Rating Vs PR

[howlingwolfpaw]

Hi, I am trying to understand the difference from the rating number such as mine in TR (1783) and my PR 24.98. The large rating number changes with every game but I do not see its effects on the game towards my rank so what is its use?. I thought my PR changed every game as well (because for a while it steadily increased or decreased with each game) and that it was related some how but the past few games it has stayed the same win or lose. So how is the in game PR towards my rank calculated/ related. the Elo rating site just explains how the large number is calculated but not what it means or anything about the PR.

Based on prior game play I thought my rank would jump up this time since the other team was a high captain and 2 lst lts and we were a low captain and a 2nd lt. and me 1st lt. When I left the game it did take me automatically out of the game and said I lost connection if that means anything but I don't think it would.

I apologize if this has been asked before but after a quick search found no results.

Thank you.

[gibson]

Pr is calculated in as similar way to elo, it has the formula on the site that had the old agecom forums. Essentially what it boils down to is for the elo rating the number of games you have played is irrelevant to how your elo fluctuates, but with pr it takes the amount of games played into account. So you can get an account to major by winning your first five games, but after you've played 5000 games your pr won't really change at all.

[howlingwolfpaw]

so basically It does not think I played enough games to level up? I'm ok with that seems to be harder to get games the higher up you are.

[gibson]

No that's not what I meant to say..... The more games you play the harder it is to level up in pr, while in elo the amount of games played is irrelevant.... I summon @metis to explain it better

[howlingwolfpaw]

ok i see... half those games or more were from years ago when i was a bigger noob they just stay with me forever? I stopped playing for a longwhile and came back. haha so thats why people make seperate accounts

[[Armag] diarouga]

When you win a game, you get x elos and your opponents looses x elos depending on your ratings.
Your pr depends on your HC level, the amount of game you've played and youe opponent's level ofc.
It changes after every games (you can see it when you look at your HCs).

[Metis]

PR was used by ESO because they thought that Elo rating fluctuated too much and that players would get too frustrated when their rating dropped significantly after only one loss.

PR is calculated thusly: Points earned = Probable Skill - K * variance

Probable skill is your average skill level before the game.

Variance begins as a high number, which is why you can rank up a new account very quickly. After you have played several games the variance value starts to decrease. Play a few hundred games and this value becomes negligible to the equation.

K is a bit more complex and takes various other things into account.

Home City tier.

Tier 1: 0-9 <'-- only those players who kept a Level 1 civ before the Level 10 patch can have these now
Tier 2: 10-24
Tier 3: 25-100+

You get more points if you win using a lower tier home city than your opponent. This is the reason that a certain noob bashers delete a HC when it reaches level 25. If both you and your opponent have the same tier HC then this variable has no affect on the equation.

Civilization Tiers

Where a civilization falls in the tier system depends on the number of times it has been played and the number of times players have won with it. You can get some idea of what this value might be by looking at the overall stats page on eso.com. For instance, in TAD treaty I can see that France has been used in 30% of games and has won in 54% of these. Therefore, a player using France will get less PR for a win than one using Sioux, which has ben played in only 1% of games and has won in only 45% of these.

In team games all the data for the team is combined into one virtual player and then used to give all players the same PR value change. Therefore, if want your PR to increase significantly in a master sergeant game, then make sure that noob-bashing colonel is playing on the other team.

As far as I can tell the exact algorithm for PR has never been published. However, the Elo rating formula that the Japan Player Community uses is published on their website.

??R : difference value of rating, each winner/loser gains/loses
RW : winner's current rating (in team, sum of team's rating)
RL : loser's current rating (in team, sum of team's rating)
N : number of players

??R = 16 + (RL - RW) x 2 / N x (16 / 400) if (??R >' 31) ??R = 31' if (??R <' 1) ??R = 1'

Example:

?1vs1
winner: 1700' loser: 1600
?? ??R = 16 + (1600 - 1700) x (16 / 400) = 12
winner: 1712, loser: 1588

?3vs3
winners: 1500, 1600, 1700' losers: 1700, 1800, 1900
?? ??R = 16 + (5400 - 4800) x 2 / 6 x (16 / 400) = 24
winner: 1524, 1624, 1724' loser: 1676, 1776, 1876


http://eso.com/faq_mp.aspx#ranks

http://aoe3.jpcommunity.com/rating2/

[yurashic]

When you win a game, you get x elos and your opponents looses x elos depending on your ratings.
Your pr depends on your HC level, the amount of game you''ve played and youe opponent''s level ofc.
It changes after every games (you can see it when you look at your HCs).

