige overhauling the ranking system and the game interface to Microsoft based/developed programs and calculations? Thanks.
From my personal experience I found Halo 2's system to suffer from streaky play. You could move up or down several ranks at any given time depending on if you got on a winning or losing streak. This system is not as prone to that issue. Also in theory you are capable of moving through the ranks more quickly (you can actually skip multiple levels at a time). This has the benefit that you don't spend as much time playing and beating up on lower levels just to get to your 45.

Understanding the system really isn't important. Although I understand why people want to understand it, none of what we know here really matters. If we let the system work it would just do it's thing. It's true the system can be a little complicated if you try to learn exactly how it works, but the truth is the basic premise is very easy to understand.

Change often does take some getting used to, but that shouldn't stop us from changing. Although the Halo 2 system might have worked okay this system is supposed to improve upon it and I think we should always seek to improve from one point to the next.

  • 11.25.2007 12:05 AM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

Posted by: A_SexyPenguin
Thanks for your indepth thought and time taken to type this up but I after reading I am still left puzzled. Why would bungie switch the ranking system (from Halo 2 to Halo 3) to this complicated mathematical calculations which probably only a very small minority of the Halo gaming community can understand, even IF they take the time to, from the previous more simple ranking system? Hindsight is 20/20; but, Shouldn't bungie have expected thousands of disgruntled Halo gamers? What were the motives behind bunige overhauling the ranking system and the game interface to Microsoft based/developed programs and calculations? Thanks.


I cannot speak to Bungie's exact intentions in replacing the Halo 2 ranking system with the Halo 3 ranking system, but I can theorize as well as give several reasons why it would be a good idea.

A nonlinear ranking system has many, many perks. Not the least of which is that it deals better with upsets. I know the Halo 2 system had SOME factoring in of player skill when it calculated XP, but it didn't have nearly the negative effect that a major skill difference has in Halo 3. If you're continuously beating up on players 5 or more skill levels below you, but lose your occasional match to even or greater-leveled players, you don't DESERVE to rank up. And Halo 3 won't let you. Even if you win 25 matches in a row against lower-leveled players. You have to win the even or underdog matches. The way it should be.

Also, the nonlinear ranking system allows players to be streaky, or even just have good and bad days, without penalizing their enjoyment.

Imagine this scenario in Halo 2. You have a good day, winning many more matches than you lose. When you finish for the day, the system has you up 5 ranks from where you were. Problem is, you can't hang at that level on a bad day. And if tomorrow's a bad day, you'll spend all day long frustrated at the seemingly poor way you're playing. Not to mention you'll go on quite the losing streak, as well as drop several skill levels, which might wreck your confidence, snowballing into a longer slump.

Now imagine the same scenario in Halo 3. You go on the same win streak, and you only gain 2 levels. Since you're only 2 levels above where you were, you can actually play somewhat effectively at that level on a bad day. As a result, you might only lose a few more games than you win, allowing you to stay at that level. Or, you might lose quite a few more than you win, but it'll only bump you back down to your previous level. Your confidence is intact, and you continue to enjoy the game because you're playing on a proper skill level.



Bungie could have also simply realized that a nonlinear system affords the best mix of gaging true player skill, preventing overcrowding at the top, and having a meaningful ranking system. With a linear system, a 50 can be a 45 that has gone on a big streak against other 45s. In Halo 3's system, a 50 will beat a 45 90% of the time. Call me crazy, but I love the idea of having skill rank actually mean something.

  • 11.25.2007 12:34 AM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

Well I still think the system is broken in some sort of way. I understand that only wins can raise your current skill level, but I've seen by my own eyes how the system is unfair in some sort of way. My cuzin which gamertag is HolyTiTsmack has racked up a pretty long winning streak of maybe about 25 in team slayer, he started playing halo 3 today and has won 38 games and only 2 losses and yes you can check his game history. He won like seven games in a row while staying at the same rank which is unfair, then goes up and loses 1 game and goes back down again i think around LEVEL 14. Then wins five in a row again and stays the same rank for some stupid reason. I dunno just throwing in my 2 cents, but i think its stupid sometimes how someone can be having running riots, rampages, perfections (yes i know stats dont matter) and WIN 15 games in a row against ppl his own level and yet stay the same or go up one level when any normal human being would know his skills are on a completely different planet... its just STUPID.. if u have the knowledge to explain y then plz go ahead and do so...

  • 11.25.2007 1:25 AM PDT

Posted by: Vorpal Sword

4) A lot of people are complaining about Team Slayer, and I don't exactly blame you--I've won 90-95% of my Team Slayer matches for weeks, and just recently got past 20. Keep this in mind: TrueSkill (like any other ranking system) is at its best in a situation like Lone Wolves, where one player is matched against multiple opponents. This is because it gets to make five comparisons at the end of every match instead of one (it compares you to each other player), and can therefore arrive at the hypothetical true skill level for you much faster.
I don't know if the system is more efficient in a Lone Wolves scenerio over a Team Slayer scenerio (maybe this is true), but I think the system is very capable of moving you fast even in Team situations. I looked at your games and I can understand why you might think it is less efficient with teams. However, I had a VERY different experience.

I looked at both our games played and here is what I found:

Note: In the Team Slayer playlist I went into matchmaking all by myself

Stats for my first 32 games:
- 26 wins and 6 losses
- 82% win rate
- Skill level achieved = 20
- Average levels moved per game played = .625

Stats for your first 34 games:
- 32 wins and 2 losses
- 94% win rate
- Skill level achieved = 13
- Average levels moved per game played = .38

Jumping ahead it took you 96 games to reach level 21 (it only took me 30) and you only had 8 losses for a win rate of 91%. You leveled .21 levels per game

I reached level 39 in only 67 games played. I had 16 losses for a win rate of 76%. I leveled .58 levels per game.

I can only come up with two theories why I leveled so much faster then you.

1) Do to your higher win % and the conservative nature of the system you actually leveled slower to this point.

2) As far as I could tell you partied with the same 2 friends for every game played. Perhaps the lack of variaty in teammates or some other factor associated with parties caused you to level slower.

The point of all this is to say that I don't really know if team games are any less efficient then lone wolve games but I do know you can level much faster then you did. Although your statement may still be accurate, the system is capable of moving someone quickly in a Team game setting.

[Edited on 11.25.2007 1:42 AM PST]

  • 11.25.2007 1:38 AM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

Nice one! This is exactly what I was loking for, very impressive. Most FAQ's I read on things are very vague and wooly, and I find they dont help much which is why I tend to shy away from them nowadays, but this is very concise. All the questions I would have asked have been directly answered and this thread gives me a much better insight into how the ranking system seems to work.
Very helpful, thanks a million!!

  • 11.25.2007 4:14 AM PDT
  • gamertag:
  • user homepage:
  • last post: 01.01.0001 12:00 AM PDT

^^^ Daves hits the nail on the head. nice thread man very helpful :)

  • 11.25.2007 5:34 AM PDT