HC level? Seriously?

That''s why all cheaters use low level home cities...

[Metis]

Your pr depends on your HC tier

HC level? Seriously?

Thats why all [strike]cheaters[/strike] noob bashers use low level home cities...

Exactly. Check out this players HCs.

http://aoe3.jpcommunity.com/rating2/player?n=azon

[drumal]

wow.. never heard this before

So if I use in sup 1v1 game spain and beat otto ( basically any sux civ vs top civ) I will get more points than from equal mu?

[gibson]

wow.. never heard this before

So if I use in sup 1v1 game spain and beat otto ( basically any sux civ vs top civ) I will get more points than from equal mu?

No the actual civs are irrelevant

[[Armag] diarouga]

Nope, that is about hc lvl, 0-9, 10-24, 25-131 as metis said.

[Wuangaga]

wow.. never heard this before

So if I use in sup 1v1 game spain and beat otto ( basically any sux civ vs top civ) I will get more points than from equal mu?

Well the devs didn''t intend civs to be way worse than others, so nope.

[drumal]

oh.. i see thx

[Metis]

wow.. never heard this before

So if I use in sup 1v1 game spain and beat otto ( basically any sux civ vs top civ) I will get more points than from equal mu?

No the actual civs are irrelevant

Actually the civs do matter (for PR, not Elo rating). Like I pointed out, the number of times a civ wins overall is also included in the calculations. In treaty, France is played more than any other civ and also wins more. France is rated as the most win civ. Conversely, in treaty, Sioux is hardly played at all and when it is has less than 50% wins. Sioux is rated as the least win civ. So, all other things equal, if you play Sioux and win against France you will gain more PR than if you play France and win against Sioux.

Here are the overall civ stats:

http://eso.com/player_stats_yp.aspx

[momuuu]

playing with like lvl 20 HC on an account that has barely played any games will level you up crazy fast. If you actually wanna infalte your pr that way, go bash noobs aswell. Thats a reliable way to profit from noobbashing. And also a reliable way to scout scrubs :p

[howlingwolfpaw]

Thank you all, So why is it that my rank has stagnated? is it that I'm so close to being a captain and there are only so many captain slots available? I do play india on a 131 deck. 55% win ratio, ( rated at 8% used 44% win) my variance just too low? but it was moving pretty quickly just prior to the stagnation.

[Metis]

Thank you all, So why is it that my rank has stagnated?

Once you play fifty games or so your PR isn''t going to change that much per win or loss, especially in team games.

Also, there are time periods when no-one''s rank changes do to MS tinkering with the servers. This rank freeze has happened dozens of times over the years.

[Einfein]

HC level? Seriously?

That''s why all cheaters use low level home cities...

I''ve always found it annoying when people with a similar rating to mine join one of my games and immediately have a sub level 25 HC picked. I used to just kick them, but now I don''t bother. True it means my rating will sometimes dip a larger amount then I would like if I lose, but in the end it just adds to the challenge.

[[Armag] diarouga]

Imperatordivi was playing with lvl10 HCs and he was colo/briga while his real rank was captain/major lol.

[_venox_]

Metis knows alot of shit wow.

[yurashic]

Imperatordivi was playing with lvl10 HCs and he was colo/briga while his real rank was captain/major lol.

He was a guy who played against pr 10 with 12/10 Aztec rush, he was surely worse than captain/major...

[tibia]

I also think you go quiker up in rank after a decay?, maybe variance resets until you reach the pr you once was.

[r4go]

I also think you go quiker up in rank after a decay?, maybe variance resets until you reach the pr you once was.

yes, if your acc be reset for inactivity, when you will be back it works as a new acc.

[[Armag] diarouga]

Imperatordivi was playing with lvl10 HCs and he was colo/briga while his real rank was captain/major lol.

He was a guy who played against pr 10 with 12/10 Aztec rush, he was surely worse than captain/major...

He wasnt a noob basher.